Aldol condensations over reconstructed Mg-Al hydrotalcites: Structure-activity relationships related to the rehydration method

Article abstract of DOI:10.1002/chem.200400409

Two different rehydration procedures in the liquid or gas phase have been applied to reconstruct mixed oxides derived from calcined hydrotalcite-like materials to be used as catalysts for aldol condensation reactions. The as-synthesized hydrotalcite, its decomposition product, as well as the reconstructed solids upon rehydration were characterized by XRD, N₂ adsorption, He pycnometry, FTIR, SEM, TEM, ²⁷Al MAS-NMR and CO ₂-TPD (TPD = temperature-programmed desorption). Compared to the Mg-Al mixed oxide rehydrated in the gas phase (HT-rg), that rehydrated in the liquid phase (HT-r1) exhibits a superior catalytic performance with respect to the aldol condensation of citral with ketones to yield pseudoionones and in the self-aldolization of acetone. The textural properties of HT-r1 and HT-rg differ strongly and determine the catalytic behavior. A memory effect led to a higher degree of reconstruction of the lamellar structure when the mixed oxide was rehydrated in the gas phase rather than in the liquid phase, although liquid-phase rehydration under fast stirring produced a surface area that was 26 times greater. This contrasts to typical statements in the literature claiming a higher degree of reconstruction in the presence of large amounts of water in the medium. CO₂-TPD shows that the number of OH^- groups and their nature are very similar in HT-rg and HT-r1, and cannot explain the markedly different catalytic behavior. Accordingly, only a small fraction of the available basic sites in the rehydrated samples is active in liquid-phase aldol condensations. Our results support the model in which only basic sites near the edges of the hydrotalcite platelets are partaking in aldol reactions. Based on this, reconstructed materials with small crystallites (produced by exfoliation during mechanical stirring), that is, possessing a high external surface area, are beneficial in the reactions compared to larger crystals with a high degree of intraplatelet porosity.

Full text of DOI:10.1002/chem.200400409

Aldol Condensations Over Reconstructed Mg–Al Hydrotalcites:
Structure–Activity Relationships Related to the Rehydration Method
Sꢀnia Abellꢁ,[a] Francesc Medina,*[a] Didier Tichit,[b] Javier Pꢂrez-Ramꢃrez,[c]
Johan C. Groen,[d] Jesffls E. Sueiras,[a] Pilar Salagre,[a] and Yolanda Cesteros[a]
Abstract: Two different rehydration
procedures in the liquid or gas phase
have been applied to reconstruct mixed
oxides derived from calcined hydrotal-
cite-like materials to be used as cata-
lysts for aldol condensation reactions.
The as-synthesized hydrotalcite, its de-
composition product, as well as the re-
constructed solids upon rehydration
were characterized by XRD, N2 ad-
sorption, He pycnometry, FTIR, SEM,
TEM, 27Al MAS-NMR and CO2-TPD
(TPD=temperature-programmed de-
sorption). Compared to the Mg–Al
mixed oxide rehydrated in the gas
phase (HT-rg), that rehydrated in the
liquid phase (HT-rl) exhibits a superior
catalytic performance with respect to
the aldol condensation of citral with
ketones to yield pseudoionones and in
the self-aldolization of acetone. The
textural properties of HT-rl and HT-rg
differ strongly and determine the cata-
lytic behavior. A memory effect led to
a higher degree of reconstruction of
the lamellar structure when the mixed
oxide was rehydrated in the gas phase
rather than in the liquid phase, al-
though liquid-phase rehydration under
fast stirring produced a surface area
that was 26times greater. This contrasts
to typical statements in the literature
claiming a higher degree of reconstruc-
tion in the presence of large amounts
of water in the medium. CO2-TPD
shows that the number of OHÀ groups
and their nature are very similar in
HT-rg and HT-rl, and cannot explain
the markedly different catalytic behav-
ior. Accordingly, only a small fraction
of the available basic sites in the rehy-
drated samples is active in liquid-phase
aldol condensations. Our results sup-
port the model in which only basic
sites near the edges of the hydrotalcite
platelets are partaking in aldol reac-
tions. Based on this, reconstructed ma-
terials with small crystallites (produced
by exfoliation during mechanical stir-
ring), that is, possessing a high external
surface area, are beneficial in the reac-
tions compared to larger crystals with a
high degree of intraplatelet porosity.
Keywords: rehydration
reaction layered compounds
memory effect · stirring speed
·
aldol
·
·
Introduction
[a] S. Abellꢀ, Dr. F. Medina, Dr. J. E. Sueiras, Dr. P. Salagre,
Dr. Y. Cesteros
In recent years, hydrotalcite-like compounds (HTlcs) have
been used in numerous reactions and applications, such as
catalyst precursors or supports, ion-exchangers, stabilisers
Dept. de Quꢁmica i Enginyeria Quꢁmica, Universitat Rovira i Virgili,
43007, Tarragona (Spain)
Fax : (+34)977-559-621
and adsorbents.[1–3] These materials are layered double hy-
E-mail: fmedina@etse.urv.es
m+
droxides with the general formula [M2+nM3m(OH)2(n+m)
]
[b] Dr. D. Tichit
[AxÀ]m/x·yH2O, where M2+ and M3+ are divalent and triva-
lent metal cations, respectively, A represents the x-valent
anion (typically carbonate) required to compensate the net
positive charge of the brucite-like layers, and y is the
number of water molecules in the interlayer space. The ratio
m/(m+n) may vary from 0.17 to 0.33, depending on the par-
ticular combination of divalent and trivalent metals.
Laboratoire des Matꢂriaux Catalytiques et Catalyse en Chimie Or-
ganique, UMR 5618 CNRS-ENSCM
8 rue de LꢃEcole Normale, 34296 Montpellier Cedex 5 (France)
[c] Dr. J. Pꢂrez-Ramꢁrez
Yara Technology Centre Porsgrunn, Catalysis and Nitric Acid Tech-
nology
P.O. Box 2560, N-3908, Porsgrunn (Norway)
[d] J. C. Groen
Thermal decomposition of Mg–Al HTlcs leads to a well-
dispersed mixture of magnesium and aluminium oxides
(MgAlO). These mixed oxides show a memory effect, a
Applied Catalyst Characterization, DelftChemTech, Delft University
of Technology
Julianalaan 136, 2628 BL, Delft (The Netherlands)
728
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200400409
Chem. Eur. J. 2005, 11, 728 – 739

FULL PAPER
property by which they can recover the original lamellar
structure if they come into contact with water vapour or are
immersed in water under a decarbonated atmosphere. This
leads to meixnerite (magnesium aluminium hydroxide hy-
drate), that is, HTlcs intercalated with OHÀ as compensating
anions in the interlayer.[4] These rehydrated materials have
been applied to a number of base-catalysed reactions on ac-
count of their Brønsted basic character.[5–7] Applications in-
clude self- and cross-aldol condensation of aldehydes and
ketones,[8–11] Knoevenagel and Claisen–Schmidt condensa-
tions,[12–14] Michael additions,[15] etc.
It has been clearly established that the precalcination
temperature is a determining parameter governing the re-
construction, which is complete when the formation of a
spinel-like phase, for example, MgAl2O4 in case of Mg–Al
HTlc, is avoided. Reconstruction cannot be completed if
this phase has been formed, even after several days of rehy-
dration.[16,17]
obtain pseudoionone or methyl-pseudoionone, respectively,
and in the self-aldolization of acetone.
Results and Discussion
Material characterization
Chemical composition: Analysis of the as-synthesized
sample, HT-as, by inductively coupled plasma (ICP) spec-
trometry as well as by thermogravimetry (TG) revealed the
following chemical composition: [Mg0.777Al0.223(OH)2]-
(CO3)0.113·0.575H2O, indicating the good agreement between
the nominal ratio in the liquid and the actual ratio in the
solid. The amount of sodium in the as-synthesized hydrotal-
cite was less than 0.1 wt%, indicating that washing of the
solid after filtration was effective. Furthermore, nitrogen
was not detected by elemental analysis.
The catalytic performance of reconstructed HTlcs as a
function of the rehydration procedure applied is not well-
understood. Corma and co-workers[18,19] approached this
aspect by carrying out the rehydration of calcined Mg–Al
hydrotalcite in a stream of nitrogen saturated with water
vapour (gas-phase rehydration), or, alternatively, by adding
water directly to the calcined material (liquid-phase rehy-
dration). Both samples, which had a very similar level of re-
generation with regard to the layered structure according to
XRD, displayed practically the same activity in the conden-
sation reactions of citral and acetone or methyl ethyl
ketone.[19] However, in the condensation reaction of citral
and methyl ethyl ketone, a higher selectivity for methylpseu-
doionones was obtained with the gas-phase rehydrated
sample, which suggests a comparatively larger ratio of
Brønsted to Lewis basic sites in the latter sample compared
to the liquid-phase rehydrated sample.
X-ray diffraction: Figure 1 shows the XRD patterns of the
as-synthesized Mg–Al hydrotalcites (HT-as), calcined hydro-
talcite (HT-c), rehydrated hydrotalcite in the gas phase
Experiments that employed the two types of rehydration
procedure over decomposed Mg–Al HTlc have also been
performed by de Jong and co-workers.[20–22] In these studies,
rehydration in water vapour led to a lower crystallinity and
specific surface area than rehydration by immersion (57
versus 200 m2 gÀ1). The as-synthesized hydrotalcites dis-
played specific surface areas in the range of 90–100 m2 gÀ1.
Figure 1. X-ray diffraction patterns of the materials investigated in this
study.
A
similar observation was also reported by Figueras
et al.;[10,23–25] the surface area of the mixed oxide (95 m2 gÀ1)
decreased to 20 m2 gÀ1 after rehydration of the sample in the
vapour phase, which led to a restoration of the layered
structure. According to these authors, this accounted for a
limitation of the rehydration process by the diffusion of
water within the particles, and the surface corresponds to
the external area of the crystals.
These features suggest that the rehydration process of the
crystallites is certainly a key parameter governing the
number and the strength of active OHÀ sites in base-cata-
lysed reactions. Accordingly, the present study was under-
taken to derive structure–activity relationships in differently
reconstructed Mg-Al-HTlcs for various aldol condensation
reactions, such as citral and acetone or 2-butanone, to
(HT-rg) and rehydrated hydrotalcite in the liquid phase
(HT-rl) (at 500 rpm for 1 h). The as-synthesized sample
shows a pure hydrotalcite phase (JCPDS 22-700). Upon cal-
cination of the material at 723 K, the layered structure was
destroyed and a mixed oxide phase Mg(Al)Ox was obtained,
corresponding to the periclase structure (JCPDS 87-0653).
After calcination and intercalation of OHÀ ions by rehydra-
tion in the gas and liquid phases, the original layered struc-
ture was recovered, which corresponded to meixnerite
(JCPDS 35-0965). However, the XRD pattern of HT-rl
shows broader peaks than that of HT-rg, indicating a lower
crystallinity, namely as a consequence of the lower crystal-
lite size in the former sample (see microscopy). This can be
tentatively attributed to the effect of mechanical stirring
Chem. Eur. J. 2005, 11, 728 – 739
www.chemeurj.org
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
729

F. Medina et al.
during the rehydration process in the liquid phase, which
can break and exfoliate the hydrotalcite-like platelets to in-
crease the surface area of the solid, and/or to the mechanism
of water diffusion in the galleries.
N2 adsorption: N2 adsorption experiments were performed
to derive detailed information on pore size distributions.
Upon calcination of HT-as, the porosity substantially in-
creases leading to a more pronounced hysteresis loop in the
isotherm of HT-c (Figure 2), which indicates formation of
extra (meso)porosity. This result is supported by the higher
surface area and total pore volume in HT-c (210 m2 gÀ1)
compared to HT-as (57 m2 gÀ1) (Table 1). In addition, forma-
tion of some microporosity in HT-c can also be concluded.
Application of the BJH model to the adsorption branch
of the isotherm further indicates that thermal decomposition
Figure 3. BJH pore size distribution derived from the adsorption branch
of the N2 adsorption isotherms.
particular, that the smaller mes-
opores which have been created
extend into lower pore sizes. It
is thought that these lower pore
sizes are probably responsible
for the significant increase in
BET surface area. This porosity
development upon calcination
is in agreement with results pre-
viously reported by de Jong and
co-workers.[21]
Rehydration in the liquid or
gas phases leads to remarkable
differences in the porous char-
acteristics of the obtained mate-
rial. The sample rehydrated in
the gas phase (HT-rg) does not
show a significant porosity, but
has a low BET surface area
(15 m2 gÀ1) and pore volume
(0.1 cm3 gÀ1). A similar but less
dramatic decrease in the BET
surface area (from 253 to
57 m2 gÀ1) was observed in ref-
erence [21], and possibly results
Figure 2. N2 adsorption isotherms of the different samples.
Table 1. Textural and chemical properties of the materials.
from slightly different rehydra-
[a]
[b]
[b]
[c]
[d]
Sample
SBET
Smeso
Vmicro
Vp
1He
[gcmÀ3
H[e]
[%]
C[e]
[%]
[mÀ2 gÀ1
]
[mÀ2 gÀ1
]
[cmÀ3 gÀ1
]
[cmÀ3 gÀ1
]
]
tion conditions. Contrarily, re-
hydration in the liquid phase
(HT-rl) leads to a remarkable
BET surface area of 270 m2 gÀ1,
which is even higher than the
surface area of the calcined ma-
terial, whereas the pore volume
of HT-rl is slightly lower than
HT-as
HT-c
HT-rg
HT-rl
57
210
15
57
180
15
0.00
0.01
0.00
0.00
0.4
0.8
0.1
0.7
1.93
2.73
2.15
2.56
4.04
0.88
4.36
3.24
1.79
0.35
0.49
1.03
270
270
[a] BET method. [b] t method. [c] Volume at p/p0 = 0.99. [d] He pycnometry. [e] Elemental analysis.
of the as-synthesized hydrotalcite induces the development
of additional porosity covering a broad range of pore sizes,
with a maximum at ~50 nm.
Interestingly, the pore-size distribution in HT-c is similar
to that in HT-as (Figure 3). This led to the conclusion, in
in the calcined material. This result suggests a markedly dif-
ferent porous structure for both materials, as confirmed by
the BJH pore size distribution in Figure 3, which shows that
the majority of pores are smaller than 50 nm with a distribu-
tion centered at ꢀ10 nm, without any appreciable micro-
730
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 728 – 739

Structure–Activity Relationships of Reconstructed Mg–Al Hydrotalcites
FULL PAPER
porosity. Although no details were provided by de Jong and
co-workers[20] on the pore size evolution upon rehydration
of Mg–Al HTlc in the liquid phase, the increase in surface
area accompanied by a decreased pore volume suggests a
similar change in the porous characteristics.
The difference in the porous characteristics of the liquid-
phase rehydrated material could be produced by exfoliation
of the meixnerite crystals during the mechanical stirring.
This was investigated in additional experiments by changing
the stirring speed and the rehydration time (Table 2); how-
ever, the same magnet was always used.
Table 2. Effect of the stirring speed and time over rehydrated hydrotal-
cites.
Figure 4. Thermogravimetric analysis of the as-synthesized and rehydrat-
ed hydrotalcites.
[a]
Sample
Rehydration
time[h]
Stirring
SBET
Reconstruction[b]
[%]
speed[rpm] [mÀ2 gÀ1
]
HT-rl-100
HT-rl-300
HT-rl
HT-rl-700A
HT-rl-700B
HT-rl-700C
1
1
1
1
0.16
6
100
300
500
700
700
700
203
229
270
374
241
401
66
74
83
89
67
92
ples. The decomposition profiles are in good agreement with
those in the literature for hydrotalcite-like compounds,[26,27]
with a total weight loss in the range of 34–45%. Similar be-
havior was observed for the three samples, with two distinct
weight-loss processes. The first weight loss at temperatures
below 473 K is attributed to the loss of physically adsorbed
and interlayer water molecules, whereas the second weight
loss (473–773 K) originates from the dehydroxylation of the
brucite-like sheets and decomposition of carbonates in the
interlayer. The first weight loss of ꢀ18% for the rehydrated
samples is similar in both cases. However, the second weight
loss was 5% higher in HT-rg (21.7%) than in HT-rl
(16.8%). This indicates a similar amount of water molecules,
but an increase of compensating OHÀ ions between the bru-
cite-like sheets of the sample rehydrated in the gas phase
compared to that rehydrated in the liquid phase. This nicely
indicates that the degree of rehydration in the gas phase is
higher than in the liquid phase. Accordingly, the degree of
reconstruction was estimated to be 94% in HT-rg and 83%
in HT-rl. These results are in agreement with the C and H
elemental analysis of the samples in Table 1 that show a
4.36 wt% and 3.24 wt% of hydrogen for HT-rg and HT-rl,
respectively. The carbon content in the samples indicates
that not all the carbonates are decomposed during the calci-
nation process. Furthermore, the increase in the amount of
carbon detected in HT-rl, indicates a major contamination
with CO2 during rehydration in liquid phase.
[a] BET method. [b] Calculated by TGA.
As expected, a higher surface area for the rehydrated ma-
terials was observed when the stirring speed was increased
from 100 to 700 rpm. This can be explained by the higher
degree of exfoliation obtained by the rupture of particles.
When a stirring speed of 700 rpm was used with different re-
hydration times, the same effect was detected.
Helium pycnometry (Table 1) provides information on the
real (skeleton) density of the material, independently of the
accessible porosity. As expected, the density of the calcined
material (2.73 gcmÀ3) is much higher than that of the as-syn-
thesized hydrotalcite (1.93 gcmÀ3). The density of the rehy-
drated hydrotalcites presents important differences: it is
higher in HT-rl (2.56 gcmÀ3) than in HT-rg (2.15 gcmÀ3).
This fact can be explained taking into account the structure
of the reconstructed hydrotalcites. If the rehydration proce-
dures would lead to total reconstruction, HT-rl and HT-rg
should possess a very similar density to HT-as. However, the
densities obtained in the rehydrated hydrotalcites lie be-
tween those of HT-as and HT-c, which strongly suggests that
the layered structure was partially recovered. That the den-
sity in HT-rg (2.15 gcmÀ3) is lower than that in HT-rl
(Table 1) indicates a more efficient reconstruction in the
gas-phase rehydrated sample, in agreement with the results
of XRD and other techniques described below. In contrast
to a previous hypothesis,[24] this indicates that the diffusion
of water within the particles is not the limiting factor of the
rehydration process. The low surface area of HT-rg is proba-
bly caused by steric hindrance produced by the OHÀ and
water molecules in the interlayer space.
IR spectroscopy: Figure 5 shows the IR spectra of the dried
samples. The FTIR spectra of various hydrotalcite-like ma-
terials have been widely discussed elsewhere.[3,28–33] The
spectrum of HT-as exhibits
a typical broad band at
3471 cmÀ1, which is attributed to the stretching mode of hy-
drogen-bonded hydroxy groups from the brucite-like layers
and interlayer water. The shoulder at ꢀ3000 cmÀ1 is as-
signed to hydrogen bonding between water and carbonate in
the interlayer, whereas the band at 1643 cmÀ1 is the H2O
bending vibration. The vibration of the carbonates (asym-
metric stretching, n3) appears at 1374 cmÀ1 and could be as-
signed to interlayer carbonates (chelating or bridging biden-
Thermogravimetric analysis: Figure 4 shows the thermogra-
vimetric analysis of the as-synthesized and rehydrated sam-
Chem. Eur. J. 2005, 11, 728 – 739
www.chemeurj.org
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
731

F. Medina et al.
tributed to the vibrations of the
MgO and Al2O3. Exposure of
the calcined sample to a flow of
wet gas (HT-rg in Figure 5) in-
duces a recovery of the hydro-
talcite-like structure. The same
effect occurs when the calcined
sample is rehydrated in the
liquid phase (HT-rl in Figure 5).
The bands at 3050 and
1640 cmÀ1 reappear in both
cases, showing a significantly
larger amount of water mole-
cules when rehydration is car-
ried out in the gas phase. This
means that the reconstruction
of the hydrotalcite-like struc-
ture is greater than in the liquid
phase. In the high-energy
region, a band at 634 cmÀ1 is
present and could be assigned
to the Mg–OH translation.[32]
The bands at 770 and 1066 cmÀ1
can be assigned to the transla-
Figure 5. Fourier-transform infrared spectra of the different samples.
tion and deformation modes re-
spectively, of the hydroxy
groups influenced by the Al3+
tate). The vibration at 1515 cmÀ1 is ascribed to a reduction
in the symmetry caused by the presence of monodentate
carbonates (nasym O-C-O) interacting with Mg2+, as reported
elsewhere,[3,29–30] whereas the band at 1740 cmÀ1 is the bend-
ing mode of water. The low frequency region shows a band
at about 560 cmÀ1, corresponding to the translation modes
of hydroxy groups, influenced by Al3+ cations. The band at
870 cmÀ1 is characteristic for the out-of-plane deformation
of carbonate (n2), whereas the in-plane bending is located at
680 cmÀ1 (n4). Calcination of this material (Figure 5, HT-c)
shows that water has been virtually removed (see C and H
analysis in Table 1) because of the disappearance of the
bands at 1643 (corresponding to the water bending vibra-
tion), 1739 (water bending vibration restricted in the inter-
ions. The band at 1370 cmÀ1, which is recovered as a sym-
metric peak, corresponds also to interlayer carbonate
(mainly bidentate carbonates),[30] indicating that a significant
amount of carbonates is still present after calcination/rehy-
2À
dration (probably, by introduction of CO3 in the recovery
process). This result is in agreement with C,H analysis
(Table 1): where the amount of carbon and hydrogen and is
1.03 wt% and 3.24 wt%, respectively, in HT-rl and
0.49 wt% and 4.36 wt%, respectively, in the HT-rg samples.
Figure 5 also shows the spectrum of the liquid-phase rehy-
drated sample after CO2 adsorption at 353 K (HT-rl-CO2 in
Figure 5), under the same conditions as the CO2-TPD
(TPD=temperature-programmed desorption) experiments.
The only difference between this latter sample and HT-rl is
the shoulder at about 1485 cmÀ1, which is assigned to the
symmetric O-C-O stretching mode of bicarbonate anions.
When CO2 is adsorbed on a base and at pH values up to 8,
2À
layer) and 3050 cmÀ1 (interaction H2O-CO3 in the interlay-
er). The intensity of the band at about 3470 cmÀ1 also de-
creases by dehydroxylation. The carbonate band suffers
from a rearrangement in the interlamellar space, the band at
1374 cmÀ1 decreases in intensity and two peaks at about
1515 and about 1400 cmÀ1 are observed because of the inter-
2À
À
CO3 formation is more favourable than HCO3 forma-
tion.[34] Accordingly, before adsorption of CO2, a rehydrated
sample shows a carbonate peak because only stronger basic
sites are evaluated. After CO2 adsorption, bicarbonates
result from remaining weaker OHÀ groups.
2À
action between CO3 and Mg2+. The band at 1458 cmÀ1 is
attributed to the adsorbed carbonate upon thermal decom-
position. It can be concluded that the applied calcination
temperature is not sufficient to completely eliminate hy-
droxyls and carbonates, although the hydrotalcite phase is
destroyed. This fact is in agreement with the results in
Table 1, which show 0.35 wt% and 0.88 wt% of carbon and
hydrogen, respectively, in the calcined sample. The bands
below 1000 cmÀ1 are the vibration modes for Mg–O and Al–
O in the mixed oxide formed.[32] The band at 457 cmÀ1 is at-
Scanning electron microscopy: SEM images were recorded
to investigate the morphology of the different samples
(Figure 6). The micrograph of the as-synthesized hydrotal-
cite shows a well-developed layered structure. The mixed
oxide obtained upon calcination at 723 K maintains the la-
mellar structure and the morphology seems to be similar to
that of HT-as. When rehydration was performed in the
732
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 728 – 739

Structure–Activity Relationships of Reconstructed Mg–Al Hydrotalcites
FULL PAPER
gesting that the (partial) recon-
struction of the HTlc structure,
which (based on the signal in-
tensities) is less complete in
HT-rl (AlOh/AlTd = 3.9) than in
HT-rg (AlOh/AlTd
= 4.4), in
good agreement with the TGA
and XRD results.
Temperature-programmed de-
sorption of CO2: CO2-TPD is
typically applied to determine
the density and strength of
basic sites in mixed oxides de-
rived from the calcination of
hydrotalcites.[35–37] To establish a
correlation between the catalyt-
ic activity and the textural and
basic properties, experiments
were carried out over both re-
hydrated samples. From the
TPD of CO2 in Figure 10, two
peaks were identified indicating
Figure 6. Scanning electron micrographs of the as-synthesized and rehydrated hydrotalcites.
liquid phase, a sample with a high degree of exfoliation was
obtained, as mentioned above. The shape is preserved with
respect to the as-synthesized material, and the smaller parti-
cles generated by mechanical stirring in the rehydration pro-
cedure formed very thin platelets. Contrarily, HT-rg is built
of complex aggregates of hydrotalcite platelets that are sig-
nificantly thicker than those in HT-rl. This is in agreement
with the higher crystallinity (XRD) and markedly lower sur-
face area (N2 adsorption) of HT-rg.
that two types of basic sites can be distinguished in the rehy-
drated Mg–Al mixed oxides. The main peak is centered
around 673–693 K, whereas the other one has its maximum
at ꢀ823 K. The total number of basic sites is higher
(803.9 mmolgcatÀ1) when rehydration is carried out in the
Transmission electron microscopy: HRTEM was performed
to obtain further information on the different morphologic
features of the samples (Figure 7). The results are in agree-
ment with those obtained by other characterization techni-
ques. In the calcined sample (HT-c), clearly discernible pla-
telets are shown. This suggests that the platelet morphology
does not collapse, even if the layered hydrotalcite structure
is no longer present. HT-rg presents much thicker platelets
than HT-rl, indicating a relatively high degree of sintering in
the former sample. Moreover, the micrographs suggest a
better reconstruction of the lamellar phase in HT-rg than in
HT-rl (see also Figure 8). The crystalline phase is observed
over a much wider range, which is in agreement with our
previous observations.
Nuclear magnetic resonance of 27Al: The as-synthesized
sample shows the majority of Al in octahedral positions, as
evidenced by the signal at dꢀ10 ppm in Figure 9. The small
contribution at dꢀ80 ppm is attributed to a spinning side
band. Upon calcination, the contribution of tetrahedrally co-
ordinated Al increases at the expense of the octahedrally
coordinated Al, correlating with the collapse of the lamellar
structure. Rehydration in both liquid and gas phases shows
a partial recovery of the initial as-synthesized pattern, sug-
Figure 7. Low-magnification transmission electron micrographs of the as-
synthesized and rehydrated hydrotalcites.
Chem. Eur. J. 2005, 11, 728 – 739
www.chemeurj.org
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
733

F. Medina et al.
liquid phase (HT-rl) than when
it is carried out in the gas phase
(HT-rg)
(437.7 mmolgcatÀ1).
Based on the CO2 profiles, the
basic nature of the OHÀ ions
appears to be similar in HT-rl
and HT-rg, but the number of
basic sites in HT-rl is double
that in HT-rg (Table 2). The
contribution of the smaller
peak with respect to the total
number of basic sites is about
20% in both cases. The CO2
uptake results from different
types of carbonate coordination
in the interlayer space. In this
sense, different species have
been identified, such as mono-
Figure 8. High-magnification transmission electron micrographs of the rehydrated hydrotalcites.
dentate, bidentate, or bicarbonate anions. Monodentate and
bidentate carbonate formation involves low-coordinate
oxygen anions, which are in turn thus strong basic sites.[38,39]
Bicarbonate requires surface hydroxy groups. We suggest
that the first peak, which decomposes at temperatures of
about 700 K, can be attributed to the contribution of mainly
bidentate carbonates, together with bicarbonate species, on
the catalyst surface, whereas the smaller peak above 800 K
is attributable to monodentate species. This result is in
agreement with our FT-IR results. Taking into account the
TGA results and comparing the amount of OHÀ with the
amount of CO2 desorbed (Table 2), the proportion of basic
sites detected by the CO2 molecules is approximately 13%
and 30% for HT-rg and HT-rl, respectively, despite the
higher degree of reconstruction, and thus the higher number
of OHÀ groups in HT-rg (see Figure 4). This indicates that
1) not all of the total OHÀ groups are probed by CO2 mole-
cules and 2) the OHÀ groups in the HT-rl sample are more
accessible (owing to the higher porosity and specific surface
area). Under our experimental conditions, it is probable that
the OHÀ ions probed by CO2 are located at the edges of the
platelets. From the quantification in Table 3, the CO2 ad-
sorption with respect to the bulk Al content shows that ap-
proximately 11.6% and 24.5% of the OHÀ sites, for HT-rg
and HT-rl, respectively, are also evaluated.
Figure 9. 27Al MAS-NMR spectra of the as-synthesized, calcined, and re-
hydrated samples.
Table 3. Quantification of the CO2-TPD profiles for the rehydrated hy-
drotalcites.
Sample
Peak at
693 K[%]
Peak at
823 K[%]
Total CO2
CO2/Al ratio[a]
evolved[mmolgÀ1
]
HT-rg
HT-rl
80
78
20
22
438
804
0.116 (0.023)[b]
0.245 (0.054)[b]
[a] Mol of total adsorbed CO2 per mol of Al in the HT. [b] Between
brackets, contribution of peak at 823 K.
These results are in fair agreement with those determined
from thermogravimetric analysis. The higher accessibility
shown by the HT-rl sample could be explained by the fact
Figure 10. CO2 uptake during temperature-programmed desorption ex-
periments over the rehydrated hydrotalcites. CO2 adsorption at 353 K for
1 h.
734
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 728 – 739

Structure–Activity Relationships of Reconstructed Mg–Al Hydrotalcites
FULL PAPER
that it has approximately 18 times more surface area than
HT-rg (Table 1). Taking into account the higher SBET of HT-
rl, the difference in basicity is just incremental. This is
strong evidence that only the edges of the platelets are oper-
ative, namely, the active sites in aldol condensations (see the
section on structure–activity relationships). Therefore, intui-
tively there are two factors that are related: porosity and
crystal size, although it seems that the relatively small crys-
tal size is the activity-directing factor in these reactions. In
accordance with these results, Figure 11 shows a representa-
tive structure that highlights the main differences between
HT-rg and HT-rl.
Catalytic performance
It is well known that the first step in the aldol condensation
reaction is the extraction of a proton from the a-carbon
atom of a ketone/aldehyde by a basic site of the catalyst to
generate a reactive enolate. The nucleophilic enolate attacks
the carbonyl group of the substrate to produce an intermedi-
ate alkoxide. Subsequently, the alkoxide deprotonates a
water molecule, and thus creates a hydroxide and a b-hy-
droxyaldehyde or aldol product. After that and depending
on the reaction conditions, a
Figure 11. Schematic representation of the different reconstruction mech-
anism in HT-rl and HT-rg based on physico-chemical characterization re-
sults.
dehydration process may occur
to yield the final conjugated al-
dehyde. In this sense, materials
obtained from hydrotalcite-like
compounds, which show both
Lewis and Brønsted basic sites,
are efficient in this type of reac-
tions. This work examines aldol
condensations between citral
and acetone or citral and
methyl ethyl ketone (MEK), as
well as the self-aldolization of
acetone over reconstructed hy-
drotalcite-like materials. A sim-
plified scheme for these reac-
tions
is
represented
in
Scheme 1. Preliminary tests
with HT-as and HT-c in the
condensation reaction between
citral and acetone at 333 K
have shown virtually no activity
over these catalysts over
a
period of 24 h. These materials
mainly contain Lewis basic
sites, which are inactive in the
reaction. Rehydration of HT-c
leads to a certain degree of con-
version, whose absolute value strongly depends on the rehy-
dration procedure applied. This fact is in agreement with
the data reported in the literature where, for the self-con-
densation of acetone, the meixnerite obtained by rehydra-
tion of Mg–Al mixed oxides, is more active than its hydroxy-
carbonate precursors (hydrotalcites).[4] It seems that Brønst-
Scheme 1. Reactions in the aldol condensation between citral and acetone and aldolization of acetone.
ed basic OHÀ ions sites play a vital role in this type of reac-
tions.
Influence of the rehydration procedure: Figure 12 shows the
conversion versus reaction time for HT-rg and HT-rl cata-
lysts for the citral/acetone and citral/methyl ethyl ketone
Chem. Eur. J. 2005, 11, 728 – 739
www.chemeurj.org
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
735

F. Medina et al.
the samples has indicated the degree of rehydration, which,
in principle, should favour the catalyst activity by creation
of active OHÀ groups, is higher in HT-rg. Similar differences
in performance between HT-rg and HT-rl samples have
been observed in the aldol condensation between citral/
MEK and the self-condensation reaction of acetone. The
conversion of citral over HT-rl was 62% after a reaction
period of 3 h. That the citral/MEK reaction is slower than
the citral/acetone reaction is attributed to the fact that
MEK is larger than acetone as well as the different acidic
character of the a-hydrogen atom in these ketones. For the
self-condensation of acetone to DAA using HT-rl as catalyst,
the thermodynamic equilibrium (23%) was achieved in less
than 0.5 h. Table 4 shows the initial reaction rates for HT-rg,
Figure 12. Conversion versus time for the reaction of citral/acetone at
333 K, citral/MEK at 333 K, and self-aldolization of acetone at 273 K
over the rehydrated hydrotalcites.
Table 4. Initial reaction rates and conversion level after 1 h for citral/
ketone condensation over HT-rl and HT-rg at 333 K.
Sample
Ketone
Initial rate[a]
Conversion of
citral[wt%]
[mmolcitral gcatÀ1 hÀ1
]
(MEK) reactions and the self-condensation of acetone to di-
acetone alcohol (DAA). A molar ketone/citral ratio of 4.4:1
was used, and the main products for the aldol reaction be-
tween acetone and citral were a mixture of cis- and trans-
pseudoionone. No b-hydroxyketone was observed under
these reaction conditions. The aldol reaction between citral
and MEK provided four isomers arising from the double
attack of the two carbanions of MEK. The selectivity of the
desired products of these reactions was always higher than
95%.
HT-rg
HT-rl
HT-rl-100
HT-rl-700A
HT-rl-700C
HT-rl
acetone
acetone
acetone
acetone
acetone
MEK
0.41
398.7
300.8
437.5
461
1
97.7
83.2
99.2
99.9
49.1
152.3
[a] Determined at 5 min of reaction.
and different HT-rl catalysts, used for these aldol condensa-
tions at T = 333 K. The data in Table 4 show the beneficial
effect of mechanical stirring during the rehydration process.
The increase of the surface area of the samples also produ-
ces an increase in the catalytic activity. These results confirm
that HT-rl samples are more active than HT-rg, accounting
on one hand for its greater amount of accessible sites as pre-
viously shown. On the other hand, sites of higher strength
should be also present, though hardly proved by TPD in
CO2 experiments, because the demanding condensation re-
action between citral/MEK only occurs with the samples re-
hydrated in liquid phase.
As mentioned above, two types of solid can be obtained
according to the different rehydration procedures of the cal-
cined hydrotalcites: HT-rl (by immersion in decarbonated
liquid water at a stirring speed of 500 rpm for 1 h, and wash-
ing with ethanol) and HT-rg (by contacting with a flow of
argon saturated with water vapour). For the gas-phase rehy-
drated samples, the flow was maintained for 10, 15, and
48 h, to obtain solids with differing amounts of water. The
loss of water detected by TGA was 29, 39, and 41 wt%, re-
spectively. The best result was achieved over the HT-rg
sample containing approximately 39% water that gave a
conversion of 4.3% in 1 h and 8.7% after 2.5 h in the self-al-
dolization of acetone. These results are not in agreement
with those previously reported (23% of conversion at 273 K
in 1–3 h).[8] This could be explained by the lower surface
areas achieved under our rehydration conditions in the gas
phase. Furthermore, other previous results with gas-phase
rehydrated hydrotalcites in this kind of reactions showed
that the presence of alkali metal ions (sodium), retained
during the synthesis of the hydrotalcite, could improve the
catalytic activity.[41] For the citral/MEK reaction, practically
no activity was detected when the reaction was carried out
over the gas-phase rehydrated catalysts. For the citral/ace-
tone reaction, only 4.5, 5, and 3.5% of citral conversion was
observed, even after a reaction period of 23 h, using HT-rg
with 29, 39, and 41 wt% of water, respectively. In contrast,
over HT-rl, the conversion of citral was 81% in only 5 min.
This fact indicates a dramatic difference in the catalytic
properties of HT-rg and HT-rl, although characterization of
Influence of the reaction temperature: Previous results for
these types of reactions showed that citral was strongly ad-
sorbed on the catalyst surface when working at low temper-
atures, indicating that a reaction temperature of about
333 K is needed to avoid this undesirable adsorption leading
to decreased catalytic activity.[19]
In this work, the aldol condensation of citral and acetone
was tested at different temperatures (283, 303, and 333 K)
over HT-rl. As expected, Figure 13 shows that the reaction
took place quickly when the temperature was increased to
333 K (96% of conversion in 30 min), whereas 24 h were
necessary for a 91% conversion of citral at 283 K. No
change in the product selectivity was observed at the differ-
ent reaction temperatures studied. An apparent activation
energy of 25 kJmolÀ1 was derived from the slope of the Ar-
rhenius plot in Figure 14. This low value for the activation
energy could indeed indicate the strong adsorption of the re-
736
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 728 – 739

Structure–Activity Relationships of Reconstructed Mg–Al Hydrotalcites
FULL PAPER
ring speed during rehydration) were obtained owing to the
lower platelet size as well as the porosity within the parti-
cles, as supported by microscopy and N2 adsorption results.
The formation of smaller platelets increases the number of
OHÀ ions near the edges, which contributes more to the en-
hanced activity. When the rehydration procedure was car-
ried out in the gas phase, the size of the platelets was larger,
although the degree of rehydration was higher. Although
the CO2-TPD results do not explain the differences in activi-
ty because the strength of the OHÀ groups in HT-rg and
HT-rl is similar, not all of the total OHÀ groups are probed
by CO2 during the TPD experiments. These OHÀ groups in
the HT-rl samples are more accessible for CO2 molecules
and for reactants than those in the HT-rg sample.
Figure 13. Conversion of citral in the aldol condensation between citral/
acetone at different temperatures.
Conclusion
This work clearly demonstrates the inactivity of as-synthe-
sized Mg–Al hydrotalcite and the product derived from its
thermal decomposition in aldol condensation reactions, and
highlights the active nature of Brønsted basic sites in pro-
ducing superior catalytic performances. To this end, two re-
hydration procedures were applied to a Mg–Al mixed
oxide: in the gas phase (HT-rg) or the liquid phase (HT-rl).
TGA, XRD, 27Al NMR spectroscopy, electron microscopy,
and FT-IR analyses show that the rehydration of HT-rg is
more efficient than that of HT-rl, with the subsequent
higher degree of lamellar structure reconstruction in the
former sample. Because the rehydration process leads to a
meixnerite-like material that contains Brønsted basic sites
(OHÀ) in the interlayer spaces, HT-rg samples can be con-
sidered to have a higher activity than HT-rl samples in aldol
condensation reactions. Nevertheless, HT-rl samples show
higher activity than HT-rg samples, indicating that the rehy-
dration method is a crucial step in the activation of the cal-
cined hydrotalcite. Gas-phase rehydration produces a virtu-
ally complete reconstruction of the layered structure; how-
ever, the final material shows a low specific surface area.
This fact means an extreme reduction in the number of ex-
posed active sites owing to a limited accessibility. The specif-
ic surface area obtained for the HT-rl samples depends on
the rehydration procedure. By increasing the rehydration
time and the stirring speed, materials with surface areas ap-
proximately 13 and 26 times higher than those of the HT-rg
samples are obtained. This fact leads to an increased
number of exposed basic sites, and thus to an improved cat-
alytic activity. This is supported by CO2-TPD, which shows
that there are more basic sites accessible probed by CO2 in
HT-rl. We can conclude that the strong difference in surface
areas for the two different types of rehydrated samples, the
number of basic sites and their accessibility are the main
factors for enhancing the catalytic activity of these reactions.
Therefore, this provides an approach for the preparation of
active catalysts for aldol condensation reactions and, in gen-
eral, for enhancing the catalytic activity of hydrotalcites by
controlling the rehydration process.
Figure 14. Arrhenius plot of the aldol condensation between citral and
acetone over HT-rl.
agents on the catalyst surface. This fact is also confirmed by
the brown color acquired for the catalysts at the end of the
reaction, even at higher temperatures, instead of the white
color in the fresh catalyst.
Structure–activity relationships
Our results show that basic OHÀ ions are required in the re-
hydrated hydrotalcite to obtain high activities in aldol con-
densation reactions; however, these active OHÀ are located
near the edges of the platelets. This finding supports the
suggestion of de Jong and co-workers[22] who stated that
only 5% of the hydroxy ions of the rehydrated samples are
active enough to carry out the reaction. By preparing and
testing hydrotalcites with a different platelet size using dif-
ferent ageing conditions,[40] we produced a disordered HT
structure that increased the basic strength. We then went a
step further and customized an enhanced rehydration proce-
dure, first by carrying out the ageing at room temperature
and after that, with a liquid-phase rehydration process, with
stirring. Consequently, higher surface areas between 200 and
400 m2 gÀ1 (depending on the variation of time and the stir-
Chem. Eur. J. 2005, 11, 728 – 739
www.chemeurj.org
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
737

F. Medina et al.
The narrow bore magnet (50 mm) was fitted with a high-speed magic
angle spinning (MAS) Doty probe. The samples were spun in 4 mm zir-
conia rotors with a spinning frequency of 8 kHz. A total of 4000 scans
were collected with a sweep width of 100 kHz and an acquisition time of
0.2 s. An acquisition delay of 1 s between successive accumulations was
selected to avoid saturation effects. The 27Al chemical shifts were refer-
Experimental Section
Materials: Mg–Al hydrotalcite (molar Mg/Al ratio = 3:1) was prepared
by a coprecipitation method at constant pH (10Æ0.2) of an aqueous so-
lution of Mg(NO3)2·6H2O (0.75m) and Al(NO3)3·9H2O (0.25m) and a
second solution of NaOH/Na2CO3 (2m). Both solutions were mixed drop-
wise under stirring at 298 K. After addition of the reactants, the slurry
was aged at 298 K for 15 h under vigorous stirring. The precipitate
formed was filtered and thoroughly washed with large amounts of deion-
enced to [Al(H2O)6]3+
.
The basic properties of the samples were determined by TPD of CO2 in
a Thermo FinniganTPDRO1100 equipped with a programmable temper-
ature furnace and a TCD detector. The gas outlet was coupled to a quad-
rupole mass spectrometer PfeifferGSD300. A method similar to that of
Padmasri et al.[42] was applied. Typically, ꢀ100 mg of solid was placed be-
tween quartz wool in a quartz reactor, pre-treated in argon at 373 K for
1 h, and then cooled to 353 K, prior to the adsorption of CO2 at this tem-
perature. After adsorption of CO2 (3 vol.% CO2 in He;
À
ized water to remove Na+ and NO3 ions. This step is essential because
Na impurities can bias the catalytic performance of the rehydrated sam-
ples by changing the basic properties of the surface.[41] The solid was then
dried at 373 K for 18 h to yield the as-synthesized hydrotalcite (HT-as).
The HT-as sample was thermally decomposed in air at 723 K for 15 h to
obtain the corresponding Mg–Al mixed oxide (HT-c). This material was
rehydrated in both gas and liquid phases. Gas-phase rehydration was car-
ried out by treating the calcined sample in an argon flow saturated with
water at room temperature for 15 h (40 mLminÀ1) to yield HT-rg (differ-
ent amounts of water in the sample were also achieved by varying the
contact time of the saturated flow (for 10 and 48 h)). Another series of
mixed oxides were rehydrated in decarbonated water (1 g of sample in
100 mL of water) for 1 h at room temperature and under mechanical stir-
ring (500 rpm). After the rehydration process, the sample was filtered,
washed with ethanol, and dried under argon to yield HT-rl. Some varia-
tions in the stirring speed (100, 300 and 700 rpm) and rehydration time
(10 min and 5 h) were also performed.
20 cm3 STPminÀ1
)
for 60 min, the catalyst was treated in He
(20 cm3 STPminÀ1) for 45 min at 373 K to remove the physically adsorbed
CO2. The CO2 uptake was measured by treating the sample from room
temperature up to 1173 K at a heating rate of 10 KminÀ1 and recording
the results with a TCD detector. Peak deconvolution was performed by
using the software of the equipment, and the number of Brønsted basic
sites was determined assuming that one molecule of CO2 adsorbs on each
basic site.
Catalyst testing: Aldol condensation reactions were performed in a 100-
mL round-bottom flask equipped with a magnetic stirrer and a condenser
system in argon under CO2-free conditions. For a typical reaction be-
tween citral (40.8 mmol) and ketone (180 mmol), a ketone/citral ratio of
4.4:1 was used. This solution was stirred at the desired temperature and
1 g of catalyst was added. The reaction was carried out in the range of
283–333 K. For the self-aldol condensation of acetone (0.25 mol), 1 g of
catalyst was also used and the reaction was conducted at 273 K. Samples
were taken at regular time intervals and were analysed off-line by gas
chromatography (GC) using a FID detector and an ULTRA2 column
(15 mꢅ0.32 mmꢅ0.25 mm). Tetradecane was used as the internal stan-
dard. Citral, pseudoionone, acetone (95%) and 2-butanone (MEK) were
purchased from Aldrich and were used without further purification.
Methods: The chemical composition of the samples was determined by
ICP-OES in a Perkin-Elmer Plasma400. The samples were diluted in
10% HNO3 before analysis. C,N,H analysis was performed in a Carlo
ErbaEA1108.
Powder X-ray diffraction patterns were collected in a SiemensD5000 dif-
fractometer with Bragg–Brentano geometry with nickel-filtered CuKa ra-
diation (l = 0.1541 nm). Data were collected in the 2q range of 5 to 708
with an angular step of 0.058 at 3 s per step, resulting in a scan rate of
18minÀ1
.
N2 adsorption and desorption isotherms at 77 K were measured on a
Quantachrome Autosorb-6B. Prior to analysis, the samples were degassed
in vacuum at 393 K for 16 h. The BET, t plot and BJH models (applied
to the adsorption branch of the isotherm using cylindrical pore geometry)
were used to derive information on the specific surface area, micro- and
mesoporosity and pore size distribution, respectively.
Acknowledgement
This work has been financially supported by the Ministerio de Ciencia y
Tecnologꢁa of Spain (REN2002–04464-CO2-01, PETRI95-0801.OP) and
Destilaciones Bordas S.A. Dr. P. J. Kooyman (TU Delft), K. Djanashvilli
(TU Delft), and T. Bach (Hydro) are acknowledged for performing the
TEM, 27Al MAS-NMR, and SEM studies, respectively.
Helium pycnometry measurements were performed at 293 K in a Quan-
tachrome pentapycnometer in order to determine the real density of the
samples.
Thermal analysis was performed in a Labsys/Setaram TG DTA/DSC
thermobalance, equipped with a programmable temperature furnace. The
sample (50 mg) was heated from room temperature up to 1173 K in
[1] F. Cavani, F. Trifiro, A. Vaccari, Catal. Today 1991, 11, 173.
[2] V. R. L. Constantino, T. J. Pinnavaia, Inorg. Chem. 1995, 34, 883.
[3] J. I. Di Cosimo, V.K. Dꢁez, M. Xu, E. Iglesia, C. R. Apesteguꢁa, J.
Catal. 1998, 178, 499.
[4] D. Tichit, B. Coq, CATTECH 2003, 7, 206.
[5] Y. Ono, J. Catal. 2003, 216, 406.
[6] W. T. Reichle (Union Carbide Corporation), US 4476324, 1984.
[7] G. Zhang, H. Hattori, K. Tanabe, Appl. Catal. 1988, 36, 189.
[8] D. Tichit, M. N. Bennani, F. Figueras, R. Tessier, J. Kervennal, Appl.
Clay Sci. 1998, 13, 401.
[9] D. Tichit, B. Coq, S. Cerneaux, R. Durand, Catal. Today 2002, 75,
197.
[10] M. J. Climent, A. Corma, V. Fornꢂs, R. Guil-Lopez, S. Iborra, Adv.
Synth. Catal. 2002, 344, 1090.
[11] K.K. Rao, M. Gravelle, J. Sꢆnchez-Valente, F. Figueras, J. Catal.
1998, 173, 115.
argon (80 cm3 STPminÀ1) at 5 KminÀ1
.
Infrared spectra were recorded on a Bruker-Equinox-55 FTIR spectrom-
eter. The spectra were acquired by accumulating 64 scans at 4 cmÀ1 reso-
lution in the range of 400–4000 cmÀ1. Samples were prepared by mixing
the powdered solids with pressed KBr disks (blank) in a ratio of 15:85.
The different samples were stored under argon to avoid absorption of
CO2. The calcined sample was analyzed in flowing Ar at 723 K in an in-
situ cell with the diffuse reflectance (DRIFT) mode.
Scanning electron microscopy images were recorded at 5 kV in a JEOL
JSM-6700F field-emission microscope. Samples were coated with palladi-
um to create contrast.
Transmission electron microscopy was carried out on a PhilipsCM30UT
electron microscope with a field emission gun as the source of electrons
operated at 300 kV. Samples were mounted on Quantifoil carbon poly-
mer supported on a copper grid by placing a few droplets of a suspension
of the ground sample in ethanol on the grid, followed by drying at ambi-
ent conditions.
[12] M. L. Kantam, B. M. Choudary, C. V. Reddy, K.K. Rao, F. Figueras,
Chem. Commun. 1998, 9, 1033.
[13] M. J. Climent, A. Corma, S. Iborra, J. Primo, J. Catal. 1995, 151, 60.
[14] A. Guida, M. H. Lhouty, D. Tichit, F. Figueras, P. Geneste, Appl.
Catal. A 1997, 164, 251.
27Al MAS-NMR experiments were performed at 9.4 T on a Varian VXR-
400S spectrometer operating at 104.2 MHz with a pulse width of 1 ms.
738
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 728 – 739

Structure–Activity Relationships of Reconstructed Mg–Al Hydrotalcites
FULL PAPER
[15] B. M. Choudary, M. L. Kantam, C. R. V. Reddy, K. K. Rao, F. Figue-
ras, J. Mol. Catal. A. 1999, 146, 279.
[16] T. Hibino, A. Tsunashima, Chem. Mater. 1998, 10, 4055.
[17] N. S. Puttaswamy, P. Vishnu Kamath, J. Mater. Chem. 1997, 7, 1941.
[18] M. J. Climent, A. Corma, S. Iborra, A. Velty, Catal. Lett. 2002, 79,
157.
[19] M. J. Climent, A. Corma, S. Iborra, A. Velty, Green Chem. 2002, 4,
474.
[20] J. C. A. A. Roelofs, A. J. van Dillen, K. P. de Jong, Catal. Lett. 2001,
74, 91.
[29] J. T. Kloprogge, R. L. Frost, Phys. Chem. Chem. Phys. 1999, 1, 1641.
[30] F. Prinetto, G. Ghiotti, R. Durand, D. Tichit, J. Phys. Chem. B. 2000,
104, 11117.
[31] J. T. Kloprogge, R. L. Frost, Appl. Catal. A 1999, 184, 61.
[32] F. Millange, R. I. Walton, D. OꢃHare, J. Mater. Chem. 2000, 10, 1713.
[33] W. Yang, Y. Kim, P. K. T. Liu, M. Sahimi, T. Tsotsis, Chem. Eng. Sci.
2002, 57, 2945.
[34] N. N. Greenwood, A. E. Earnshaw, Chemistry of the elements, Perga-
mon Press, New York, 1984, p. 329.
[35] J. Shen, M. Tu, C. Hu, J. Solid State Chem. 1998, 137, 295.
[36] J. I. Di Cosimo, C. R. Apesteguꢁa, M. J. L. Ginꢂs, E. Iglesia, J. Catal.
2000, 190, 261.
[21] J. C. A. A. Roelofs, A. J. van Dillen, K. P. de Jong, Catal. Today
2000, 60, 297.
[22] J. C. A. A. Roelofs, D. J. Lensveld, A. J. van Dillen, K. P. de Jong, J.
Catal. 2001, 203, 184.
[37] M. A. Aramendꢁa, V. Borau, C. Jimꢂnez, J. M. Marinas, J. R. Ruiz,
F. J. Urbano, Appl. Catal. A 2003, 255, 301.
[23] F. Figueras, J. Lꢀpez, J. Sꢆnchez-Valente, T. T. H. Vu, J-M. Clacens,
J. Palomeque, J. Catal. 2002, 211, 144.
[38] V.K. Dꢁez, C. R. Apesteguꢁa, J. I. Di Cosimo, J. Catal. 2003, 215,
220.
[24] J. Sꢆnchez-Valente, F. Figueras, M. Gravelle, P. Khumbar, J. Lꢀpez,
J-P. Besse, J. Catal. 2000, 189, 370.
[39] A. Alejandre, F. Medina, X. Rodrꢁguez, P. Salagre, J. E. Sueiras, J.
Catal. 1999, 188, 311.
[25] J. Lꢀpez, R. Jacquot, F. Figueras, Stud. Surf. Sci. Catal. 2000, 130,
491.
[26] A. Morato, C. Alonso, F. Medina, Y. Cesteros, P. Salagre, J. E. Suei-
ras, D. Tichit, B. Coq, Appl. Catal. B 2001, 32, 167.
[27] F. Prinetto, G. Ghiotti, P. Graffin, D. Tichit, Microporous Mesopo-
rous Mater. 2000, 39, 229.
[40] W. T. Reichle, J. Catal. 1985, 94, 547.
[41] S. Abellꢀ, F. Medina, D. Tichit, J. Pꢂrez-Ramꢁrez, X. Rodrꢁguez,
J. E. Sueiras, P. Salagre, Y. Cesteros, Appl. Catal. A 2004, in press.
[42] A. H. Padmasri, A. Venugopal, V. Durga Kumari, K. S. Rama Rao,
P. Kanta Rao, J. Mol. Appl. Catal. A. 2002, 188, 255.
[28] J. Pꢂrez-Ramꢁrez, G. Mul, F. Kapteijn, J. A. Moulijn, J. Mater. Chem.
2001, 11, 821.
Received: April 27, 2004
Published online: December 6, 2004
Chem. Eur. J. 2005, 11, 728 – 739
www.chemeurj.org
ꢄ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
739