Impact of the memory effect on the catalytic activity of Li-Al hydrotalcite-like compounds for the cyanoethylation reaction

Article abstract of DOI:10.1016/j.materresbull.2010.06.003

The target of this work is to investigate structural and textural modifications of Li-Al hydrotalcite-like compounds by memory effect and their impact on the catalytic activity of these solids for the cyanoethylation reaction. The samples have been characterized by XRD, FT-IR, TG-DTG, surface area and porous structure. The catalytic test results showed that the reconstruction of the hydrotalcite leads to a decrease of the catalytic activity compared to the parent hydrotalcite. This fact may be related either to the smaller specific surface area of the reconstructed form or to the leaching of Li⁺ during reconstruction. It has been also noticed that the calcination of Li-Al hydrotalcite leads to the formation of a α-LiAlO ₂ stable phase.

Full text of DOI:10.1016/j.materresbull.2010.06.003

Materials Research Bulletin 45 (2010) 1106–1111
Contents lists available at ScienceDirect
Materials Research Bulletin
journal homepage: www.elsevier.com/locate/matresbu
Impact of the memory effect on the catalytic activity of Li–Al hydrotalcite-like
compounds for the cyanoethylation reaction
a
b
a
O.D. Pavel a, , R. Za˘voianu , R. Bırjega , E. Angelescu
*
ˆ
a University of Bucharest, Faculty of Chemistry, Department of Chemical Technology and Catalysis, 4-12 Regina Elisabeta Blv., Bucharest 030018, Romania
b National Institute for Lasers, Plasma and Radiation Physics, 409 Atomistilor Street, PO Box MG-16, Magurele 077125, Romania
A R T I C L E I N F O
A B S T R A C T
Article history:
The target of this work is to investigate structural and textural modifications of Li–Al hydrotalcite-like
compounds by memory effect and their impact on the catalytic activity of these solids for the
cyanoethylation reaction. The samples have been characterized by XRD, FT-IR, TG–DTG, surface area and
porous structure. The catalytic test results showed that the reconstruction of the hydrotalcite leads to a
decrease of the catalytic activity compared to the parent hydrotalcite. This fact may be related either to
the smaller specific surface area of the reconstructed form or to the leaching of Li+ during reconstruction.
Received 12 October 2009
Received in revised form 8 April 2010
Accepted 2 June 2010
Available online 10 June 2010
Keywords:
It has been also noticed that the calcination of Li–Al hydrotalcite leads to the formation of a
stable phase.
a-LiAlO2
Cyanoethylation
Layered double hydroxides
Li–Al hydrotalcite
Memory effect
ß 2010 Elsevier Ltd. All rights reserved.
1. Introduction
In Li–Al hydrotalcites, the structure is constituted by the layer
containing the cations in which octahedral Al sites are arranged as
in gibbsite and the octahedral vacancies are filled by lithium ions
[8] resulting in a positively charged layer that requires an
interlayer anion for charge balance.
Li–Al hydrotalcites have been used as adsorbents for photo-
chromic dye compounds [17], for Cr(VI) [9], or in the adsorption
and controlled release of pesticides [18]. Also these compounds
were used as precursor for LiAlO2 which is applied in various areas
of electrochemical energy conversion and storage, in particular as
an electrolyte support material for molten carbonate fuel cells [15]
or as irradiation blanket for nuclear fusion reactors [14]. In what
concerns the catalytic activity of Li–Al hydrotalcites, these solids
have been used as precursors for mixed-oxides catalysts used in
biodiesel production [12], or for NiLi–Al catalysts used in hydrogen
production [19].
For the conventional Mg–Al hydrotalcites, the reconstruction of
the structure through memory effect was explained by two
mechanisms: a retro-topotactic process involving Al3+ and bivalent
cations located both in the octahedral sites of mixed oxide [20]
and/or a dissolution-recrystallization process [21]. The dissolu-
tion-recrystallization process has been noticed by Thomas and
Kamath [22] for Li–Al hydrotalcite obtained from LiOH and
different Al precursors such as AlCl3 or Al(NO3)3.
The aim of this paper was to study the impact of the memory
effect on the modification of the structure and texture of a Li–Al
hydrotalcite obtained from Al(i-C3H7O)3 and LiOH precursors as
well as on the catalytic activity of these solids for the
cyanoethylation of ethanol with acrylonitrile. This type of test
In the last decade many efforts were dedicated to explain the
memory effect of hydrotalcites which have the general formula:
3+
[M2+
M
1ꢀx
x(OH)2]x+[Anꢀx/n]ꢁmH2O, where M2+ and M3+ are respec-
tively divalent and trivalent metals like Mg2+ and Al3+, Anꢀ is an
anion, x can have values between 0.2 and 0.33, and m is the amount
of water moles [1–5]. The possibility of preparing hydrotalcites with
general formula [M1+
M
1ꢀx
x(OH)2](2xꢀ1)+[Anꢀ2xꢀ1/n]ꢁmH2O is limit-
3+
ed only to Li+, which has an ionic radius comparable to that of Mg2+
.
With other monovalent cations, such as Na+, K+ and (NH4)+, the
corresponding double hydroxide of Al are dawsonite-type com-
pounds [6–8].
The most common methods for the obtaining Li–Al hydro-
talcites are either the intercalation of Li+ into gibbsite, bayerite and
nordstrandite [9–11] or the precipitation route using
Al(NO3)3ꢁ9H2O and LiOHꢁH2O as starting materials [7,8,12].
LiAl2(OH)7ꢁ2H2O can be obtained following the reaction between
hydrated alumina gel and LiOH in alcoholic medium [13]. Other
preparation routes are sol–gel methods using lithium methoxide
and aluminum-sec-butoxide with ethylacetoacetate ligand [14] or
mechanical activation of mixtures of Al(OH)3 and various lithium
salts, such as LiCl, LiF, LiBr, Li2SO4 and Li2CO3 [15,16].
In usual hydrotalcites the M2+ and M3+ are placed in the
octahedral sites and there is no evidence for ordering the cations.
* Corresponding author. Tel.: +40 214103178x138; fax: +40 213159249.
E-mail address: octaviandumitru.pavel@g.unibuc.ro (O.D. Pavel).
0025-5408/$ – see front matter ß 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2010.06.003

O.D. Pavel et al. / Materials Research Bulletin 45 (2010) 1106–1111
1107
reaction has been chosen since it represents a selective route to
obtain beta-ethoxy-proprionitrile and is also influenced by the
at 30 8C. This method has been successfully applied for the
determination of base sites in Mg–Al hydrotalcites, and Co-
modified Mg–Al hydrotalcites [23,24] and it is more suitable for
the comparison of the basicity of all the catalysts investigated in
this study taking into account that the determination of the base
sites using TPD of CO2 can be applied only for the calcined
hydrotalcite samples and cannot be applied to the dried
hydrotalcite-type samples. The total number of base sites is
related to the amount of adsorbed acrylic acid, the number of
strong surface base sites is related to the amount of adsorbed
phenol, while the number of weak base sites is given by the
difference between the amount of adsorbed acrylic acid and the
adsorbed phenol. Catalyst samples (0.05 g) previously degassed
under vacuum were introduced in brown sealable bottles and
10 mL of freshly prepared solution containing the organic acid (e.g.
phenol and acrylic acid, respectively) in methylcyclohexane were
added. The bottles were sealed and were shaken for 2 h at 30 8C,
since the time required for the system to reach the equilibrium at
constant temperature was checked for each solid and it was never
longer than 1 h. It was assumed that the interaction of the solids
with atmospheric CO2 and water was negligible since their
exposure to atmosphere was limited to the weighing period.
After 2 h the concentration of the organic acid remained in
solution was determined spectrophotometrically at lmax (e.g.
276.5 nm for phenol and 231.5 nm for acrylic acid) using the
calibration curves absorption versus acid concentration in the
range where the Lambert–Beer law is fitted. The amount of
adsorbed organic acid, is given by the difference between the
amount introduced in the test and the amount that remained in
solution after 2 h. For each catalyst sample 2–3 measurements
were performed in order to check the reproductibility of the
method. The standard deviation for a set of 2–3 determinations
was approximately 5%.
base character of the solid.
A correlation of the catalytic
performances with the physico-structural characteristics of the
catalysts has been done.
2. Experimental
2.1. Preparation of the catalysts
A
layered double hydroxide with chemical composition
LiAl2(OH)7ꢁ2H2O was prepared by a wet chemical route of gel to
crystallite conversion at 80 8C involving the reaction of hydrated
alumina gel Al(OH)3ꢁxH2O with LiOH (Li2O/Al2O3 > 0.5) in pres-
ence of hydrophilic solvents such as ethanol under refluxing
conditions, as indicated by Nayak et al [13].
The hydrated alumina gel was obtained by hydrolysis of 25 g
Al(i-C3H7O)3 (p.a. Merck) in 500 mL bi-distilled water during 24 h
at 40 8C. Then Al(OH)3ꢁxH2O gel was filtered, washed free of alcohol
using hot bi-distilled water, and further was dried at 90 8C for 12 h.
To obtain Li–Al hydrotalcites, the alumina gel (5 g Al(OH)3ꢁxH2O
with 11.25 Al wt.%) was contacted with an ethanolic solution of
lithium hydroxide (fourfold excess of Li over the aluminum
content Li/Al = 4 in a mixture of water–ethanol, volume ratio 1:1)
in a conical flask. The reaction vessel was provided with a water-
cooled condenser, and an alkali guard tube to prevent the
contamination with CO2. The reaction mixture was refluxed for
8 h under continuous magnetic stirring. The solid product obtained
was washed under nitrogen with decarbonized water until free of
un-reacted LiOH (pH = 8–8.5) then filtered and dried at 90 8C for
24 h. The hydrotalcite obtained was called A1. The sample of mixed
oxides obtained by calcination of A1 at 460 8C for 18 h under
nitrogen flow was named A2. The reconstructed layered double
hydroxide (sample named A3) was obtained by re-hydration of A2
by immersion in decarbonized water at room temperature during
24 h, followed by drying in nitrogen at 90 8C for 24 h. A3 sample
was calcined in nitrogen flow for 18 h at 460 8C to obtain the mixed
oxides (sample denoted A4).
2.3. Catalytic measurements
Catalytic tests were performed during 5 h reaction time in a
glass batch reactor equipped with a condenser system. Acryloni-
trile (0.01 mol) and ethanol (0.03 mol) were stirred and heated to
the reflux temperature (ca. 90 8C). Once the reaction temperature
was stabilized, the catalysts (3% weight reported to the weight of
the reagents) were added and the reaction started up. These
reaction conditions (e.g. molar ratio ethanol/acrylonitrile = 3,
catalyst concentration in the reaction mixture 3% wt.) have been
chosen taking into account that they lead to higher acrylonitrile
conversions without the formation of by-products, as it has been
shown in one of our previous publications [25]. The reaction
mixture was analyzed by GC using a K072320 Termo-Quest
chromatograph equipped with a FID detector and a capillary
column of 30 m length and 0.324 mm diameter and DB-5
stationary phase. Highly pure N2 (99.999%) was used as carrier gas.
2.2. Characterization of catalysts
The chemical composition of A1–A4 samples before and after
the catalytic tests was determined using inductively coupled
plasma-optical emission spectroscopy, ICP-OES for Li and Al
determination and an elemental analyzer Carlo-Erba for the
determination of C content. For the characterization of the used
samples we have performed catalytic tests during 5 h at 90 8C
using 3 g of catalyst and 100 g reaction mixture containing ethanol
and acrylonitrile in a molar ratio 3/1.
The X-ray powder diffraction patterns were collected on a
DRON-3 X-ray diffractometer with
a
nickel filtered Cu-K
a
˚
radiation (lCu-K = 1.5418 A) in a 2
u range of 5–708, a step width
a
of 0.058 and an acquisition time of 2 s on each step.DRIFTS spectra
were recorded in the domain 400–4000 cmꢀ1 with a Varian 3100
Excalibur spectrometer equipped with diffuse reflectance acces-
sory Harrick Praying Mantis. DRIFTS spectra averaged over 200
scans were refined by substracting the spectrum of KBr used as
background.
3. Results and discussion
3.1. Catalysts characterization
The results of the chemical analyses for A1–A4 samples are
presented in Table 1.
The TG–DTG analysis was performed up to 500 8C on a
Setaram—92 equipment with a heating rate of 10 8C/min using
As it may be seen from these data, all catalysts present low
concentrations of C due to the contamination with CO2 during
handling. It may also be noticed that the concentration of Li in A3
sample is lower than the one in the A1 sample and consequently
the concentration of Li in the sample A4 is lower than the
concentration in A2. This fact may be due to leaching of Li during
the rehydration step. The analyses of the catalysts used in the
catalytic tests did not reveal a significant leaching under the
a
-Al2O3 as reference.
The N2 adsorption–desorption isotherms for textural analysis
were collected on Micromeritics ASAP 2020 Instrument.
The surface base sites of the catalysts were determined using a
method based on the irreversible adsorption of organic acids, e.g.
acrylic acid, pKa = 4.2 and phenol pKa = 9.9 on the surface base sites

1108
O.D. Pavel et al. / Materials Research Bulletin 45 (2010) 1106–1111
Table 1
Table 2
Chemical compositions of Li–Al catalysts before and after the catalytic tests (wt.%).
Structural parameters of hydrotalcite-like phases before (A1) and after reconstruc-
tion by memory effect (A3).
Catalyst sample
Li
Al
C
N
Li/Al molar ratio
Samples
Structural parameters
I0 0 3
I3 0 0
/
D
(%)a
Ih0 0 3
/
Ih3 0 0
/
A1 – fresh
A1 – used
A2 – fresh
A2 – used
A3 – fresh
A3 – used
A4 – fresh
A4 – used
3.28
3.18
9.85
9.72
2.85
2.79
8.62
8.55
25.50
24.71
38.33
37.85
25.10
24.93
37.73
37.51
1.13
5.84
0.22
5.03
0.71
5.47
0.52
5.38
0
0.50
0.50
1.00
1.00
0.44
0.44
0.89
0.89
b
b
I0 0 3
I3 0 0
˚
˚
0.58
0
a (A)
c (A)
A1
A3
5.088
5.096
22.584
22.715
4.16
4.32
3.03
7.21
0.42
0.43
0.65
0
a
0.55
0
Degree of disorder along ‘‘c’’ axis calculated from the ratio between FWHM (full
width half maximum) and the area of the diffraction maximum (0 0 3).
b
0.62
Ratios between the intensities of the diffraction lines (0 0 3), and (3 0 0)
respectively of the hydrated sample (A3) and the initial sample (A1).
reaction conditions since the molar ratio Li/Al was the same as in
the fresh catalysts. The slight increase of the carbon content and
the occurrence of nitrogen may be a consequence of the adsorption
of the organic compounds from the reaction mixture.
The X-ray diffraction patterns of the samples A1–A4 are
presented in Fig. 1.
diffraction lines in the samples, since almost the same values
are obtained for both the ratios of the diffraction maxima
corresponding exclusively to ‘‘a’’ axis, e.g. I3 0 0 and to ‘‘c’’ axis,
e.g. I0 0 3, showing that we are dealing with a global behavior
instead of an exclusive modification of the interlayer composi-
tion.
A1 sample has a structure typical for LiAl2(OH)7ꢁ2H2O layered
double hydroxides (JCPDS 40-0710). The XRD pattern may be
indexed as a primitive hexagonal elementary cell belonging to the
spatial group P32 (1 4 5) (JCPDS 31-0704 or JCPDS 40-0710) or to
the spatial group P63/mcm (1 9 3) [10].
(b) the effect of the reconstruction is the enhancement of the
lattice parameters;
(c) the composition of the interlayer anions indicated by the ratio
I0 0 3/I3 0 0 is not significantly modified;
(d) the degree of disorder along ‘‘c’’ axis (
D (%)) is significantly
In order to obtain the distances corresponding to the basal
reflections (along ‘‘c’’ direction) which may be compared to those
characteristic for the classical Mg–Al hydrotalcite structure, the
diffractogram of A1 has been indexed as a spatial group P32 (1 4 5).
The structure of Li–Al hydrotalcite-like compound is a matter
that is still under debate since there are two different assumptions
for the localization of Li+ ion: either in tetrahedral sites outside the
gibbsite layer of alumina or in octahedral sites of the gibbsite-type
layer similarly to the positions occupied by Al3+ in the brucite-type
layer of Mg–Al hydrotalcite-like compounds [10,13].
higher in the reconstructed structure;
(e) the reconstruction valorizes both LiAlO2ꢁ0.25H2O type phase as
well as the amorphous material which appears after calcina-
tion;the
a
-LiAlO2 phase is especially stable.
The stability of
a
-LiAlO2 phase (identified in sample A2)
following the rehydration (sample A3) and re-calcination (sample
A4) is indicated by the structural data presented in Table 3.
The examination of the evolution of the oxide phase
upon calcination, hydration and re-calcination lead to the
following observations: the oxide phase formed presents a typical
a-LiAlO2
The diffraction pattern of the hydrated sample A3 reveals that
almost 50% of the initial hydrotalcite-type phase has been
reconstructed. The result may be explained by the high stability
a
-LiAlO2 structure with lattice parameters higher than those
˚
˚
corresponding to the standard (a = 2.799 A and c = 14.180 A
respectively (JCPDS 44-0224). The -LiAlO2 phase is preserved
of
a-LiAlO2, since more than 90% of
a
-LiAlO2 from A2 sample
a
which was obtained by calcination of A1 is still present in the
hydrated sample A3.
The comparison between the structural data of the hydro-
talcite-like phases in A1 and A3 presented in Table 2 lead to the
following findings:
upon hydration (sample A3) exhibiting a tendency to preferred
orientation along the direction perpendicular to (1 0 1) plane and
broader peaks (see for example the single (1 1 0) reflection around
2u
= 688).
The results indicate the formation of a defective structure, with
possible vacancies or the presence of amorphous
g-Al2O3-type
(a) the partial reconstruction may be quantified by the values of
the ratios between the intensities of the corresponding
domains covered by the more intense maxima of the crystalline a-
LiAlO2-phase, which made difficult – the deconvolution of the
spectrum. These characteristics are also maintained after hydra-
tion (sample A3) and subsequent calcination (sample A4). The
[(Fig._1)TD$FIG]
hydration alters slightly the structure of
a-LiAlO2 initially formed
after calcination indicating that this phase does not contribute to
the reconstruction of the hydrotalcite-like form. The re-calcination
after hydration leads to an increased crystallinity of the
a-LiAlO2
phase. The enhanced cristallinity is also revealed by the lower
degree of disorder for the same maximum (1 0 4). As a brief
conclusion it may be stated that the memory effect is limited
Table 3
Structural parameters of
a-LiAlO2 in the original mixed oxide sample and those
obtained by memory effect.
Samples
Structural parameters
D
(%)a
˚
˚
a (A)
c (A)
A2
A3
A4
2.812
2.811
2.805
14.212
14.207
14.214
11.3
11.8
9.4
a
Degree of disorder along the axis perpendicular to (1 0 4) plane calculated from
the ratio of FWHM (full width half maximum) and the area of the diffraction
maximum (1 0 4).
Fig. 1. XRD patterns of A1–A4 samples, ", LiAl2(OH)7ꢁ2H2O; o,
LiAlO2ꢁ0.25H2O.
a LiAlO2; and #,

O.D. Pavel et al. / Materials Research Bulletin 45 (2010) 1106–1111
1109
probably due to the system lower affinity for water as well as to the
partial extraction of the lithium from the cationic layers which
leads to the decrease of their positive charge and subsequently to
the modification of the capacity to attract the anion layers and to
assemble in layered structures.
shifted at higher wavenumbers 3630 and 3537 cmꢀ1 indicating the
presence of isolated hydroxyl groups with increased base character
while the shoulders of these maxima at 3432 cmꢀ1 indicate the
presence of OH groups in hydrogen bond stretching in the
interlayer OHꢀ groups and those at lower wavenumbers are
characteristic to the stretching of the hydrogen bond formed
between interlayer water and the carbonate ion [28]. The bands
characteristic to the deformation of water molecules around 1600–
1700 cmꢀ1 are clearly defined in the spectra of the calcined
samples A2 and A4 while they are overlapped by the stronger
bands of Li2CO3 in the spectra of dried samples A1 and A3.
As it may be seen from al DRIFT spectra, the carbonate anion is
mainly found in its monodentate form showing the two
characteristic bands at 1400 and 1510 cmꢀ1 [29] while the other
two forms, bidentate and hydrogen–carbonate, respectively, are
negligible.
Fig. 2 presents the DRIFTS spectra of A1–A4 samples.
The DRIFT spectra of the dried and the rehydrated samples A1
and A3, respectively, show more intense bands than those of the
calcined samples A2 and A4. There is a strong similarity between
the spectra of A1 (the fresh hydrotalcite) and A3 (the rehydrated
sample) indicating the reconstruction of the structure through
memory effect. However it may be noticed that the intensity of the
bands in the spectrum of the rehydrated sample is lower probably
due to the fact that the reconstruction has not been complete.
Both calcined samples show the bands characteristic to
a-
LiAlO2 at 835 and 666 cmꢀ1 [26,27]. These two maxima may be
also observed in the DRIFT spectra of the rehydrated sample A3,
while they are absent in the spectrum of the fresh sample A1. In the
region 400–550 cmꢀ1 there are two bands at 450, 484 cmꢀ1 and a
shoulder at 510 cmꢀ1 in the spectra of A1, A3, and A4 related to the
vibrations of AlO6 and LiO6–AlO6 combination interactions [14]
and only one band and a shoulder in the spectrum of A2 at 470 and
500 cmꢀ1. In all samples one may notice the presence of Li2CO3,
impurity due to the presence of the doublet at 1418 and
1510 cmꢀ1. The intensity of these bands is significantly lower
for the calcined samples and it decreases in the following order:
A1 > A3 > A4 > A2. The presence of Li2CO3, impurity was unex-
pected since the reaction mixture used in the synthesis did not
include any carbonate containing compound (such as Na2CO3) and
the installation was equipped with a CO2-trap tube. This fact
suggests that the contamination with CO2 takes place during the
short contact with the atmosphere while handling the sample.
In the region characteristic to hydroxyl groups it may be noticed
From the data presented in Table 4, it may be noticed that the
sample A4 presents a decrease of the specific surface area
compared to the sample A2. This fact is going to be reflected
also by its influence on the catalytic activity performances of this
sample.
The determinations of the surface base sites of the catalysts
using the method based on the adsorption of acrylic acid, pKa = 4.2
and phenol pKa = 9.9 lead to the results presented in Table 5. It may
be seen that both the fresh sample A1 and the corresponding mixed
oxide A2 exhibt higher basicity compared to the rehydrated
sample A3 and the recalcined one A4. This trend is rather different
to that observed in the case of classical Mg3Al hydrotalcites where
the rehydrated sample has a higher basicity than the fresh
hydrotalcite. However, considering the chemical composition of
A1–A4 samples, the lower basicity of A3 and A4 samples may be a
consequence of the loss of Li during the rehydration step.
Fig. 3 presents the TG–DTG diagram corresponding to A1 and A3
samples.
[(Fig._2)TD$FIG]
that the spectra of both calcined samples present a maximum
It may be seen that the solid A1 exhibits only a decrease of
weight due to the loss of water following the condensation of
different OH groups. The decrease of weight due to the removal of
the tiny amount of carbonate adsorbed by the solid during its
handling in atmosphere is very low and cannot be distinguished in
the TG–DTG diagram. For A1, the total loss of weight in percentages
Table 4
The textural properties of the mixed oxides.
Samples
Specific surface
area (m2/g)
Pore volume
(cm3/g)
Mean pore
˚
diameter (A)
A2
A4
151
124
0.4230
0.2318
56
58
Table 5
Distribution of base sites for the investigated catalysts.
Catalysts
Total number of base Distribution of base sites
sites (mmol acrylic
Strong base sites Weak and medium
acid/g catalyst)
(mmol phenol/g
catalyst)
strength base sites
(mmol/g catalyst)a
A1
8.87
9.25
7.32
7.58
6.73
0.37
0.42
0.35
0.31
0.26
0.34
0.38
8.5
A2
8.83
6.97
7.27
6.47
6.87
7.98
A3
A4
Mg3Al HTb
Mg3Al RHTb 7.21
Mg3Al CHTb 8.36
a
Weak and medium strength base sites =total number of base sites ꢀ strong base
sites).
b
Fig. 2. DRIFT spectra of A1–A4 samples.
Previously published results [24].

[
O.D. Pavel et al. / Materials Research Bulletin 45 (2010) 1106–1111
4
Fig. 4. The evolution of the catalytic activity for the samples A1–A4 during 5 h
reaction time.
Fig. 3. TG–DTG diagram of A1 and A3 samples, (A1 continuos line; A3 dashed line).
A1–A4.
CA1 ¼ 6:56 t þ 57:052 correlation coefficient R2
represents 40% and the residue consisting of LiAl mixed oxides is
60%. Meanwhile for the reconstructed hydrotalcite structure A3
the total loss of weight is only 35% probably due to the fact that in
this case there is a lower amount of CO3 anions and structural
water as it was also revealed by the chemical composition of these
compounds showed in Table 1.
¼ 0:9979; CA1;0 ¼ 57:052 %
(1)
(2)
(3)
(4)
CA2 ¼ 5:784 t þ 70:406 correlation coefficient R2
¼ 0:9737; CA2;0 ¼ 70:406 %
2ꢀ
CA3 ¼ 10:746 t þ 24:38 correlation coefficient R2
¼ 0:9747; CA3;0 ¼ 24:38 %
Considering the data published by Cavani et al. [1], the weight
loss for common hydrotalcites Mg–Al occurs in three stages
namely: (a) an initial weight loss due to the removal of physisorbed
water in the temperature range 20–100 8C; (b) a weight loss due to
the removal of structural water in the temperature range 100–
190 8C; and (c) the weight loss due to the dehydroxylation of the
surface at higher temperatures. The last stage is characterized in
the DTG profile by a wide peak occurring in the temperature range
190–500 8C. According to Serna and co-workers [8,28] lithium
hydrotalcite with interlayer carbonate anions exhibits a thermal
behavior rather different from that of its Ni or Mg analogues with
an M2+/A13+ ratio of 2 and an octahedral cation arrangement
similar to that of hydrotalcite. Their results showed that in these
compounds, interlayer water is lost between 200 and 300 8C and
that during the calcination up to 600 8C there is a loss of all the
hydroxyl groups, in the form of H2O, and 2/3 of the carbonate, as
CO2. The same authors indicate that when heating the samples
between 150 and 1000 8C, there are three regions where the loss of
weight occurs, e.g. 30% weight loss between 150 and 290 8C, 16% in
the temperature range 290–500 8C, and 3.4% for the temperature
interval 500–1000 8C.
CA4 ¼ 7:27 t þ 53:893 correlation coefficient R2
¼ 0:9853; CA4;0 ¼ 53:893 %
After 5 h reaction time, the sample A1 exhibits a satisfactory
activity leading to a conversion of 90%. This fact is due both to the
strong basicity of OHꢀ groups and the presence of Li+. For the
sample A2, the conversion increases up to 98%. Meanwhile, the
reconstructed hydrotalcite sample A3 obtained following the
hydration of A2 presents a lower activity compared to the original
hydrotalcite sample A1. This behavior is due especially to the fact
that during the rehydration a leaching of 12.5 wt.% Li+ (reported to
the initial concentration of Li in A1 sample) occurs. Another
important factor that affects the activity is the presence of
a-LiAlO2
phase which has high stability under hydration and re-
a
calcination. The decrease of the catalytic activity of the sample
A4 compared to A2 is related both to the lower amount of Li and the
decrease of the specific surface area.
The TG diagrams of A1 and A3 samples show that our Li–Al
hydrotalcites with low concentrations of carbonate anions present
only a 10% weight loss between 150 and 290 8C, and 17.2% weight
loss in the temperature range 290–500 8C. The differences between
these results and those obtained by Serna and co-workers [8,28]
may be a consequence of the fact that our samples were heated
only between 30 and 500 8C.
Taking into account that the cyanoethylation reaction is
promoted by the base sites of the catalysts according to the
reaction mechanism described in Scheme 1, it has been attempted
to find a correlation between the number of base sites and the
values of the conversions reached in the presence of A1–A4
catalysts. The base sites which may be implicated in the process
have weak, medium and high strength and they correspond to O2ꢀ
,
Li–O ion pairs, and OH groups. The first two types possess Lewis
base character while OH groups act as Bro¨nsted base sites [12]. The
results plotted in Fig. 5 were also compared to those obtained in
our previous work [2] when using classical Mg3Al hydrotalcites,
3.2. Catalytic tests
Fig. 4 presents the evolution of the catalytic activity for the
samples A1–A4 during 5 h reaction time.
[(cShem_)1DT$FI]G
The obtained results showed that for all catalysts there is
almost a linear trend of the conversion increase up to 5 h. The
analyses of the reaction mixture showed that the selectivity to
beta-ethoxypropionitrile was always higher than 99%. The
eqs. (1–4) describing the linear trends for the increase of the
conversion (CAi; i = 1–4) as a function of the reaction time (t)
allowed us to estimate the initial conversion (CAi,0) for the samples
Scheme 1.

[
O.D. Pavel et al. / Materials Research Bulletin 45 (2010) 1106–1111
1111
without carbonate anions due to the contamination with
atmospheric CO2. In the case of Li–Al hydrotalcite compounds,
there is a partial reconstruction of the layered double hydroxide
phase through memory effect which is accompanied by a leaching
of Li+. The partial reconstruction may be due to the fact that the
calcination of Li–Al hydrotalcite leads to the formation of a
a-
LiAlO2 stable phase in A2 and this phase is present also in the
samples A3 and A4 as indicated by XRD.
Following the re-calcination a solid with a lower specific surface
area than the initial mixed oxide is obtained. The initial mixed
oxide A2 leads to high a conversion (98%), which makes it an
excellent catalyst for the cyanoethylation of ethanol with
acrylonitrile.
Li–Al catalysts are more active than the corresponding Mg–Al
hydrotalcite compounds for the cyanoethylation of ethanol with
acrylonitrile due to their higher overall basicity.
Acknowledgments
This work was supported by CNCSIS –UEFISCSU, project number
PNII – IDEI 534/15.01.2009, code 1907/2008.
References
[1] F. Cavani, F. Trifiro, A. Vaccari, Catal. Today 11 (1991) 173–301.
[2] O.D. Pavel, R. Bˆırjega, M. Che, G. Costentin, E. Angelescu, S. S¸erban, Catal. Commun.
9 (2008) 1974–1978.
[3] E. Angelescu, O.D. Pavel, R. Bˆırjega, M. Florea, R. Za˘voianu, Appl. Catal. A: Gen. 341
(2008) 50–57.
[4] T. Kawabata, N. Fujisaki, T. Shishido, K. Nomura, T. Sano, K. Takehira, J. Mol. Catal.
A: Chem. 253 (2006) 279–289.
[5] A.E. Palomares, J.G. Prato, F. Rey, A. Corma, J. Catal. 221 (2004) 62–66.
[6] M.J. Hernandez-Moreno, M.A. Ulibarri, J.L. Rendon, C.J. Serna, Phys. Chem. Miner.
12 (1985) 34–38.
[7] I. Sissoko, E.T. Iyagba, R. Sahai, P. Biloen, J. Solid State Chem. 60 (1985) 283–288.
[8] C.J. Serna, J.L. Rendon, J.E. Iglesias, Clay Clay Miner. 30 (1982) 180–184.
[9] S.L. Wang, R.J. Hseu, R.R. Chang, P.N. Chiang, J.H. Chen, Y.M. Tzou, Colloid Surface
A: Physicochem. Eng. Aspect 277 (2006) 8–14.
[10] A.V. Besserguenev, A.M. Fogg, R.J. Francis, S.J. Price, D. O’Hare, V.P. Isupov, B.P.
Tolochko, Chem. Mater. 9 (1997) 241–247.
Fig. 5. Correlation of the catalytic activity with the basicity of the catalysts: (A) as a
function of the total number of base sites and (B) as a function of the ratio between
weak and medium strength base sites and strong base sites.
[11] A.M. Fogg, A.J. Freij, G.M. Parkinson, Chem. Mater. 14 (2002) 232–234.
[12] J.L. Shumaker, C. Crofcheck, S.A. Tackett, E. Santillan-Jimenez, T. Morgan, Y. Ji, M.
Crocker, T.J. Toops, Appl. Catal. B: Environ. 82 (2008) 120–130.
[13] M. Nayak, T.R.N. Kutty, V. Jayaraman, G. Periaswamy, J. Mater. Chem. 7 (1997)
2131–2137.
[14] F. Oksuzomer, S.N. Koc, I. Boz, M.A. Gurkaynak, Mater. Res. Bull. 39 (2004) 715–
724.
[15] O.A. Kharlamova, R.P. Mitrofanova, V.P. Isupov, Inorg. Mater. 43 (2007) 645–650.
[16] L.T. Menzheres, N.P. Kotsupalo, E.V. Mamylova, J. Mater. Synth. Process. 7 (1999)
239–244.
[17] T. Kuwahara, H. Tagaya, K. Chiba, Microporous Mater. 4 (1995) 247–250.
[18] A. Ragavan, A.I. Khan, D. O’Hare, J. Phys. Chem. Solid 67 (2006) 983–986.
[19] F. Melo, N. Morlane´s, Catal. Today 133–135 (2008) 374–382.
[20] A.J. Marchi, C.R. Apesteguia, Appl. Clay Sci. 13 (1998) 35–48.
[21] K. Takehira, T. Kawabata, T. Shishido, K. Murakami, T. Ohi, D. Shoro, M. Honda, K.
Takaki, J. Catal. 231 (2005) 92–104.
e.g. fresh (Mg3Al HT), calcined (Mg3Al CHT) and rehydrated (Mg3Al
RHT) under similar reaction conditions. The number of base sites in
the above mentioned Mg3Al samples has been determined using
the same method and the values were presented in our previous
publication [24].
From Fig. 5A it may be seen that the conversions increase with
the total amount of base sites in the catalyst and conversions above
90% are obtained for the catalysts exhibiting a total number of base
sites higher than 7.5 mmol/g.
Fig. 5B shows that there is also a correlation of the catalytic
activity and the base sites strength distribution, reflected by the
ratio between weak plus medium strength base sites and the
strong base sites. The results plotted in Fig. 5B indicate that 21 is
the optimum value of the above mentioned ratio of base sites
required to obtain conversions higher than 95%.
[22] G.S. Thomas, P.V. Kamath, Mater. Res. Bull. 37 (2002) 705–713.
[23] K. Parida, J. Das, J. Mol. Catal. A: Chem. 151 (2000) 185–192.
[24] R. Ionescu, O.D. Pavel, R. Bˆırjega, R. Za˘voianu, E. Angelescu, Catal. Lett. 134 (2010)
309–317.
[25] E. Angelescu, O.D. Pavel, R. Za˘voianu, R. Birjega, Rev. Roum. Chim. 49 (2004) 367–
375.
[26] P. Tarte, Spectrochim. Acta, Part A 23 (1967) 2127–2143.
[27] K. Nakomoto, Infrared and Raman Spectra of Inorganic and Coordination Com-
pounds part A. Theory and Applications in Inorganic Chemistry, 5th edition,
Wiley, New York, 1997.
[28] M.J. Hernandez, M.A. Ulibarri, J. Cornejo, M.J. Pen˜a, C.J. Serna, Thermochim. Acta
94 (1985) 257–266.
4. Conclusions
The synthesis of Li–Al hydrotalcite-like compounds from
carbonate free starting materials cannot lead to obtain a structure
[29] J.I. Di Cosimo, V.K. Diez, M. Xu, E. Iglesia, C.R. Apesteguia, J. Catal. 178 (1998) 499–
510.