Structure and surface and catalytic properties of Mg-Al basic oxides

Article abstract of DOI:10.1006/jcat.1998.2161

Mg-Al mixed oxides with Mg/Al molar ratios of 0.5-9.0 were obtained by thermal decomposition of precipitated hydrotalcite precursors. The effect of composition on structure and surface and catalytic properties was studied by combining several characterization methods with ethanol conversion reactions. The nature, density, and strength of surface basic sites depended on the Al content. On pure MgO, strong basic sites consisted predominantly of O^2- anions. Calcined hydrotalcites contained surface sites of low (OH^- groups), medium (Mg-O pairs), and strong (O^2- anions) basicity. The relative abundance of low and medium strength basic sites increased with the Al content. The addition of small amounts of Al to MgO diminished drastically the density of surface basic sites because of a significant Al surface enrichment. Formation of surface amorphous AlO_y structures in samples with low Al content (Mg/Al > 5) partially covered the Mg-O pairs and decreased the concentration of surface O^2- anions. At higher Al contents (5 > Mg/Al > 1), the basic site density increased because the Al³⁺ cations within the MgO lattice created a defect in order to compensate the positive charge generated, and the adjacent oxygen anions became coordinatively unsaturated. In samples with Mg/Al < 1, segregation of bulk MgAl₂O₄ spinels occurred and caused the basic site density to diminish. The catalyst activity and selectivity of Mg-Al mixed oxides in ethanol conversion reactions depended on composition. The dehydrogenation of ethanol to acetaldehyde and the aldol condensation to n-butanol both involved the initial surface ethoxide formation on a Lewis acid-strong base pair. Pure MgO exhibited poor activity because the predominant presence of isolated O^2- basic centers hindered formation of the ethoxide intermediate by ethanol dissociative adsorption. The incorporation of small amounts of Al³⁺ cations to MgO drastically increased the acetaldehyde formation rate because of the generation of new surface Lewis acid-strong base pair sites. Acetaldehyde condensation toward n-butanol is a bimolecular reaction between adjacent adsorbed acetaldehyde species that requires not only acid-strong base pair sites but also a high density of basic sites; these pathways were favored on Mg-Al samples with higher Al content (5 > Mg/Al > 1). The dehydration of ethanol to ethylene, and the coupling and dehydration to diethyl ether increased with Al content, probably reflecting the density increase of both Al³⁺-O^2- pairs and low- and medium-strength basic sites. Pure Al₂O₃ displayed the highest dehydration activity.

Full text of DOI:10.1006/jcat.1998.2161

JOURNAL OF CATALYSIS 178, 499–510 (1998)
ARTICLE NO. CA982161
Structure and Surface and Catalytic Properties of Mg-Al Basic Oxides
,1
ꢀ
ꢀ
ꢀ
J. I. Di Cosimo, V. K. Dı´ez, M. Xu,† E. Iglesia,† and C. R. Apesteguı´a
ꢀ
Instituto de Investigaciones en Cata´lisis y Petroqu´ımica (INCAPE), Santiago del Estero 2654, (3000) Santa Fe, Argentina;
and †Department of Chemical Engineering, University of California at Berkeley, Berkeley, California 94720
Received December 11, 1997; revised May 15, 1998; accepted May 18, 1998
INTRODUCTION
Mg-Al mixed oxides with Mg/Al molar ratios of 0.5–9.0 were ob-
tained by thermal decomposition of precipitated hydrotalcite pre-
cursors. The effect of composition on structure and surface and
catalytic properties was studied by combining several characteriza-
tion methodswith ethanolconversion reactions. Thenature, density,
and strength of surface basic sites depended on the Al content. On
pure MgO, strong basic sites consisted predominantly of O2ꢁ anions.
Calcined hydrotalcites contained surface sites of low (OHꢁ groups),
medium (Mg-O pairs), and strong (O2ꢁ anions) basicity. The relative
abundance of low and medium strength basic sites increased with
the Al content. The addition of small amounts of Al to MgO dimin-
ished drastically the density of surface basic sites because of a signif-
icant Al surface enrichment. Formation of surface amorphous AlOy
structures in samples with low Al content (Mg/Al > 5) partially cov-
ered the Mg-O pairs and decreased the concentration of surface O2ꢁ
anions. At higher Al contents (5 > Mg/Al > 1), the basic site density
increased because the Al3+ cations within the MgO lattice created a
defect in order to compensate the positive charge generated, and the
adjacent oxygen anionsbecamecoordinativelyunsaturated. In sam-
ples with Mg/Al < 1, segregation of bulk MgAl2O4 spinels occurred
and caused the basic site density to diminish. The catalyst activity
and selectivity of Mg-Al mixed oxides in ethanol conversion reac-
tions depended on composition. The dehydrogenation of ethanol
to acetaldehyde and the aldol condensation to n-butanol both in-
volved the initial surface ethoxide formation on a Lewis acid–strong
Layered double hydroxides are anionic clays in which
divalent cations within brucite-like layers are replaced by
trivalent cations. The resulting positive charge is compen-
sated by hydrated anions located in the interlayer space be-
tween two brucite sheets. The general formula for these ma-
II
terials is [M
MIxII (OH)2]x+ (Ax/n
)
nꢁ ꢃ m H2O, where MII
1ꢁx
is a divalent cation, such as Mg2+, Mn2+, Fe2+, Co2+, Cu2+
Ni2+, Zn2+, or Ca2+; MIII is a trivalent cation, such as Al3+
,
,
Cr3+, Mn3+, Fe3+, Co3+, or La3+, and Anꢁ is the anion (1).
Hydrotalcites and hydrotalcite-like compounds belong to
this class of materials.
Hydrotalcite-derived mixed oxides are used as antacids,
ion exchangers, and absorbers, and catalysts and catalyst
supports, because of their high surface area, phase pu-
rity, basic surface properties, and structural stability (2).
Hydrotalcite-derived oxides catalyze higher alcohol syn-
thesis from CO/H2 (3), oxidation of mercaptans (4), water
gas shift reactions (5). Also, Mg-Al mixed oxides prepared
by oxidative decomposition of hydrotalcite precursors have
been used in base-catalyzed aldolcondensations(6, 7), alky-
lations (8), Knoevenagel condensations (9), double-bond
isomerization (7), and alcohol dehydrogenation and cou-
base pair. Pure MgO exhibited poor activity because the predomi- pling (10).
nant presence of isolated O2ꢁ basic centers hindered formation of
The structure and surface properties of Mg-Al hydro-
the ethoxide intermediate by ethanol dissociative adsorption. The
incorporation of small amounts of Al3+ cations to MgO drastically
increased the acetaldehyde formation rate because of the genera-
tion of new surface Lewis acid–strong base pair sites. Acetaldehyde
condensation toward n-butanol is a bimolecular reaction between
adjacent adsorbed acetaldehyde species that requires not only acid–
strong base pair sites but also a high density of basic sites; these
pathways were favored on Mg-Al samples with higher Al content
(5 > Mg/Al > 1). The dehydration of ethanol to ethylene, and the
coupling and dehydration to diethyl ether increased with Al con-
talcites and of the resulting mixed oxides depend strongly
on chemical composition and synthesis procedures (2). In
catalysis, surface area, intimate contact between two or
more oxide components, and the size of the resulting metal
oxide clusters strongly influence the rate and selectivity of
chemical reactions. Shaper et al. (11) studied the activity of
calcined Mg-Al hydrotalcites with low Al content (Mg : Al
atomic ratios between 5 and 15) in the double-bond isomer-
ization of 1-pentene. They found that the sample activity
tent, probably reflecting the density increase of both Al3+-O2ꢁ pairs depends strongly on chemical composition and calcination
and low- and medium-strength basic sites. Pure Al2O3 displayed the
temperature. Mg-Al hydrotalcites calcined at 773 K dis-
played strong basicity and they were more active than pure
MgO. Velu and Swamy (8) studied the alkylation of phenol
with methanolon Mg-Alcalcined hydrotalciteswith Mg : Al
ratios from 3 to 10. The Mg-Al 4 : 1 ex-hydrotalcite was the
c
highest dehydration activity.
ꢂ 1998 Academic Press
1 Corresponding author. E-mail: capesteg@fiqus.unl.edu.ar.
499
0021-9517/98 $25.00
c
Copyright ꢂ 1998 by Academic Press
All rights of reproduction in any form reserved.

500
DI COSIMO ET AL.
most active sample, exhibiting also the highest selectivity Methods for the Characterization of Structure
for 2,6-xylenol formation. The authors concluded that the
participation of both acidic and basic sites is required for
alkylation reactions and that surface acid-base properties
are determined by the Al content. The influence of chemi-
cal composition of Mg-Al calcined hydrotalcites on the de-
hydrogenation of isopropanol was investigated by Corma
et al. (12). They prepared samples with different Mg/Al ra-
tios and found a maximum in dehydrogenation activity for
samples with a Mg/Al ratio of 3. For reactions requiring
stronger basicity, maximum rates were observed at higher
Mg/Al ratios. This work showed that on calcined Mg/Al hy-
drotalcites the surface acid–base properties and, as a con-
sequence, the catalytic activity and selectivity depend on
chemical composition. Intermediate Mg/Al compositions
may show unique catalytic properties, which depend on the
reaction requirements for the density and strength of basic
sites.
In this study, we prepared calcined Mg-Al hydrotalcites
with Mg/Al molar ratios of 0.5–9.0. The effect of com-
position on structure and surface and catalytic properties
was studied by combining several characterization meth-
ods with chemical reactions using probe molecules. The
crystal structure and crystallite size were determined by
X-ray diffraction (XRD). Surface composition was mea-
sured by X-ray photoelectron spectroscopy (XPS). The
density and reactivity of basic sites were determined using
13CO2/12CO2 isotopic switch methods and catalytic reac-
tions of ethanol. The density, binding energy, and struc-
ture of CO2 chemisorbed on basic sites were examined
by temperature-programmed desorption and infrared spec-
troscopy of adsorbed CO2.
and Surface Properties
The solid structure and crystal size in HT and CHT sam-
ples were determined by powder X-ray diffraction methods
using a Shimadzu XD-D1 diffractometer and Ni-filtered
CuKꢀ radiation.
The density of basic sites available for reversible ad-
sorption-desorption of CO2 on CHT samples were mea-
sured at 473 and 573 K by isotopic 13CO2/12CO2 exchange
methods (13). In this method, a sample is exposed to a 0.1%
13CO2/0.1% Ar/He stream at 473 K or 573 K until 13CO2
concentrations in the effluent reach a constant value and
then the stream is rapidly replaced with one consisting of
0.1% 12CO2/He. The number and rate ofevolution ofpread-
sorbed 13CO2 is then measured by continuous monitoring
of the effluent stream by mass spectrometry as a function of
time elapsed since the isotopic switch. Ar is present in one
of the streams in order to correct the exchange dynamics
for gas holdup and hydrodynamic delays within the exper-
imental apparatus.
CO2 binding energies were determined by temperature-
programmed desorption (TPD) of CO2 preadsorbed at
room temperature. Samples (50 mg) were treated in He at
723 K for 0.3 h and exposed to a 0.1% CO2/0.1% Ar/He
stream at room temperature until saturation coverages
were reached. Weakly adsorbed CO2 was removed by flush-
ing with He at room temperature for about 0.3 h. The tem-
perature was then increased at a linear rate of 0.5 K/s from
298 to 723 K and the rate of CO2 evolution was monitored
by mass spectrometry.
The structure of CO2 chemisorbed on hydrotalcite-
derived samples was determined by infrared spectroscopy
(IR). Data were obtained using a Shimadzu FTIR-8101M
spectrophotometer after CO2 adsorption at room temper-
ature and sequential evacuation at 298, 373, 473, 573, and
623 K. Spectra were taken at room temperature. An in-
verted T-shaped Pyrex cell containing the sample pellet was
used. The two ends of the short arm of the T were fitted with
CaF2 windows. The absorbance scales were normalized to
50-mg pellets.
The surface composition of CHT samples was measur-
ed by X-ray photoelectron spectroscopy (XPS) using a
Shimadzu ESCA-750 spectrometer and the integrated ar-
eas corresponding to the O 1s, Mg 2s, and Al 2s lines. XPS
binding energies were calibrated by reference to the posi-
tion of the C 1s signal at 284.6 eV. Bulk elemental analysis
of Mg, Al, and K was carried out by atomic absorption
spectroscopy (AAS). Total surface areas (Sg) were mea-
sured by N2 physisorption at 77 K using a Quantachrome
Nova-1000 sorptometer and BET analysis methods. Prior to
N2 physisorption, the hydrotalcite precursors and calcined
hydrotalcites were outgassed for 1 h at 393 K and 623 K,
respectively.
METHODS
Synthesis Procedures
Five Mg-Al hydrotalcite precursors with different Mg/Al
atomic ratios were prepared by co-precipitation of an aque-
ous solution of Mg(NO3)2 ꢃ 6 H2O and Al(NO3)3 ꢃ 9 H2O
containing the desired Mg/Al ratio at a constant pH of
10. A solution with a total [Mg + Al] cation concentration
of 1.5 M was contacted with a basic solution of K2CO3
and KOH by dropwise addition of both solutions into a
stirred beaker containing 350 cm3 of distilled deionized
water held at 333 K. The precipitates formed were aged
in their mother liquor for 2 h at 333 K and then filtered,
washed with boiling distilled water until K+ was no longer
detected in the filtrate, and dried at 348 K overnight.
These hydroxycarbonate precursors (HT samples) were
decomposed in N2 at 673 K overnight in order to obtain
the corresponding Mg-Al mixed oxides (CHT samples).
Pure alumina and MgO were prepared following the same
procedure.

PROPERTIES OF Mg-Al BASIC OXIDES
501
Ethanol Reaction Procedures
Catalytic testing was carried out at 573 K and 100 kPa in
a plug-flow fixed-bed reactor. Ethanol (Merck ACS, 99.8%
purity) was introduced as a 1 : 10 ethanol/N2 mixture. Be-
fore catalytic tests, the samples were pretreated in nitrogen
at 673K for 1h in order to remove water and carbon dioxide.
Catalyst loadings of 200 mg, sieved at 0.35–0.42 mm, and
contact times of 46 g catalyst h/mol of ethanol were used.
Diffusional limitations were ruled out by varying particle
sizes and contac times between 0.15–0.49 mm and 4–120 g
catalyst h/mol of ethanol, respectively. The reaction prod-
ucts were analysed by on-line gas chromatography using
a ATI Unicam 610 chromatograph equipped with flame
ionization detector and a 0.2% Carbowax 1500/80-100
Carbopack C column. Data were collected every 0.5 h
for 10 h. The predominant reaction products were acetal-
dehyde, n-butanol, diethyl ether, ethylene, and n-butyral-
dehyde. Traces of ethyl acetate, 2-propanol, 2-pentanone,
and acetone were also detected.
FIG. 1. HT samples: unit cell parameter, a, as a function of the chem-
ical composition. Brucite included as a reference.
tures with hexagonal 3R symmetry were calculated on all
HT samples from X-ray diffraction patterns and values of
a are shown in Fig. 1 as a function of the Al/(Al + Mg)
atomic ratio. The value of a decreases linearly with increas-
ing Al content from its initial value for brucite structures
˚
not containing Al (3.148 A). Sharp X-ray diffraction lines
were detected in sample HT3, which has the composition of
the stoichiometrichydrotalcite (x = 0.25). Allother samples
show broader XRD lines, corresponding to smaller crystal-
lites or less ordered structures, as shown by the crystallite
diameter values calculated from XRD line breadths and re-
ported in Table 1 for HT samples. The BET surface areas
of HT samples were between 65 and 85 m2/g and showed
no systematic dependence on composition.
RESULTS
Structural Characterization of Hydrotalcite
Precursors (HT)
The nominal chemical composition, BET surface area
and crystalline phases are shown in Table 1 for Mg(1ꢁx)Alx
samples before hydrotalcite decomposition (HT samples).
X-ray diffraction patterns show a single crystalline
phase with hydrotalcite structure for values of x = Al/
(Al + Mg) between 0.1 and 0.67. Outside this compo-
sition range, phases containing only Mg or Al were
also detected (e.g., Al(OH)3 (bayerite, ASTM 20-11),
Mg(OH)2 (brucite, ASTM 7-239)). Diffraction patterns
are consistent with the proposed hydrotalcite structure,
Structural Characterization of CHT Mixed Oxide Samples
The chemical composition, BET surface area, and crys-
talline phases for Mg(1ꢁx)Alx O samples formed by de-
composition of hydrotalcite precursors (CHT) are shown
in Table 2. Data on pure MgO and Al2O3 are also in-
cluded in Table 2. Elemental analysis of CHT samples gave
Al/(Al + Mg) atomic ratios very similar to those present
in the precursor solution, consistent with complete precip-
itation of Mg and Al salts during synthesis. The potassium
content in all CHT samples was below 0.1 wt% , which con-
firms that K+ ions were effectively removed by filtration
and washing of the precipitated precursors.
consisting of layered double hydroxides with brucite-like
2ꢁ
layers and [Mg1ꢁx Alx (OH)2]x+(CO3) ꢃ mH2O compo-
x/2
sition. The stoichiometric hydrotalcite structure, Mg6Al2
(OH)16CO3 ꢃ 4H2O (ASTM 14-191) is reached when x is
0.25. The unit cell constants (a and c) for hydrotalcite struc-
TABLE 1
Thermal decomposition of hydrotalcite precursors led
to the formation of high surface area Mg-Al oxides. Sur-
face areas were significantly higher than in HT samples
(Tables 1 and 2); it appears that the removal of CO2 during
decomposition leads to the formation of significant poros-
ity. This is consistent with the increase in surface area as
the Al (and the CO23ꢁ) content on HT samples increases.
All CHT samples showed diffraction patterns consistent
with the presence of a crystalline MgO periclase phase
(ASTM 4-0829). Sample CHT5 also has an additional
MgAl2O4 spinel phase (ASTM 21-1152). Crystalline AlOy
phases were not detected in any of the CHT samples. The
HT Samples: Chemical Composition, BET Surface Areas,
and XRD Characterization
XRD analysis
Nominal molar
composition
Sg
Crystallite
size (A)
Sample x = Al/(Al + Mg) (m2/g)
Phase
˚
HT1
HT2
HT3
HT4
HT5
0.10
0.17
0.25
0.50
0.67
71
69
76
66
85
Hydrotalcite + brucite
Hydrotalcite
Hydrotalcite
Hydrotalcite
Hydrotalcite + bayerite
54
101
106
99
75

502
DI COSIMO ET AL.
TABLE 2
CHT Samples: Chemical Composition, BET Surface Areas and XRD Characterization
Elemental analysisa
XRD analysis
Crystallite size
Al/(Al + Mg)
Mg
(wt% )
Al
(wt% )
K
(wt% )
Sg
Lattice
parameter (A)
Sample
(molar)
(m2/g)
Phase
(A)
˚
˚
MgO
0.00
0.11
0.18
0.24
0.47
0.65
1.00
—
45.3
39.7
34.4
22.2
15.0
-
—
6.2
9.7
12.1
21.5
31.1
-
0.003
0.008
0.020
0.020
0.080
0.085
0.060
125
114
184
238
231
298
388
MgO
MgO
MgO
MgO
MgO
75
124
91
28
21
4.213
4.208
4.211
4.216
4.209
-
CHT1
CHT2
CHT3
CHT4
CHT5
Al2O3
MgO + MgAl2O4
<20
-
Amorphous
-
a By AAS.
crystallinity of CHT samples and the apparent crystallite surface oxygen also depends on chemical composition. In
size decreased with increasing Al content (Table 2). Higher Fig. 2, the Os/Ob ratio is shown as a function of Al content.
surface areas and smaller crystallites are strongly favored Pure MgO exhibits the highest surface oxygen concentra-
by high Al contents in CHT samples.
tion, with a value of Os/Ob = 0.92. In CHT samples, the
Os/Ob ratio depends on chemical composition, with a min-
imum value reached at a low Al content and then a broad
maximum appearing at higher Al contents.
Surface Properties of CHT Mg(1ꢁx)AlxO Samples
As expected, Als and Mgs values follow qualitative trends
that parallel the bulk sample composition. As x increases,
Als increased and Mgs decreased (Table 3). As the Al con-
tent in CHT samples increased, however, Als/Alb ratios
decreased and Mgs/Mgb ratios increased. Samples CHT1
and CHT2 showed segregation of Al to the surface and
the Als/Alb ratio (2.30–2.68) is much higher than in pure
Al2O3 (1.13) (Table 3). In CHT samples with low Al con-
tent (x < 0.18), Mgs/Os ratios were very similar to those
in MgO, but these ratios decreased as the Al content in-
creased. Als/Os ratios were between 0.3 and 0.6 in all sam-
ples; these values are much lower than in pure Al2O3 (0.82).
13CO2/12CO2 isotopic exchange measurements at 473 K
and 573 K give the density of basic sites available for
The surface composition of MgO, Al2O3, and CHT
Mg(1ꢁx)Alx O samples was measured by X-ray photoelec-
tron spectroscopy. The surface atomic fraction of oxygen,
aluminum, and magnesium are designated as Os, Als, and
Mgs, respectively. The corresponding bulk molar fractions,
Ob, Alb, and Mgb, were obtained from bulk elemental anal-
ysis (Table 2). XPS results are shown in Table 3. The O 1s
binding energy (Table 3, column 2) changes with chemi-
cal composition, suggesting that the chemical form of the
surface oxygen species depends on chemical composition.
The measured O 1s binding energy increases as the Al con-
tent increases, from values typical of surface O2ꢁ (14, 15)
on pure MgO (529.5 eV), to that characteristic of surface
hydroxide (16) on pure Al2O3 (531.8 eV). The amount of
TABLE 3
Characterization of CHT Samples: XPS and TPD Results
Surface analysis by XPS
TPD of adsorbed CO2
Desorption peaksb (area % )
O 1s
B. E.a
(eV)
Total
evolved CO2
(ꢁmol/m2)
Surface composition (atomic ratio)
Sample
Mgs/Mgb
Als/Alb
Als/Os
Mgs/Os
l.t.p.
m.t.p.
h.t.p.
MgO
529.5
530.5
530.1
530.6
530.8
531.2
531.8
1.08
1.35
1.42
1.47
1.55
2.50
-
—
—
1.20
1.26
1.16
0.90
0.50
0.70
-
4.9
20.0
16.7
14.6
21.9
23.2
48.6
35.9
37.3
37.6
39.6
44.5
44.2
51.4
59.3
42.7
45.7
45.8
33.6
32.6
0.0
1.63
1.17
0.46
0.59
0.83
0.73
0.34
CHT1
CHT2
CHT3
CHT4
CHT5
Al2O3
2.68
2.30
1.74
1.38
1.20
1.13
0.31
0.41
0.33
0.40
0.63
0.82
a Binding energy.
b L.t.p: low-temperature peak; m.t.p.: middle-temperature peak; h.t.p.: high-temperature peak.

PROPERTIES OF Mg-Al BASIC OXIDES
503
FIG. 2. Surface/bulk oxygen ratio in MgO and CHT samples as a func-
tion of the chemical composition.
FIG. 4. Isotopic exchange experiments at 473 K and 573 K; density
of basic sites on MgO and CHT samples as a function of the chemical
composition.
reversible CO2 adsorption at these temperatures. Figure 3
shows isotopic exchange rates as a function of time on pure
MgO and CHT samples at 573 K and the corresponding
relaxation curve for the inert Ar tracer. The areas under
the relaxation curves in Fig. 3, after corrections for gas
holdup, physical hydrodynamic delays, and CO2 mass spec-
trometric response factors, represent the exchange capacity,
i.e. the density of basic sites that reversibly chemisorb CO2
at 573 K. Basic site densities measured on pure MgO and
CHT samples at 473 K and 573 K are shown in Fig. 4 as a
function of chemical composition. Basic site densities mea-
sured at 473 K are slightly higher than those obtained at
573 K, because weaker binding sites are unable to main-
tain measurable CO2 coverages at the higher temperature.
Basic site densities were higher on MgO than on CHT sam-
ples. At 573 K, the chemisorption capacity on MgO was
0.40 ꢁmol/m2, while it was only 0.17 ꢁmol/m2 on sample
CHT1, which contains 6.2 wt% Al. Increasing the Al con-
tent led to a second maximum in site densityat 0.33ꢁmol/m2
for sample CHT4 (x = 0.47).
A qualitative measure of the basic site strength and the
structure of chemisorbed CO2 were obtained by TPD and
infrared measurements of preadsorbed CO2. The rate of
CO2 evolution as a function of sample temperature is shown
in Fig. 5 for MgO, Al2O3, and CHT samples. The total
FIG. 3. Isotopic exchange experiments at 573 K; relaxation of 13CO2
as a function of time on MgO and CHT samples. Relaxation of Ar included
as a reference.
FIG. 5. TPD profiles of CO2 on MgO, Al2O3, and CHT samples.

504
DI COSIMO ET AL.
evolved CO2 was obtained by integration of TPD curves.
The resulting values are reported in the last column in
Table 3and showthat the amount ofevolved CO2 on MgO is
clearly higher than on CHT samples. The CO2 TPD profiles
on pure MgO and CHT samples are qualitatively similar
and can be deconvoluted in three desorption bands, reach-
ing maximum desorption rates at about 373, 443, and 543 K.
We calculated by integration the relative contribution of
each individual desorption peak. Results are presented in
Table 3. The high-temperature peak is predominant on pure
MgO, representing 59.3% of the total evolved CO2, but the
relative contribution of the low- and middle-temperature
peaks increase with increasing the Al content in CHT sam-
ples. Thus, on sample CHT5 the high-temperature peak ac-
counts for only 32.6% of the total desorbed CO2.
Figure 6 shows the infrared spectra obtained on sample
CHT4 after CO2 adsorption at room temperature and se-
quential evacuation at 373, 473, and 573 K. Similar spectra
were observed on all CHT samples. The spectra of MgO and
Al2O3 samples are also included as reference. Three species
of adsorbed CO2, which are shown in Scheme 1, were de-
tected (17, 18), apparently reflecting three different types
of surface basic sites. Unidentate and bidentate carbonate
formation requires surface basic oxygen atoms. Unidentate
carbonates exhibit a symmetric O-C-O stretching at 1360–
1400 cmꢁ1 and an asymmetric O-C-O stretching at 1510–
SCHEME 1. IR bands of adsorbed CO2 surface species [after (17, 18)].
1560 cmꢁ1. Bidentate carbonates show a symmetric O-C-O
stretching at 1320–1340 cmꢁ1 and an asymmetric O-C-O
stretching at 1610–1630 cmꢁ1. Bicarbonate species forma-
tion involves surface hydroxyl groups. Bicarbonates show
a C-OH bending mode at 1220 cmꢁ1 as well as symmetric
and asymmetric O-C-O stretching modes at 1480 cmꢁ1 and
1650 cmꢁ1, respectively.
Overlapping broad infrared bands were observed on
pure MgO and CHT samples in the 1700–1500 cmꢁ1 re-
gion following the CO2 adsorption at room temperature.
Similar qualitative spectra were obtained for all the sam-
ples after evacuation at each temperature, as is illustrated
in Fig. 6 for pure MgO and sample CHT4. The bicarbonate
species disappeared after evacuation at 373 K, but uniden-
tate and bidentate carbonates remained on the surface even
after evacuation at 573 K; after evacuation at 623 K, only
unidentate carbonates species were detected. Figure 6 also
shows that bicarbonate was clearly the predominant species
formed on Al2O3 upon CO2 adsorption at room tempera-
ture. After sequential evacuation at 373, 473, and 573 K, the
bicarbonates were removed completely from the alumina
surface.
Catalytic Conversion of Ethanol on Mg-Al Mixed Oxides
The rate and selectivity for ethanol conversion reactions
were determined on CHT samples at 573 K. Ethanol is
converted on Mg-Al calcined hydrotalcites via three main
groups of reactions: dehydrogenation, coupling, and dehy-
dration. In order to identify primary and secondary prod-
ucts, the effect of contact time on the product distribution
was investigated using sample CHT4. It was found that
dehydrogenation of ethanol forms acetaldehyde as a pri-
mary product. The consecutive aldol condensation of ac-
etaldehyde yields n-butanol. Dehydration reactions lead to
FIG. 6. Infrared spectra of CO2 adsorbed on MgO, Al2O3, and sample
CHT4 upon increasing evacuation temperatures: (a) room temperature;
(b) 373 K; (c) 473 K; (d) 573 K.

PROPERTIES OF Mg-Al BASIC OXIDES
505
like compounds consisting of positively charged brucite-
like layers [M6IIM2III(OH)16]2+ alternating with disordered
negatively charged interlayers (19). The interlayer spacing
is occupied by charge-compensating CO23ꢁ anions and by
weakly bonded water molecules. Several studies have dis-
cussed the successful synthesis of Mg(1ꢁx)-Alx hydrotalcites
with low Al content, but few reports have addressed the
synthesis of HT samples with high Al content (x > 0.50).
The synthesis conditions used in this work led to the for-
mation of Mg-Al layered double hydroxides with uniform
hydrotalcite structure and a broad range of Al content (x =
0.10–0.67).
The lattice parameter a of HT samples decreased mono-
tonically, following Vegard’s law dependence as the Al con-
tent increases(Fig. 1), suggestingthat the brucite-like sheets
can accommodate Al3+ cations within a wide compositional
range. This contraction of the HT unit cell was previously
reported for similar Mg-Al hydrotalcites(20–22). It appears
to be caused by the formation of a solid solution, which is
structurally stable in brucite layers because Al3+ cations
replace larger Mg2+ cations in the brucite structure. Thus,
Al3+ can randomly and uniformly replace Mg2+ in octahe-
dral sites within brucite layers without disrupting the layer
structure.
SCHEME 2. Reaction network of ethanol conversion.
diethyl ether and ethylene. A simplified reaction network
of the ethanol conversion is shown in Scheme 2.
The catalytic results obtained on MgO, Al2O3, and CHT
samples are shown in Table 4. Samples deactivated dur-
ing the experiments and lost about 50% of activity after
10 h on stream. Rates in Table 4 are initial activities ob-
tained by extrapolating conversion versus time curves to
zero. Pure MgO shows very low reaction rates. On CHT
samples, ethanol conversion rates depend on the Al con-
tent. Acetaldehyde formation is favored at low Al content
while condensation to form n-butanol reaches a maximum
rate on sample CHT4. The formation rate of ethylene and
diethyl ether on CHT samples increase with the Al con-
tent. The alumina sample is especially active for producing
ethylene. Table 4 shows that the ethylene formation rate
on Al2O3 was almost two orders of magnitude higher than
on sample CHT5, the most active CHT sample. Dehydra-
tion reactions appear to occur on Lewis or Brønsted acid
sites present on Al2O3 surfaces that become abundant in
Al-rich CHT samples.
Influence of the Chemical Composition on the Structure
of the Mixed Oxides
The thermal decomposition of Mg(1ꢁx)Alx hydrotalcites
occurs in two steps (23). Below 453 K, weakly held water
of hydration is desorbed. In the 453–673 K temperature
range, water and CO2 are formed by dehydroxylation of
vicinal OH groups within brucite layers and by decarboxy-
lation of interlayer CO23ꢁ, respectively. The evolution of
these gaseous products creates a significant porous struc-
ture and leads to the marked increase in surface area ob-
served upon decomposition of HT samples. Al-rich hydro-
talcite precipitates contain larger amounts of CO2, because
higher densities of CO23ꢁ anions are required to compensate
the net positive charge in the brucite layer. The resulting
larger amounts of CO2 evolved during decomposition cause
the higher surface areas observed in Al-rich CHT samples
(Table 2). Diffraction patterns of CHT samples showed no
residual hydrotalcite or hydroxide phase, confirming that
heating at 673 K leads to complete decomposition of HT
samples. A poorly crystalline MgO phase was detected in
all CHT samples, but no crystalline AlO y phase was ob-
served. The latter shows that Al3+ cations remain closely
associated with the MgO structure after decomposition of
HT samples within the entire compositional range of our
study.
DISCUSSION
Structure and Homogeneity of Hydrotalcite Precursors
The structure of precipitated samples is consistent with
the predominant presence of rhombohedral hydrotalcite-
TABLE 4
Catalytic Results on MgO, Al2O3, and CHT Samples
Formation rate (ꢁmol/h m2)
Ethanol
conversion
rate
Diethyl
ether
Sample (ꢁmol/h m2) Acetaldehyde n-Butanol
Ethylene
MgO
0.89
8.35
4.24
4.55
9.17
9.84
48.15
0.28
4.64
2.55
1.10
0.29
0.26
0.08
0.21
0.96
0.49
1.00
1.82
0.52
0.00
0.03
0.26
0.08
0.19
2.02
3.89
4.84
0.04
0.06
0.07
0.08
0.57
0.60
28.57
CHT1
CHT2
CHT3
CHT4
CHT5
Al2O3
Several studies based on NMR analysis have proposed
that Al coordination partially changes from octahedral to
tetrahedral upon decomposition of hydrotalcites (11, 12,
23–25). Corma et al. (12) detected, in calcined hydrotalcites,
Note. T = 573 K, P = 100 kPa, W/F0 = 46 g h/mol, N2/ethanol = 10 (mo-
lar).

506
DI COSIMO ET AL.
the presence of tetrahedrally coordinated Al3+ ions (AlTd), sites with widely different desorption rate constants. As ex-
which causes a 27Al NMR band different from that of AlTd pected, isotopic exchange methods cannot detect weakly
located in a pure Al2O3 structure. Consequently, they pro- adsorbed bicarbonate species, because OH surface sites
posed that part of the AlTd ions are in tetrahedral sites are largely uncovered at the temperatures of the exchange
within the MgO lattice. They found, however, that the experiments and of catalytic interest. Bidentate species
amount of AlTd increases with the Al content in the calcined account for the rapid exchange processes at short times
hydrotalcite. The presence of AlTd in the MgO structure and more strongly adsorbed unidentate species for the ex-
leads to a decrease in the lattice constant of MgO (26). The change process occurring at longer times.
diffraction patterns of our CHT samples, however, did not
Compositional Effects on Basic Site Density and Strength
show any measurable lattice contraction with the Al con-
tent (Table 2), probably because of the poor crystallinity of
The 13CO2/12CO2 isotopic exchange data in Fig. 4 show
hydrotalcite precursors decomposed at low temperatures that basic site densities (per unit surface area) on CHT sam-
(673 K). At high Al contents (sample CHT5, x = 0.67) a plesdepend on Alcontent. The densityofbasicsiteson pure
separate MgAl2O4 phase begins to form; its structure con- MgO decreases when small amounts of Al are added. Basic
sists of a partially inverse spinel containing 44% of the Al3+ site densitieson CHT sampleswith xvaluesbetween 0.1and
ions within tetrahedral positions (27). Thus, it is possible 0.2 are less than half of the corresponding values measured
that some of the Al3+ cations are located in tetrahedral sites on pure MgO. The density of basic sites then increases as
within MgO structures even for samples with lower Al con- additional Al is added, reaches a maximum for CHT sam-
tent. This unstable structure may well generate small and ples with x values of about 0.5, but never recovers the site
distorted structures with partially inverse spinel ordering, density values characteristics of pure MgO. Qualitatively
which are difficult to detect by X-ray diffraction methods.
similar trends were obtained when site densities are calcu-
lated from the area under CO2 temperature-programmed
desorption profiles (Table 3), which include both strongly
Surface Properties and Reactivity of the Basic Sites
Temperature-programmed desorption profiles of chemi- held and weakly held CO2 .
sorbed CO2 suggest that several binding sites are available
The effect of the Al content on the basicity of calcined
on the surface of MgO and of CHT samples. TPD profiles Mg-Al hydrotalcites was studied by Fishel et al. (25) (for
(Fig. 5) consist of three overlapping CO2 desorption peaks, x = 0.17–0.33) and by Nakatsuka et al. (29) (x = 0.10–0.44).
reaching maximum desorption rates at about 373, 443, and Nakatsuka et al. (29) reported that the basic site density
543 K. The presence of three distinct adsorbed structures measured by titration with benzoic acid (strength H ꢄ 15)
ꢁ
is also apparent from the infrared spectra of CO2 adsorbed depends on the Al content, reaching a maximum for x val-
on CHT samples after partial desorption by evacuation at ues of about 0.28. Fishel et al. (25) measured the number
various temperatures (Fig. 6). In TPD experiments, the low- of basic sites by TPD of CO2 and observed that the basic
temperature desorption peak corresponds to bicarbonate site density of pure MgO decreases when small amounts of
species adsorbed on weakly basic OH groups; bicarbonate Al are added; then it increases for higher x values, reach-
infrared bands disappear after treatment at 373 K. Biden- ing a maximum at ca x = 0.25. Derouane et al. (24) studied
tate carbonates adsorb on Mg-O site pairs and remain ad- the physicochemical properties of Mg-Al calcined hydrotal-
sorbed until samples are treated above 573 K. Unidentate cites with x = 0.17–0.33 and found that both the crystallinity
carbonates form on strongly basic surface O2ꢁ anions and and the unit cell of pure MgO decrease with the Al content.
remain adsorbed even after evacuation at 623 K. Bidentate These results were attributed to the migration of Al3+ ions
and unidentate carbonates have been assigned to CO2 ad- into the MgO framework, which creates a cationic vacancy;
sorbed on sites with intermediate and high basic strength, the surface then acts as a sink for the vacancy and one Mg2+
respectively (28).
is removed from the surface leading to the formation of an
13CO2/12CO2 isotopic exchange reactions follow first- additional surface O2ꢁ ion. They postulated that oxygen-
order kinetics and relaxation curves follow straight lines terminated, negatively charged (111) planes in MgO would
in semi-logarithmic plots when adsorption–desorption pro- be then more suitable than magnesium-oxygen-terminated
cesses occur on a single type of surface site. The slope in (100) planes to accommodate additional oxygen atoms.
these plots reflects the rate constant for isotopic exchange.
Using Derouane’s model, Fishel et al. (25) explained
Large valuesofthisrate constant reflect short 13CO2 surface the CO2 chemisorption capacity decrease observed on
lifetimes corresponding to weaker binding of CO2 on basic ex-hydrotalcites of low Al content by postulating that
sites. The nonlinear behavior of the data in Fig. 3 confirms the oxygen-terminated (111) planes would be less likely
that the surface of MgO and CHT samples is nonuniform to chemisorb CO2. They assumed the need of exposed
and that it contains sites with varying exchange kinetics Mg2+-O2ꢁ pairs, which are not favored on (111) planes, in
and basic strength. The data in Fig. 3 can be described by order for the chemisorption of CO2 to take place. How-
first-order exchange processes occurring on two types of ever, our TPD and IR results show that at low x values,

PROPERTIES OF Mg-Al BASIC OXIDES
507
the CO2 chemisorption takes place not only on medium- to values corresponding to OHꢁ groups as the Al content
strength Mg2+-O2ꢁ pairs, but also on strongly basic sur- increases (Table 3). Also, the infrared spectra of adsorbed
face O2ꢁ ions. About 45% of the chemisorbed CO2 gave CO2 in Fig. 6 show that only weakly bonded bicarbonate
rise to unidentate carbonates formed on high-strength O2ꢁ species on basic OHꢁ groups are formed on pure Al2O3.
ions (Table 3). Furthermore, X-ray photoelectron results Therefore, we conclude that the formation of weakly basic
show that Mgs/Os ratios on CHT samples with low Al con- OH groups on CHT samples increases with the Al content
tent (1.16–1.26) were similar to that on pure MgO (1.20) and, as a result, the average surface basic strength dimin-
(Table 3). The nearly equimolar amounts of Mg and O on ishes.
the surface of these samples suggest the presence of abun-
dant exposed Mg2+-O2ꢁ pairs and excludes a preferential
Compositional Effects on Catalytic Activity
(111) surface orientation.
The product distribution in ethanol conversion reactions
Our X-ray photoelectron results show significant Al sur-
is influenced by the surface acid–base properties. We have
face enrichment in CHT samples with low Al content
shown that the CHT Mg1ꢁx Alx O samples present three
(Als/Alb = 2.30–2.68, Table 3), suggesting a more plausible
explanation for the decrease in basic site density observed
on these samples. The Al3+ migration to the surface in sam-
ples with low Al content was predicted by Rohrer et al. (30)
using Monte Carlo simulations of Mg(Al)O solid solutions.
These simulations showed that for a sample with x = 0.17
the effects of Schottky defects and Al3+ ions on the atomic
relaxations at the MgO surface cause substitutional Al3+
ions and cation vacancies to segregate to the surface re-
gion, in accordance with experimental observations. The
presence of more electronegative amorphous AlO y struc-
tures or isolated Al centers on MgO surfaces blocks CO2
chemisorption sites, which are five times more abundant on
pure MgO than on pure Al2O3 (Table 3, last column). At
higher Al contents, the Als/Alb ratio decreases, suggesting
that surface segregation of Al is less likely to occur, and
basic site densities then increase (Fig. 4) concurrently with
increases in surface oxygen concentrations (Fig. 2). While
no separation of phases takes place, the Al3+ cations located
in the MgO framework create a defect in order to compen-
sate the positive charge generated, and the adjacent oxygen
anions become coordinatively unsaturated; as a result, the
solid solution displays higher basicity. As the bulk spinel is
formed, the Al3+ cations leave the MgO lattice and phase
segregation occurs causing the density of basic sites to drop.
The increase in Al content influences both the basic site
density, as measured by 13CO2/12CO2 isotopic exchange,
and the surface oxygen concentration (Figs. 4 and 2), which
strongly suggests that surface oxygen (as O2ꢁ anions and
in Mg2+-O2ꢁ pairs) provides the binding sites for adsorbing
CO2 at the temperatures of the exchange measurements.
The effect of Al content on the distribution of basic site
strength in CHT samples was investigated by decomposing
the CO2 TPD curves of Fig. 5 into three desorption peaks
and measuring their respective areas. Results presented in
Table 3 show that pure MgO exhibits the highest relative
concentration of high-strength basic sites. Consequently,
the relative contribution of low- and medium-strength basic
sites on CHT samples is higher than on MgO. This is con-
sistent with the observed shift in the binding energy of the
O 1s XPS line from its value for O2ꢁ species on pure MgO
different types of basic sites: isolated O2ꢁ ions, Mg2+-O2ꢁ
pairs, and OH groups; their relative concentrations depend
on the Al content. Also, we found that the contribution
of Al3+-O2ꢁ pairs to the surface basicity is not important.
In contrast, the surface Lewis acidic sites are provided by
both Mg2+ and Al3+ cations. The Lewis acidity supplied by
the metal cations is stronger than that of the Brønsted OH
groups (38).
Table 4 shows that acetaldehyde was the predominant
product on CHT samples with low Al content. Dehydro-
genation of a primary alcohol to the corresponding alde-
hyde is usually faster than alcohol dehydration on basic
oxides and this reaction appears to require large polariz-
able cations, such as the alkaline earths ions (32–34). Dehy-
drogenation involves the initial adsorption of ethanol on an
acid–strongbase pair site, which breaksthe O-H bond form-
ing a surface ethoxy intermediate. Then, the ꢀ-hydrogen in
the ethoxy group is abstracted by another strong basic site.
The acetaldehyde formation mechanism is represented in
Scheme 3, where A and B are Lewis acid sites and Brønsted
basic sites, respectively:
SCHEME 3
Pure MgO catalyzes the dehydrogenation of ethanol to ac-
etaldehyde at very low rates. This is consistent with the
predominant presence of isolated surface O2ꢁ ions on MgO
surfaces. Isolated O2ꢁ basiccenterswould be unable to form
ethoxide intermediates, which requires acid–strong base
pair sites. The rate of formation of acetaldehyde on CHT
samples with low Al content (x < 0.2) was about one order
of magnitude higher than on pure MgO. It appears that the
addition of small amounts of Al drastically increases the
surface density of active acid–strong base pair sites. Our
XPS and CO2 TPD results showed significant Al surface
enrichment in CHT samples with low Al content, which

508
DI COSIMO ET AL.
counterbalanced the contribution of high-strength basic
Dehydration of alcohols occurs on oxides containing
sites. The enhanced acetaldehyde production observed on small-highly charged cations (32) which exhibit acidic prop-
samples CHT1 and CHT2 may be interpreted then by as- erties. Olefin formation is the predominant reaction on
suming that these samples contain a much larger number of acidic oxides, but ether formation can also take place
properly positioned Al3+ Lewis acid sites and Mg2+-O2ꢁ ba- (35–37). Ethanol dehydration to ethylene on Mg-Al mixed
sic pairs, which catalyze the rate-limiting step in the ethanol oxides proceeds via two mechanisms. The E1cB mechanism
dehydrogenation mechanism.
requires both strongly basic sites and weak Lewis acid sites
Condensation of adsorbed acetaldehyde species is faster (34, 38). Initially, the adsorption of the ethanol molecule
than dehydrogenation (32), so that formation of coupling and the O-H bond rupture on the acid–strong base pair site
products on basic oxides is expected to be rather important. gives rise to a surface ethoxy group. Then, the E1cB mech-
However, whereas the acetaldehyde formation on CHT anism leads to the abstraction of the most acidic proton in
samples decreases monotonically with Al content, its con- the ethoxy intermediate on the basic site, with formation of
secutive coupling to n-butanol reaches a maximum for sam- a carbanion by Cꢂ-Hꢂ bond rupture (39). Formation of the
ple CHT4. Similarly to ethanol dehydrogenation, the aldol olefin takes place in the final step:
condensation of acetaldehyde to n-butanol on basic oxides
involves formation of a carbanion intermediate on an acid-
strong base pair site:
SCHEME 5
The E2 elimination is a concerted single-step mechanism
which involves Lewis acidic and basic sites of balanced
strength so that no formation of ionic intermediate takes
place. The acidic site responsible for the OH abstraction,
should be stronger than the one participating in the E1cB
mechanism:
SCHEME 4
Acetaldehyde condensation, however, is a bimolecular
reaction between adjacent adsorbed acetaldehyde species,
and the formation of n-butanol requires not only acid–
strong base pair sites but also a high density of basic sites.
SCHEME 6
We found that sample CHT4 contains the highest density
Table 4 shows that ethylene is readily formed on alu-
of basic sites among the CHT samples (Fig. 4) and this may mina, which is a typical E2 catalyst (33). Alumina was two
explain the high rate of n-butanol synthesis on this sample. or three orders of magnitude more active for ethylene pro-
For similar reasons, the aldol condensation to n-butanol is duction than CHT samples. On CHT samples, ethylene for-
unlikely to take place on pure Al2O3 surface.
mation rates increased with increasing Al content. Sam-
In addition to n-butanol, at least two other prod- ples CHT5 and CHT4 contain a relatively high density of
ucts are predicted to form from acetaldehyde condensa- Al3+-O2ꢁ pairs and a lower density of strong basic sites
tion. The partially hydrogenated intermediate compound, (Table 3). Ethylene formation on these samples would oc-
n-butyraldehyde, was detected on CHT samples of high cur then through an E2 mechanism. In contrast, on CHT
Mg contents in concentrations that represent less than samples with low Al content, the density of strong basic
3% of the overall reaction rate. On the other hand, the sites is much higher than that of the acid sites generated by
ꢀ,ꢂ-unsaturated aldol-type species (2-butenal) depicted in the aluminum surface enrichment and the ethylene forma-
Scheme 4, which is expected to interact strongly with the tion rates decrease. On Mg-rich samples (samples CHT1
basic surface, was never observed. It appears that the migra- and CHT2), dehydration of ethanol to ethylene probably
tion and recombination of surface hydrogen atoms result- occurs via an E1cB mechanism, which requires strong basic
ing from formation of adsorbed acetaldehydic species take sites. The olefin formation on pure MgO has already been
place faster than the ꢀ,ꢂ-unsaturated intermediate des- ascribed to an E1cB mechanism (38).
orption, thereby favoring the complete hydrogenation of
2-butenal to n-butanol.
Formation of diethyl ether from ethanol on alumina is
a second-order reaction that involves two adjacent alcohol

PROPERTIES OF Mg-Al BASIC OXIDES
509
molecules adsorbed on two different types of site (35, 40).
CONCLUSIONS
One alcohol molecule (I) adsorbs on a Lewis acid site
through the oxygen of the OH group, and the Cꢀ atom
becomes slightly positive. The other ethanol molecule (II)
adsorbs on a basic site via a H-bond that increases the nu-
cleophilicity of the oxygen in the OH group. The adsorbed
molecule II is an incipient alkoxide ion (35) which then
performs a nucleophilic displacement on the Cꢀ atom of
the molecule I to form diethyl ether:
The nature, density, and strength of surface basic sites on
ex-hydrotalcite Mg-Al mixed oxides depend on the sample
composition. In samples of low Al content (x < 0.2), the ba-
sic site density is clearly lower as compared to pure MgO
because the formation of a surface AlO y phase partially
covers the Mg-O pairs and decreases the concentration of
surface O2ꢁ anions. At higher Al contents (0.2 < x < 0.5),
a highly interacting Mg-Al oxide phase is formed by incor-
poration of Al3+ ions into the MgO matrix and causes the
number of surface defects to increase and the partial recov-
ery of the basic site density. In samples with x > 0.5, demix-
ing of the Mg-Al phase occurs leading to bulk MgAl2O4
spinels. As a result, the basic site density diminishes.
On pure MgO, strong basic sites consist predominantly
of O2ꢁ anions. Calcined hydrotalcites contain surface sites
of low (OHꢁ groups), medium (Mg-O pairs), and strong
(O2ꢁ anions) basicity. Their relative abundance depends on
the chemical composition; the surface concentration of low
and medium strength basic sites increases with increasing
Al content in the sample.
The catalyst activity and selectivity of CHT samples in
ethanol reactions also depend on composition. Reactions
requiring strong basic sites, such as dehydrogenation, con-
densation, and E1cB dehydration, involve the initial sur-
face ethoxide formation on a Lewis acid–strong base pair
site. Pure MgO exhibits poor activity because the pre-
dominant presence of isolated O2ꢁ basic centers hinders
ethanol dissociative adsorption steps that form the ethoxide
intermediate. MgO forms predominantly acetaldehyde and
n-butanol via reactions favored on strongly basic materials.
The incorporation of small amounts of Al to MgO
increases ethanol reaction rates by about one order of
magnitude. It appears that Al3+ cations drastically increase
the surface density of active Lewis acid–strong base pair
sites. Calcined Mg-Al hydrotalcites with low Al content se-
lectively dehydrogenate ethanol to acetaldehyde. The con-
secutive coupling of acetaldehyde to n-butanol demands,
in addition to the Lewis acid–strong base center, a higher
density of basic sites; therefore these pathways are favored
on CHT samples with higher Al content (0.2 < x < 0.5).
Ethanol dehydration rates increase abruptly on samples
containing a high density of both Al3+-O2ꢁ pairs and low-
and medium-strength basic sites. Thus, the formation of
ethylene and diethyl ether from ethanol occurs mainly on
CHT samples with high Al content. Pure alumina displays
the highest total activity, however, dehydrogenation and
condensation products do not form because of the absence
of acid–strong base pair sites.
SCHEME 7
It has been debated in the literature whether formation
of olefin and ether from primary alcohols takes place on
acid sites of similar or different acid strengths. Ross et al.
(41) studied the effect of Na+ doping on the dehydration of
ethanol over alumina. They showed that the increase in the
Na content leads to a shift in the selectivity from ethylene
to diethyl ether. Consistently, Jain et al. (35) reported that
the alcohol dehydration to olefins has a higher activation
energy than the competitive dehydration to ethers and will
be favored then by the presence of stronger surface acidic
sites or the use of higher reaction temperatures.
Table 4 shows that the formation rates of both diethyl
ether and ethylene increase with increasing Al content
in calcined hydrotalcites. However, results of Table 4
suggest different acid–base site requirements for ethylene
and diethyl ether formation. Low selectivities to ethylene
were found in the entire compositional range of calcined
hydrotalcites; diethyl ether was one of the main products
only on CHT samples of high Al/(Al + Mg) ratios. This sug-
gests that at high Al contents, the combination of surface
properties, as discussed above, generates acid–base pair
sites acidic enough to catalyze the synthesis of diethyl ether
rather than the olefin formation through an E2 mechanism.
It has to be noted here that several authors have found that
the adsorption of alcohols over basic oxides takes place
also on weaker Brønsted acidic sites provided by surface
OH groups (31, 34). Noller et al. (31) reported that the
ethanol adsorption mode shown for the ethanol molecule I
in Scheme 7, takes place on surface hydroxyl groups rather
than on Lewis sites such as Mg2+ or Al3+ ions. From our
results, the participation of surface OH groups in the syn-
thesis of diethyl ether cannot be ruled out because diethyl
ether formation rates on CHT samples parallel the extent
of surface hydroxylation determined by XPS (Table 3).
ACKNOWLEDGMENTS
The financial support for this work by CONICET (Argentina), the
Universidad Nacional del Litoral (Santa Fe, Argentina), the National

510
DI COSIMO ET AL.
Science Foundation (U.S.A.), and the U.S. Department of Energy (U.S.A.)
is gratefully acknowledged.
18. Philipp, R., and Fujimoto, K., J. Phys. Chem. 96, 9035 (1992).
19. Taylor, H. F. W., Min. Mag. 39(304), 377 (1973).
20. Pausch, I., Lohse, H. H., Schurmannand, K., and Allmann, R., Clays
Clay Min. 34, 507 (1986).
REFERENCES
21. Miyata, S., Clays Clay Min. 28, 50 (1980).
22. Kooli, F., Kosuge, K., and Tsunashima, A., J. Mater. Sci. 30, 4591
(1995).
23. Reichle, W. T., Kang, S. Y., and Everahardt, D. S., J. Catal. 101, 352
(1986).
24. Derouane, E. G., Jullien-Lardo, V., Davis, R. J., Blom, N., and Hojlund-
Nielsen, P. E., in “New Frontiers in Catalysis” (L. Guczi, L. Solymosi
and P. Tetenyi, Eds.), Vol. B, p. 1033. Elsevier, Amsterdam, 1993.
1. Constantino, V. R. L., and Pinnavaia, T. J., Inorg. Chem. 34(4), 883
(1995).
2. Cavani, F., Triffiro, F., and Vaccari, A., Catal. Today 11, 173 (1991).
3. Sugier, A., and Freund, E., U.S. Patent 4,112,110, Institut Franc¸ais du
Pe´trole, 1978.
4. Alcaraz, J. J., Arena, B. J., Gillispie, R., and Holmgren, J. S., in “Pro-
ceedings, 15th Meeting of the North American Catalysis Society, 25. Fishel, C. T., and Davis, R., Langmuir 10(1), 159 (1994).
Chicago, 1997.” p. 114.
5. Gine´s, M. J. L., Amadeo, N., Laborde, M., and Apesteguı´a, C. R., Appl.
Catal. 131, 283 (1995).
26. Sato, T., Fujita, H., Endo, T., and Shimada, M., Reactivity of Solids 5,
219 (1988).
27. Mc Clure, D., J. Phys. Chem. Solids 3, 311 (1957).
6. Di Cosimo, J. I., Dı´ez, V. K., Apesteguı´a, C. R., Xu, M., and Iglesia, 28. Kanno, T., and Kobayashi, M., in “Acid-Base Catalysts II” (M. Misono
E., in “Proceedings, 15th Meeting of the North American Catalysis
Society, Chicago, 1997.” p. 58.
7. Reiche, W. T., J. Catal. 94, 547 (1985).
and Y. Ono, Eds.), p. 207. Kodansha-Elsevier, Tokyo, 1994.
29. Nakatsuka, T., Kawasaki, H., Yamashita, S., and Kohjiya, S., Bull.
Chem. Soc. Japan 52, 2449 (1979).
8. Velu, S., and Swamy, C. S., Appl. Catal. 119, 241 (1994).
30. Rohrer, C. L., and Rohrer, G. S., Chem. Mater. 6, 501 (1994).
9. Corma, A., Iborra, S., Primo, J., and Rey, F., Appl. Catal. 114, 215 31. Noller, H., and Ritter, G., J. Chem. Soc. Faraday Trans. 1(80), 275
(1994).
(1984).
10. Gine´s, M. J. L., Iglesia, E., Di Cosimo, J. I., and Apesteguı´a, C. R., in
preparation.
11. Schaper, H., Berg-Slot, J. J., and Stork, W. H. J., Appl. Catal. 54, 79
(1989).
32. Kibby, C. L., and Hall, W. K., J. Catal. 29, 144 (1973).
33. Tanabe, K., Misono, M., Ono, Y., and Hattori, H., “New Solid Acids
and Bases,” Kodansha-Elsevier, Tokyo, 1989.
34. Canesson, P., and Blanchard, M., J. Catal. 42, 205 (1979).
35. Jain, J. R., and Pillai, C. N., J. Catal. 9, 322 (1967).
12. Corma, A., Forne´s, V., and Rey, F., J. Catal. 148, 205 (1994).
13. Xu, M., Gine´s, M. J. L., Hilmen, A., Stephens, B. L., and Iglesia, E., 36. Padmanabhan, V. R., and Eastburn, F. J., J. Catal. 24, 88 (1972).
J. Catal. 171, 130 (1997). 37. Luy, J. C., and Parera, J. M., Appl. Catal. 26, 295 (1986).
14. Padro´, C. L., Grosso, W. E., Baronetti, G. T., Castro, A. A., and Scelza, 38. Gervasini, A., Fenyvesi, J., and Aroux, A., Catal. Lett. 43, 219
O. A., in “New Developments in Selective Oxidation” (V. Corte´s (1997).
Corbera´n and S. Vic Bello´n, Eds.), p. 411. Elsevier, Amsterdam, 1994. 39. Winterbottom, J. M., in “Catalysis,” Vol. 4, p. 141. The Royal Society
15. Peng, X. D., Richards, D. A., and Stair, P. C., J. Catal. 121, 99 (1990).
16. Balakrishnan, K., and Schwank, J., J. Catal. 127, 287 (1991).
17. Morterra, C., Ghiotti, G., Boccuzzi, F., and Coluccia, S., J. Catal. 51,
299 (1978).
of Chemistry, Burlington House, London, 1981.
40. Balaceanu, J. C., and Jungers, J. C., Bull. Soc. Chim. Belg. 60, 476
(1951).
41. Ross, R. A., and Bennett, D. E. R., J. Catal. 7, 289 (1967).