Ind. Eng. Chem. Res. 2004, 43, 4559-4570
4559
Thermal Evolution of the Structure of a Mg-Al-CO3 Layered
Double Hydroxide: Sorption Reversibility Aspects
Yongman Kim,† Weishen Yang,‡ Paul K. T. Liu,§ Muhammad Sahimi,† and
Theodore T. Tsotsis*,†
Department of Chemical Engineering, University of Southern California, Los Angeles, California 90089,
State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences,
Dalian 116023, China, and Media and Process Technology, Inc., Pittsburgh, Pennsylvania 15238
In our prior study, we employed several in-situ techniques to investigate the thermal evolution
of the structure of a Mg-Al-CO3 layered double hydroxide (LDH) under an inert atmosphere
and proposed a model to describe the structural evolution of the Mg-Al-CO3 LDH structure.
In this paper we present results of our ongoing investigations with these materials pertaining
to their sorption characteristics and thermal reversibility under both inert and reactive
atmospheres. The experimental observations are shown to be consistent with the structural
model for the LDH previously proposed. The structure, sorption characteristics, and thermal
reversibility of these materials are of importance in their use for the preparation of CO2permselective membranes for high-temperature membrane reactor applications.
1. Introduction
Layered double hydroxides (LDH), also known as
anionic clays or hydrotalcites,1-3 consist of two types of
metallic cations accommodated with the aid of a closepacked configuration of OH- groups in a positively
charged brucite-like layer. The charge typically results
from the substitution of the lower valence metal in the
brucite structure by a metal of higher valence. The
interlayer space in the LDH is, typically, occupied by
water and various anions for charge compensation. The
general LDH chemical structure is
[M1-xIIMxIII(OH)2]x+[Xx/mm-‚nH2O]
where MII is a divalent (Mg, Mn, Fe, Co, Ni, Cu, Zn,
Ga) and MIII a trivalent metal cation (Al, Cr, Mn, Fe,
Co, Ni, La). Xm- represents m-valence inorganic (CO32-,
OH-, NO3-, SO42-, ClO4-), heteropolyacid (PMo12O403-,
PW12O403-), or even organic acid anions. In the above
formula, typically 0.2 e x e 0.33, but LDH with
significantly higher x values have also been reported.4
The most frequently studied LDH is [Mg1-xAlx(OH)2][(CO3)x/2‚nH2O].
Since these materials have a well-defined layered
structure with nanometer (0.3-3 nm) interlayer distances and contain important functional groups, they
are widely investigated for use as adsorbents for liquid
ions5-7 and gas molecules8,9 and as catalysts for oxidation,10-12 reduction,13 and other catalytic reactions.14-16
Recently, LDH materials have also been investigated
in novel reactive separation processes to increase the
conversion of catalytic reactions by removing one of the
products from the reactor.17,18 In our research these
materials are used for the preparation of nanoporous
* To whom correspondence should be addressed. Tel.:
+1-213-740-2069. Fax: +1-213-740-8053. E-mail: ttsotsis@
worldnet.att.net.
†
University of Southern California.
‡
Chinese Academy of Sciences.
§
Media and Process Technology, Inc.
membranes, intended for the separation of gases, particularly CO2, at high temperatures.19
LDH find applications in a broad range of temperatures, from room-temperature adsorption to high-temperature catalytic reactions and separations. The membranes for CO2 separation, as an example, are being
evaluated for membrane reactor applications in the
water gas shift (WGS) reaction. Some of the functional
groups that LDH contain are known to be sensitive to
high temperatures. Therefore, it is important to understand the thermal evolution of the LDH structure, to
better correlate their performance with their structure
and functional groups. We have recently reported1 the
use of a combination of several in-situ techniques to
investigate the thermal evolution of the structure of a
Mg-Al-CO3 LDH under an inert atmosphere. The
techniques utilized included diffuse reflectance infrared
Fourier transform spectroscopy (DRIFTS), for investigating changes in the functional groups, TG/DTA, for
monitoring the weight and energetic changes, TG/MBMS, for monitoring the weight change simultaneously
with the gaseous products generated, and high-temperature X-ray diffraction (HTXRD), for in-situ monitoring
of the structure evolution. On the basis of the results
of this study, a model was proposed to describe the
structural evolution of the Mg-Al-CO3 LDH. According
to this model,1 as the temperature is increased, loosely
held interlayer water is lost in the temperature range
of 70-190 °C, but the LDH structure still remains
intact. The OH- group, likely in an Al-(OH)-Mg
configuration, begins to disappear at 190 °C and is
completely lost around 280 °C; a gradual transformation
of the LDH structure begins in the same range of
temperatures. The OH- group, likely in a Mg-(OH)Mg configuration, begins to disappear around 280 °C
and is completely lost around 440 °C. A gradual
degradation of the LDH structure is also observed in
the same range of temperatures. Although some CO32loss is observed at lower temperatures, its substantial
loss (∼90%) begins at 400 °C and is completed around
600 °C. At these temperatures the material becomes an
amorphous metastable, mixed solid oxide solution.
10.1021/ie0308036 CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/24/2004
4560 Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004
Table 1. Weight Loss from the TG/MB-MS Studies and
Calculated Weight Loss Based on the ICP Data for the
Samples (a) LDH1 and (b) LDH2
wt loss (wt %)
H2O
ICP value
exptl from
TG/MS
ICP value
exptl from
TG/MS
10.8
12.93
2
2
OH- from
Al
OH- from
Mg
CO2 from
CO32-
tot.
Sample a
10.4
13.2
8.93
12.65
7.2
7.72
41.6
42.23
Sample b
11.0
16.1
12.08
15.94
10.9
11.06
40
41.08
Prior to our study, other investigators had studied the
thermal evolution of the Mg-Al-CO3 structure but did
not propose a complete structural model (for further
discussion, see ref 1). The reasons for that are because
(i) they carried out their surface measurements20 exsitu and the results obtained, therefore, did not really
reflect the actual LDH structure or its functional groups,
as they exist at any given temperature, (ii) they did not
simultaneously monitor the gaseous species evolved21-23
and, therefore, reached the wrong conclusions in the
interpretation of the surface measurements (e.g.,
HTXRD), or (iii) they used a limited number of in-situ
techniques or whose signal was weak (e.g., in-situ
infrared emission spectroscopy) that provided little or
misleading information concerning the interpretation of
the mechanism of the structure evolution.24-26
The advantage in the approach of Yang et al.1 is in
that they have utilized a combination of different insitu techniques, which can provide simultaneous complimentary information on the structural changes, the
type/number of functional groups present, and the
various species generated. This has allowed them to
elucidate the changes that occur in the structure, type,
and number of functional groups in the Mg-Al-CO3
LDH at different temperatures.
In this paper, the thermal evolution of the
Mg-Al-CO3 LDH structure is studied further. The
focus in this study is on validating the structural model
of Yang et al.1 under more realistic reactive environments, in which the LDH materials may find eventual
application. The emphasis in the paper is on the study
of the sorption characteristics and thermal reversibility
of these materials under both inert and reactive atmospheres. In addition to providing a test of the validity
of the LDH thermal evolution model, these experimental
observations are of importance in their own right for
the use of LDH in the preparation of CO2-permselective
membranes for high-temperature membrane reactor
applications.
2. Experimental Section
2.1. Mg-Al-CO3 (LDH) Samples. We have utilized
two LDH samples provided by Media and Process
Technology, Inc., of Pittsburgh, PA. Their composition
was Mg0.71Al0.29(OH)2(CO3)0.15‚0.46(H2O) for sample no.
1 (hereinafter referred to as LDH1; this was the sample
previously studied1) and Mg0.645Al0.355(OH)2(CO3)0.178‚
0.105(H2O) for sample no. 2 (hereinafter referred to as
LDH2), as determined by ICP. The weight loss for the
same samples was also studied by TGA/MB-MS (see
further discussion below), and the results were shown
to be consistent with the ICP results, as Table 1
Figure 1. XRD spectra of the LDH samples (a) LDH1 and (b)
LDH2.
indicates. XRD characterization of these LDH samples
indicates that the materials have the typical LDH XRD
spectra (see Figure 1).21,22,27 The LDH2 sample has a
higher Al/Mg ratio (x being near the higher end of the
typical LDH range) and contains less interlayer water;
it has, in addition, an XRD spectrum that is more noisy
and weaker than that of LDH1 (compare Figure 1a with
Figure 1b). The differences in the XRD spectra of these
two LDH samples of similar composition are also
accompanied by differences in the crystallite sizes, as
TEM studies indicates (Figure 2). The average crystallite size for LDH1 is 0.31 µm, while that for LDH2 is
0.13 µm. The TGA spectra for the two LDH generated
with a scan rate of 5 °C/min (to avoid the presence of
kinetic effectsssee discussion to follow) together with
the CO2 MS signals are shown in Figure 3a, while the
cumulative amounts of H2O evolved are shown in Figure
3b. Though differences in the TGA and mass evolution
spectra exist (Hibino et al.20 also observed differences
in the TGA spectra of Mg-Al-CO3 LDH with a different
Al/Mg ratios but even between LDH with the same
Al/Mg ratio and different crystallite sizes), the sequence
of events during the structural evolution of the two LDH
remain the same; only the range of temperatures where
the different phenomena occur differs. For the LDH2,
the loosely held interlayer water is lost in the temperature range 80-190 °C, the OH- group begins to
Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004 4561
Figure 2. (a) SEM picture of a membrane tube prepared by the
deposition of a thin LDH layer on an underlying macroporous
support, (b) TEM picture of LDH1, and (c) TEM picture of LDH2.
disappear at 190 °C and is completely lost around 520
°C, and while some CO32- loss is observed at lower
temperatures, its substantial loss begins at 450 °C and
is completed at 720 °C. Table 2 shows the fractions (in
terms of the total) of H2O and CO2 that are evolved in
different temperature ranges for both LDH1 and LDH2.
2.2. In-Situ DRIFTS. This is a sensitive and powerful technique that has been utilized in this study to
monitor in-situ the changes of functional groups in the
Mg-Al-CO3 LDH samples as a function of temperature
and other experimental conditions.1 DRIFTS spectra
were recorded in-situ using a Genesis II (Mattson, FTIR) instrument equipped with a DRIFTS COLLECTOR
II chamber (SpectraTech, Inc.) with ZnSe windows,
capable of operating under high temperatures (up to 900
°C) and pressures (up to 1500 psi). The experimental
operating conditions were a DRIFTS scan range from
4000 to 500 cm-1, scan numbers 16, and a scan resolution of 2 cm-1. The spectra were calibrated for background with KBr. To obtain a strong signal intensity
and better resolution for quantitative measurements,
the sample was first ground to 2-10 µm, diluted with
KBr to 5-10 wt %, placed in the sample cup, and leveled
with a spatula. The other experimental conditions will
be indicated, whenever appropriate, throughout the
paper.
2.3. In-Situ TG/MB-MS. These techniques are used
to monitor the weight changes of the LDH samples and
the gases generated during their thermal evolution as
a function of temperature and other conditions. Combined with DRIFTS, TG/MB-MS gives additional quantitative insight into functional group changes at various
conditions. In this study the thermogravimetric (TG)
curve was recorded on a Cahn TGA 121 instrument. The
MB-MS instrument is custom-made, using a MKS UTI
100C Precision Gas Analyzer.
Figure 3. TGA spectra and CO2 MS signal for the two LDH
samples generated with a scan rate of 5 °C/min and (b) cumulative
amount of H2O evolved.
Table 2. Fractions of H2O and CO2 (Percent of the Total
Sample Weight) That Are Evolved in Different
Temperature Ranges for Both LDH1 and LDH2
temp range
(°C)
RTa-100
100-200
200-300
300-400
400-500
500-600
600-750
tot.
a
tot.
LDH1
H2O
2.61
11.39
8.03
12.64
4.8
2.76
2.61
11.31
7.94
11.95
0.7
42.23
34.51
CO2
0.08
0.09
0.69
4.1
2.76
7.72
tot.
0.59
1.48
7.41
8.31
10.42
11.38
1.49
41.08
LDH2
H2O
0.59
1.46
7.33
8.04
9.91
2.69
30.02
CO2
0.02
0.08
0.27
0.51
8.69
1.49
11.06
Room temperature.
2.4. Other Techniques. Though not specifically
discussed here, during our ongoing study of the use of
LDH as high-temperature membrane materials we have
utilized a variety of other techniques. These include insitu differential thermal analysis (DTA) to monitor
energetic changes during the structural evolution of the
LDH materials and in situ high-temperature XRD to
detect the structural changes of Mg-Al-CO3 LDH as
a function of temperature (see ref 1 for further discussion on the use of these techniques for the characterization of the LDH1 sample). Various microscopies (highresolution optical, SEM, TEM, and AFM) provide useful
information on crystallinity and surface characteristics
4562 Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004
of these materials. Figure 2a, for example, shows a SEM
picture of a membrane tube prepared by the deposition
of a thin LDH layer on an underlying macroporous
support tube. Clear is the growth on the membrane
support of a layer of LDH crystals along a preferential
direction. As previously noted, Figure 2b,c shows TEM
pictures of the LDH1 and LDH2 crystallites.
3. Results and Discussion
3.1. Effect of Heating Rate and Gas Atmosphere
on Weight Loss. One of the issues of interest about
the use of the LDH materials in the preparation of CO2permselective membranes and adsorbents pertains to
their ability to function stably under thermal cycling
and other temperature changes in a variety of gaseous
atmospheres. To study the structural stability of these
materials, typically one raises their temperature in a
linear fashion while monitoring weight loss and other
structural characteristics in situ. In the study of Yang
et al.,1 for example, the temperature of the LDH1
sample was raised linearly under vacuum and in an
inert Ar atmosphere, while the sample was studied insitu by a variety of techniques, as previously outlined.
For the heating rates utilized for the LDH1 sample by
Yang et al.,1 the observations that were made up to a
temperature of 250 °C were of equilibrium nature (no
effect of the heating rate), but for some of the heating
rates utilized kinetic effects were apparent above this
temperature. Similar observations were previously made
by Costantino and Pinnavaia28 and most recently by
Rhee and Kang.29
To investigate these issues further, we have studied
the weight-loss characteristics of the LDH2 sample
for 4 different heating rates, namely, 1, 3, 5, and 10 °C/
min and in three different atmospheres. The weightloss results in an inert Ar atmosphere and the corresponding CO2 MS signals are shown in Figure 4a (for
the results in Figure 4, a fresh LDH2 sample ∼110120 mg was used in every experiment). For the results
in Figure 4a, UHP dry inert Ar was utilized as a purge
gas at a flow rate of 20 sccm. For heating rates below 5
°C/min the weight-loss curves and the MS signals
coincide (the results with the 10 °C/min heating rate
contain kinetic artifacts), indicating that the structural
changes (loss of interlayer water, hydroxyl, and CO32-)
occur rapidly enough, so that the LDH structure equilibrates within the time frame allotted by the changing
temperature (similar observations were also made in the
presence of dry and humid CO2 atmospheres, with the
results showing absence of kinetic effects for heating
rates below 5 °C/minssee further discussion below).
The effect of varying the heating rate on weight loss
for the LDH2 sample was also studied in the presence
of a reactive atmosphere. For the experiments in Figure
4b, in addition to the weight loss of the LDH2 sample
in the presence of inert Ar, we also show the weightloss curve for the case in which dry CO2, instead of Ar,
was utilized as a purge gas atmosphere at a flow rate
of 20 sccm/min and a heating rate of 5 °C/min. The
weight-loss results in the presence of a humidified CO2
atmosphere are also shown in the same figure. For the
latter experiments, the other experimental conditions
were the same as those with the other two weight-loss
curves shown in the same figure, with the exception that
the CO2 stream in this case was humidified by bubbling
it through a beaker containing distilled water. Measurements of the water concentration of the gas exiting the
Figure 4. Effects (a) of varying the heating rate and (b) of using
different purging gases on the weight loss for the LDH2 sample.
beaker indicate that the relative humidity (RH) of the
CO2 stream was ∼70%. The results in Figure 4b indicate
little effect of the gaseous atmosphere on the weightloss curve in the first region of temperatures associated
with the evolution of interlayer water. Differences exist,
however, in the other regions. In the region where
mostly CO2 evolves, and DRIFTS indicates that all
hydroxyls in the LDH structure have already disappeared, the weight-loss curves for the humidified and
dry CO2 atmospheres coincide but are still different from
the weight-loss curve under inert conditions; the presence of CO2 in the purge atmosphere appears to slow
down somewhat the rate of CO2 evolution. The dry and
humidified CO2 weight-loss curves are different in the
region associated with hydroxyl evolution, particularly
in the range of temperatures associated with loss of
hydroxyls in a Mg-(OH)-Mg configuration. Previously,
Ding and Alpay,8 who studied CO2 adsorption on a
K-promoted commercial Mg-Al hydrotalcite at 400 °C,
noted a small (∼10%) beneficial effect of the presence
of water on CO2 adsorption. They also noted, however,
that the actual partial water pressure did not really
matter, with even traces of water vapor being capable
of providing the same beneficial effect. Ding and Alpay8
attribute this beneficial effect to the ability of water
vapor to either maintain the hydroxyl concentration on
the surface or to prevent the sites from poisoning
through carbonate or coke deposition.
Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004 4563
Figure 5. Weight gain or loss: (a) weight gain during adsorption for various temperatures as a function of the cycle number; (b) weight
loss during desorption for various temperatures as a function of the cycle number; (c) weight change due to loss of water or CO2 during
desorption as a function of temperature.
3.2. Sorption Reversibility Studies. The ability of
the LDH to reversibly adsorb CO2 and H2O is of
significance in the use of these materials as adsorbents
and membranes. In the former case, the ability to
reversibly adsorb CO2 is critical from the standpoint of
being able to regenerate the adsorbent; in the latter case
the presence of a relatively mobile CO2 phase within
the LDH structure is important in determining the
permeation rate through the membrane. We investigated, therefore, the ability of the LDH materials to
reversibly adsorb CO2 under a broad range of experimental conditions.
To initiate these studies (carried out in the flow
TG/MB-MS system), we first investigated the effect of
the conditions utilized during the experiments. For each
series of experiments, we utilized 100-120 mg of a fresh
LDH sample. During the adsorption part of the cycle,
30 sccm of CO2 was bubbled through a beaker containing distilled water (the CO2 stream’s RH being ∼70%),
and the sample was exposed to this humidified CO2
stream for varying periods of time. Subsequently to
adsorption, the flow of CO2 was shut down and the
desorption part of the cycle was initiated (see below).
The effect of varying the duration of adsorption was
investigated. Increasing the time of adsorption from 1
to 2 h increased the total amount adsorbed by about
5%, but a subsequent increase from 2 h to 3 h had no
additional significant effect. For the remainder of the
study we utilized, therefore, an adsorption step time of
3 h. We also investigated two different methods to carry
out the desorption step. In the first method, upon
termination of the CO2 flow, the sample was exposed
to flowing UHP dry Ar at a rate of 30 sccm. Typically,
after 30 min the weight change of the sample ceased.
Subsequently, we allowed the flow of Ar to continue for
a total desorption period of 1 h. In the second method,
the chamber was evacuated for a period of 1 h at a
pressure below 40 mTorr. Evacuation was shown to be
a more effective means for carrying out the desorption
step (an ∼10% increase in weight gain upon subsequent
readsorption) and was utilized in the remainder of the
study.
Upon completion of the preliminary runs, we studied
the effect of temperature on the adsorption/desoprtion
behavior of the LDH2. The results are shown in Figure
5. Figure 5a shows the total weight gain (as percent
fraction of the original weight of the LDH sample)
during the adsorption part of the cycle. Figure 5b shows
the corresponding total weight loss during the desorption part of the cycle. Figure 5c shows the weight change
corresponding to either the H2O or CO2 lost during the
desoprtion part in the cycle (the TG/MB-MS system
4564 Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004
allows one to monitor not only the total weight change
but in-situ the gaseous species that are emitted during
desoprtion). For each experiment, at any given temperature, a fresh LDH2 sample (100-120 mg) was used.
The temperature was increased from room temperature
at a rate of 5 °C/min in an Ar atmosphere (flow rate of
30 sccm) to the preset point (e.g., 200, 250 °C, etc.) and
kept at this temperature until the sample weight became constant, typically for 20-30 min. Subsequently,
the cyclic sorption/desoprtion experiments were initiated.
One can distinguish three distinct regions of different
behavior in Figure 5. The first region is for temperatures
below 190 °C, where one observes that the LDH2 sample
reversibly adsorbs only water and slight amounts of
CO2. The results of the cyclic adsorption/desorption
experiments in this region of temperatures are consistent with prior observations1 with the LDH1 sample and
our current studies with the LDH2 sample (e.g., Figure
3 and Table 2), which indicate that, in the same region
of temperatures, the interlayer water is removed. The
cyclic adsorption/desorption experiments indicate, in
addition, that the exchange of interlayer water is a fairly
reversible process (see further discussion on the longer
duration cyclic experiments below).
In the second temperature range, from 190 to 280 °C,
our prior1 and current studies with the LDH2 indicate
that the water that leaves the sample is from the
hydroxyl groups that are bonded with Al cations. In
addition, some CO2 is also emitted in this region. The
cyclic sorption/desoprtion studies show that the same
two species are also emitted during the period in which
the sample’s temperature is raised to the desired level.
Upon initiation of the sorption/desorption cycle, however, only CO2 appears to be reversibly adsorbed in this
region, with very little H2O emitted; the sample weight
change can be fully attributed to the reversibly adsorbed
CO2.
In the temperature range 280-440 °C, our prior
studies,1 under inert conditions, have indicated that
the OH- group bonded with Mg2+ begins to disappear
at 280 °C and is completely lost at 440 °C (for the
LDH2 the upper temperature extends higher; see
Table 2); a degradation of the hydrotalcite structure is
also observed in the same region. The cyclic adsorption/desorption experiments indicate that the same
two species are also emitted during the period in which
the sample temperature is raised to the desired level.
Upon initiation of the sorption/desorption cycle, however, only CO2 appears to be reversibly adsorbed in
this region, with very little H2O emitted; hence, the
sample weight change, once more, can be mostly attributed to the reversibly adsorbed CO2. As can be
seen in Figure 5c, the amount of CO2 that is reversibly adsorbed in this region decreases as the temperature increases, consistent with the observations that
the crystallinity of the hydrotalcite material also decreases, and its structure begins to fall apart in this
region.1
To further validate the cyclic sorption/desoprtion
behavior observed using the TG/MB-MS, we also carried
cyclic flow sorption/desorption experiments in-situ using
the DRIFTS system, following the same experimental
protocol as with the TG/MB-MS experiments described
above. During these studies we monitored a number of
distinct peaks corresponding to various functional groups
in the hydrotalcite. During our prior study,1 the various
DRIFTS peaks were assigned as follows: (1) the signal
at ∼3470 cm-1 to the OH- group vibration in the MgAl-CO3 LDH sample; (2) the signal at ∼3070 cm-1 to
hydrogen bonding between water and the carbonate
species in the LDH interlayer space; (3) the signal at
∼1620 cm-1 to the H2O bending vibration of interlayer
water in the LDH sample; (4) the signals at ν3 ) 1370
cm-1, ν2 ) 940 cm-1, and ν4 ) 680 cm-1 at room
temperature to the CO32- group vibration bands in the
LDH sample. (The CO32- group in the hydrotalcite at
room temperature behaves more like it would in a
aqueous environment, in which the bands of CO32- are
observed at ν3 ) 1415 cm-1, ν2 ) 880 cm-1, and ν4 )
680 cm-1, without a ν1 mode vibration at ∼1080 cm-1
and no splitting of the ν3 band.)
In the prior experiments,1 starting from room temperature, the sample temperature was increased at a
rate of 0.5 °C/s in an inert atmosphere. Every 20 °C or
so the temperature increase was stopped, and the
DRIFTS spectra were recorded after keeping the sample
isothermal for a period of ∼2 min. The testing was continued until the temperature of the sample had reached
580 °C. The following summarizes the observed behavior:1
The intensities of the interlayer water bands at 3070
and 1620 cm-1 gradually decrease with increasing
temperature and disappear around 190 °C. The intensity of the OH- vibration band at 3470 cm-1 begins to
decrease at 190 °C and completely disappears at 440
°C. The band at 1370 cm-1 for the CO32- ν3 vibration
begins to decrease in size as the temperature increases
and also shifts to ∼1350 cm-1. Gradually a band at 1530
cm-1 begins to form at temperatures higher than 170
°C (the splitting of the ν3 band relates to the decrease of the amount of interlayer water, as a result
of which the CO32- group begins to interact more
strongly with the backbone of the hydrotalcite itself;
bands at 1530 and 1350 cm-1 have been previously
assigned to a monodentate carbonate species adsorbed
on Al2O330). The band size at the lower wavenumber
(∼1350 cm-1) decreases as the temperature increases
(and so are the peaks at 940 and 680 cm-1). At higher
temperatures all peaks corresponding to CO32- species
in Mg-Al-CO3 hydrotalcite disappear. Experiments
with the LDH2 sample, for which the temperature was
increased linearly (5 °C/min) in an inert atmosphere,
are shown in Figure 6a. The sequence of events and the
way the various bands evolve during the structural
evolution of the two LDH again remain the same; only
the range of temperatures where these different phenomena occur differs.
Figure 6b shows the integrated peak area reflecting
the CO32- ν3 vibration during the cyclic sorption/
desorption experiments (expressed as percent fraction
of the peak area at the beginning of the experiment
at room temperature). Figure 6c shows the 3470 cm-1
band corresponding to the OH- vibration and the
combined integrated peak areas for the interlayer
water peaks (3070 cm-1 and 1620 cm-1), again expressed as percent fraction of the same peak areas at
the beginning of the experiment at room temperature.
In these figures we show experimental data at the end
of first and second sorption and desorption cycles for
various temperatures. In Figure 6b,c one notes that at
150 °C only the combined integrated peak area for the
interlayer water peak changes in a reversible manner.
No substantial changes are observed during the sorp-
Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004 4565
Figure 6. (a) In-situ DRIFTS of LDH2 as a function of temperature and (b) change in the CO 32- integrated peak area (left) and
change in the OH- and H2O integrated peak areas as a percent fraction of the original peak area (right) during the sorption/desorption
cycles.
tion/desoprtion cycles in the integrated peak areas
reflecting the CO32- ν3 vibration or the 3470 cm-1 band
corresponding to the OH- vibration. The hydrotalcite
during the cyclic sorption/desorption experiments simply
exchanges reversibly only interlayer water. Above the
temperature of 190 °C, the interlayer water disappears
during the heating period in Ar before reaching the
preset temperature. During the cyclic sorption/desoprtion process the hydrotalcite exchanges reversibly
only CO2 to any substantial extent. Furthermore, the
amount of CO2 that reversibly adsorbs and desorbs
decreases as a function of temperature. Clearly, the
hydrotalcite under the conditions of the cyclic experiments described here is not capable of reversibly
exchanging the OH-, after a certain amount of OH- is
emitted during the heating period to reach the preset
temperature. These observations are consistent with the
TGA/MB-MS experiments and the observations of Yang
et al.1
We also carried out longer term cyclic sorption/
desorption experiments. For the experiments with the
LDH1, 10 mg of sample was utilized, and two different
temperatures 150 and 250 °C were investigated (the
same sample was used for both experiments). For each
experiment, the sample was first heated in UHP dry
Ar (20 sccm; 5 °C/min) and the weight of the sample
and the gas composition were monitored. Upon reaching
the desired temperature, the feed was switched to
humidified carbon dioxide (20 sccm, 70% RH) and kept
there for 3 h. Then the sample was evacuated for 1 h,
switched back to humidified carbon dioxide for 3 h, and
so on. Figure 7a shows the total sample weight (as
percent fraction of the original sample weight) observed
at the conclusion of the adsorption and desoprtion parts
4566 Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004
Figure 7. (a) Total sample weight during the sorption/desorption
cycles and (b) H2O and CO2 MS signals during the heating and
the desorption part of the cycles as a function of time.
Figure 8. (a) Total sample weight during the sorption/desorption
cycles and (b) H2O and CO2 MS signals during heating and the
desorption part of the cycles as a function of time.
of each cycle, for a total of 14 cycles. Figure 7b shows
the corresponding MS signals during the heating and
evacuation parts of the cycle. Only water was detected
coming out of the sample at 150 °C, an observation
consistent with the structural model for the LDH
presented by Yang et al.1 The experiments at 150 °C
indicate that the system reaches a steady-state reversible behavior after the 11th cycle, with the corresponding weight change being 0.23%.
Upon completion of the 14 cycles at 150 °C, the LDH1
sample was heated in UHP dry Ar (20 sccm; 5 °C/min)
until its temperature reached 250 °C; the feed was then
switched to humidified CO2 (20 sccm, 70% RH) and kept
there for 3 h. Then the sample was evacuated for 1 h,
switched back on to humidified CO2 for 3 h, and so on.
Figure 8a shows the total sample weight observed for a
total of 14 cycles for the LDH1 sample at 250 °C. The
sample reaches a steady-state behavior after the 9th
cycle, with the corresponding reversible weight change
being 0.31%. Figure 8b shows the corresponding MS
signals during the heating and evacuation parts of the
cycle. During the heating part of the experiment, both
water and CO2 are emitted; however, subsequently to
that only CO2 is emitted during the cyclic experiments,
indicating that under these relatively low RH conditions
the OH- are not reversibly exchanged to any substantial
extent (as noted previously, a similar observation also
made for the LDH2 sample during the shorter term
reversibility experiments). Once again, the observations
are consistent with the structural model previously
proposed.1
Figure 9 describes a long-term cycling experiment
using 113 mg of LDH2 at 250 °C, following the same
other experimental conditions as with the experiments
involving the LDH1 sample. The sample reaches a
steady-state behavior after the 14th cycle, with the
corresponding reversible weight change being 0.32%.
3.3. Sorption Reversibility Studies: Effect of
Thermal Cycling. After the cyclic sorption/desoprtion
experiments with the LDH samples were completed, a
number of experiments were initiated in which the
weight gain/loss of the sample was monitored as its
temperature was cycled from room temperature to a
preset temperature and back down to room temperature. In the first experiment (the same LDH1 sample
was used in all of temperature cycling experiments
reported in this section) the sample was heated in
flowing Ar (20 sccm) with a heating rate of 3 °C/min
from room temperature to a temperature of 150 °C;
subsequently the flowing Ar feed was substituted with
a humidified CO2 feed (20 sccm) and was cooled to room
temperature with a 3 °C/min cooling rate. Upon reaching room temperature, the sample was kept at this
temperature for an additional 2 h. The weight-loss/-gain
Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004 4567
Figure 11. MS signals for H2O (top) and CO2 (bottom) during
the temperature cycling experiment from room temperature to
200 °C.
Figure 9. Weight gain or loss (top) and total sample weight
(bottom) during the sorption/desorption cycles.
Figure 10. Weight loss/gain during the temperature cycling
experiments. Solid lines are the experiments from room temperature to 150 °C; dotted lines are experiments from room temperature to 200 °C.
data are shown in Figure 10. The total weight loss for
the LDH1 sample was ∼5.5% at 150 °C; upon cooling
in the humidified CO2 atmosphere, the sample recovered 98.4% of its original weight. The gaseous components evolved during the heating step were monitored
by mass spectrometry; only water is detected during the
experiment.
Upon termination of the experiment at 150 °C, the
humidified CO2 atmosphere was switched back to
flowing UHP dry argon and the temperature was slowly
(3 °C/min) increased to 200 °C. When this temperature
was reached, the same experimental protocol was followed. The weight-loss curve leveled off after 240
min. The sample, upon cooling, recovered 99.8% of its
original weight (Figure 10). The gases evolved during
the heating part of the cycle were also monitored (Figure 11). Water was detected throughout the whole
temperature range, and trace amounts of carbon dioxide were detected from 195 to 200 °C. Most of the
weight loss was observed below 150 °C; only ∼0.2%
of the weight loss was observed between 180 and
200 °C.
Upon termination of the experiment at 200 °C, the
humidified CO2 atmosphere was switched back to dry
argon, and the temperature was slowly (1 °C/min)
increased to 250 °C (in the 250 °C and higher temperature cycling experiments, the cooling/heating rates
were changed from 3 to 1 °C/min). When this temperature was reached, the flowing Ar feed was substituted
with a humidified CO2 feed (20 sccm) and the sample
was cooled to room temperature with a 1 °C/min cooling
rate. Also the sample was allowed to equilibrate at room
temperature for as long as necessary for the weightgain curve to level off. Upon cooling, 99.2% of original
4568 Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004
Figure 12. Weight loss/gain during the temperature cycling experiments: (a) from room temperature to 250 °C; (b) from room temperature
to 300 °C; (c) from room temperature to 350 °C.
weight of the sample was recovered; most of the weight
loss was again observed below 150 °C (Figure 12). The
components evolved during the heating part of the cycle
were also monitored (Figure 13). Water was detected
throughout the whole temperature range, similar to the
previous experiments, and smaller amounts of carbon
dioxide were detected in the range from 195 to 250 °C.
The experiment was repeated with the temperature
raised (1 °C/min) to 300 °C, cooled in humidified CO2
to room temperature with a 1 °C/min cooling rate, and
left there for as long as necessary for the weight-gain
curve to level off. As shown in Figure 12, it took a much
longer time than in the previous experiments for the
sample to regain the weight, which leveled off at 96%
of original weight. The composition of outlet gas showed
water being evolved through the whole range of temperatures; carbon dioxide was again detected between
195 and 300 °C. The experimental results with the
temperature raised to 350 °C are also shown in Figure
12. Once more, it took a much longer time for the sample
to recover its weight, which leveled off at 93.8% of its
original weight; water was evolved through the whole
region of temperatures, and carbon dioxide was detected
between 195 and 350 °C. These observations are clearly
consistent with the structural model for the LDH
developed, based on observations made under inert or
vacuum conditions.1
3.4. Sorption Reversibility Studies at Moderate
Pressures in a Flow System. To further study the
sorption reversibility behavior of the LDH, and to
establish a connection between the TG/MB-MS and
DRIFTS experiments carried out at atmospheric pressure conditions and the conditions one may encounter
in the use of these materials in the WGS membrane
reactor environment, we carried out similar experiments
using a high-pressure adsorption flow system. The
experimental system is equipped with mass flow controllers and a flow control valve at the exit to maintain
the system pressure constant under flow conditions.
Two types of experiments were performed. In the first
series of experiments the flow system was first pressurized with flowing dry argon (50 sccm) to 50 psig, then
the temperature was increased to 250 °C, using a 5 °C/
min heating rate (it takes ∼45 min), and the system
was kept at 250 °C for 1 h as a desorption step. When
the desorption step was over, the system was cooled to
150 °C (cooling rate 5 °C/min, ∼20 min) in flowing dry
argon (50 sccm). Subsequently, the inlet gas was
changed to dry CO2 (50 sccm) from argon while keeping
the same pressure of 50 psig for 3 h as an adsorption
step. During the adsorption step, the outlet flow rate
was monitored by a digital flow meter (while the reactor
pressure was maintained constant at 50 psig), and from
the flow rate change, the amount of adsorption was
Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004 4569
Table 4. Weight Gain (wt %) during the Sorption Step
for the Moderate-Pressure Flow Experiments at Various
Temperatures Using Dry CO2
cycle
wt gain during
the sorptn step
1st
2nd
2.02
1.97
At 150 °C
3rd
4th
1.87
1.84
1st
2nd
1.84
1.79
At 200 °C
3rd
4th
1.71
1.69
1st
2nd
1.72
1.69
At 250 °C
3rd
4th
1.68
1.66
cycle
wt gain during
the sorptn step
Table 5. Weight Gain (wt %) during the Sorption Step
for the Moderate-Pressure Flow Experiments at 200 °C
Using Humidified CO2
Figure 13. MS signals for H2O (top) and CO2 (bottom) during
the temperature cycling experiment from room temperature to
250 °C.
Table 3. Weight Gain (wt %) during the Sorption Step
for the Moderate-Pressure Flow Experiments Using Dry
CO2
cycle
wt gain during
the sorptn step
1st
2nd
2.621
2.542
cycle
wt gain during
the sorptn step
3rd
2.476
calculated. Table 3 shows the weight gain during the
sorption step for the first 3 sorption/desoprtion cycles
with the LDH2 sample.
In the second series of experiments the sample (∼14
g) was first heated to a preset temperature in Ar gas
(50 sccm). Each cycle involved first evacuating the
sample for 1 h as a desorption step. After the evacuation
step, the flow system was then again pressurized to 50
psig in flowing Ar. When the outlet flow rate was
stabilized in flowing Ar at 50 psig, the inlet gas was
then changed to either dry or humidified CO2 (50 sccm)
for 3 h while maintaining the same pressure of 50 psig.
Upon completion of the first sorption/desoprtion cycle,
the procedure was repeated for a number of additional
cycles and for a number of temperatures. The cyclic
sorption/desoprtion experimental results for the various
temperatures for dry CO2 are summarized in Table 4,
while those for humidified CO2 for one temperature are
shown in Table 5 (a fresh LDH sample was utilized for
each set experiments at every new temperature). During
the humidified CO2 experiments at moderate pressures,
a syringe pump was used (instead of a bubbler) to
cycle
wt gain during
the sorptn step
cycle
wt gain during
the sorptn step
1st
2nd
1.85
1.79
3rd
4th
1.73
1.71
deliver a predetermined amount of water into the CO2
stream. For the experiments in Table 5, the CO2 stream
contains 2 mol % of water, which corresponds approximately to 70% RH at the temperature and pressure
of the experiment. For the run at 150 °C the weight lost
during the desoprtion part corresponds to water with
only traces of CO2 being lost. The opposite observation
is true for the runs at 200 and 250 °C. Comparing the
run in humidified CO2 (Table 5) with the corresponding
run with dry CO2, one observes that there is little
difference in the weight change, which is consistent with
the observation under atmospheric conditions that, beyond the initial heating step, the LDH does not exchange water reversibly under these temperature conditions. The whole set of observations under moderate
pressure conditions are again in agreement with the
TG/MB-MS and DRIFTS data under atmospheric conditions and the LDH structural model.1 This lends confidence in the usefulness of these fundamental techniques in predicting the materials behavior under
realistic process conditions.
4. Conclusions
In our prior study,1 we have employed several in-situ
techniques to investigate the thermal evolution of the
structure of a Mg-Al-CO3 LDH under an inert atmosphere. On the basis of the results of that study, a model
was proposed to describe the structural evolution of the
Mg-Al-CO3 LDH structure. In this paper we have
presented results of our ongoing investigations with
these materials pertaining to their sorption characteristics and thermal reversibility under both inert and
reactive atmospheres. The experimental observations
are shown to be consistent with the structural model
for the LDH previously presented. The LDH are shown
capable of exchanging reversibly CO2 for a broad region
of conditions. In addition to providing a test of the
validity of the LDH thermal evolution model, these
experimental observations are of importance in their
own right for the use of the LDH materials in the
preparation of CO2 permselective membranes and adsorbents for high-temperature membrane reactor applications. One particular promising application is their
use in reactive separations with the water gas shift
reaction. The ability of the LDH to reversibly adsorb
4570 Ind. Eng. Chem. Res., Vol. 43, No. 16, 2004
CO2 is of significance in the use of these materials in
such an application as adsorbents and membranes. In
the former case, the ability to reversibly adsorb CO2 is
critical from the standpoint of being able to regenerate
the adsorbents; in the latter case the presence of a
relatively mobile CO2 phase within the LDH structure
is important in determining the permeation rate through
the membrane.
Acknowledgment
The support of the U.S. Department of Energy is
gratefully acknowledged.
Note Added after ASAP Posting
This article was released ASAP on 3/24/04 with the
x-axis labels in Figure 1 shifted slightly to the right.
The correct version was posted on 7/22/04.
Literature Cited
(1) Yang, W.; Kim, Y.; Liu, P. K. T.; Sahimi, M.; Tsotsis, T. T.
A Study by in-situ techniques of the thermal evolution of the
structure of a Mg-Al-CO3 layered double Hydroxide (LDH).
Chem. Eng. Sci. 2002, 57, 2945.
(2) Rives, V.; Ulibarri, M. A. Layered double hydroxides (LDH)
intercalated with metal coordination compounds and oxometalates.
Coord. Chem. Rev. 1999, 181 (1), 61.
(3) Roy, A.; de Forano, C.; Malki, K. E.; Besse, J. P. Anionic
clays: Trends in pillaring chemistry. Expanded Clays and Other
Microporous Solids; Van Nostrand Reinhold: New York, 1992;
Chapter 7.
(4) Tsuji, M.; Mao, G.; Yoshida, T.; Tamaura, Y. Hydrotalcites
with an extended Al3+ substitution: Synthesis, simultaneous TGTDA-MS study, and their CO2 adsorption behaviors. J. Mater. Res.
1993, 8 (5), 1137.
(5) Goswamee, R. L.; Sengupta, P.; Bhattacharyya, K. G.;
Dutta, D. K. Adsorption of Cr(VI) in layered double hydroxides.
Appl. Clay Sci. 1998, 13 (1), 21.
(6) Pavan, P. C.; Gomes, G. de A.; Valim, J. B. Adsorption of
sodium dodecyl sulfate on layered double hydroxides. Microporous
Mesoporous Mater. 1998, 21 (4-6), 659.
(7) Shin, H. S.; Kim, M. J.; Nam, S. Y.; Moon, H. C. Phosphorus
removal by hydrotalcite-like compounds (HTLCs). Water Sci.
Technol. 1996, 34 (1-2), 161.
(8) Ding, Y.; Alpay, E. Equilibria and kinetics of CO2 adsorption
on hydrotalcite adsorbent. Chem. Eng. Sci. 2000, 55, 3461.
(9) Yong, Z.; Mata V.; Rodrigues, A. E. Adsorption of carbon
dioxide onto hydrotalcite-like compounds (HTLCs) at high temperatures. Ind. Eng. Chem. Res. 2001, 40, 204.
(10) Alejandre, A.; Medina, F.; Rodriguez, X.; Salagre, P.;
Cesteros, Y.; Sueiras, J. E. Cu/Ni/Al layered double hydroxides as
precursors of catalysts for the wet air oxidation of phenol aqueous
solutions. Appl. Catal., A 2001, 30 (1-2), 195.
(11) Bahranowski, K.; Bueno, G.; Corberan, V. C.; Kooli, F.;
Serwicka, E. M.; Valenzuela R. X.; Wcislo, K. Oxidative dehydrogenation of propane over calcined vanadate-exchanged Mg,
Al-layered double hydroxides. Appl. Catal., A 1999, 185 (1), 65.
(12) Pinnavaia, T. J.; Gardner, E. On the nature of selective
olefin oxidation catalysts derived from molybdate- and tungstateintercalated layered double hydroxides. Appl. Catal., A 1998, 167
(1), 65.
(13) Auer, S. M.; Grunwaldt, J. D.; Koppel, R. A.; Baiker, A.
Reduction of 4-nitrotoluene over Fe-Mg-Al lamellar double
hydroxides. J. Mol. Catal., A 1999, 139 (2-3), 305.
(14) Malherbe, F.; Depege, C.; Forano, C.; Besse, J. P.; Atkins,
M. P.; Sharma, B.; Wade S. R. Alkoxylation reaction catalyzed by
layered double hydroxides. Appl. Clay Sci. 1998, 13 (5-6), 451.
(15) Santhanalakshmi, J.; Raja, T. Selective N-methylation of
aniline by calcined MgII-AlIII layered double hydroxides. Appl.
Catal., A 1996, 147 (1), 69.
(16) Tichit, D.; Bennani, M. N.; Figueras, F.; Tessier, R.;
Kervennal, J. Aldol condensation of acetone over layered double
hydroxides of the meixnerite type. Appl. Clay Sci. 1998, 13 (5-6),
401.
(17) Ding, Y.; Alpay, E. Adsorption-enhanced steam-methane
reforming. Chem. Eng. Sci. 2000, 55, 3929.
(18) Hufton, J. R.; Mayorga, S.; Sirkar S. Sorption-enhanced
reaction process for hydrogen production. AIChE J. 1999, 45, 248.
(19) Kim, Y.; Yang, W.; Liu, P. K. T.; Sahimi, M.; Tsotsis, T. T.
Layered double hydroxide (LDH) materials for CO2 adsorption and
separation. Presented at A&WMA’s 96th Annual Conference &
Exhibition, San Diego, CA, June 22-26, 2003.
(20) Hibino, T.; Yamashita, Y.; Kosuge, K.; Tsunashima, A.
Decarbonation behavior of Mg-Al-CO3 hydrotalcite-like compounds during heat treatment. Clays Clay Miner. 1995, 43 (4),
427.
(21) Kanezaki, E. Effect of atomic ratio Mg/Al in layers of Mg
and Al layered double hydroxide on thermal stability of hydrotalcite-like layered structure by means of in-situ high-temperature
powder X-ray diffraction. Mater. Res. Bull. 1998, 33 (5), 773.
(22) Kanezaki, E. Thermal behavior of the hydrotalcite-like
layered structure of Mg and Al-layered double hydroxides with
interlayer carbonate by means of in-situ powder HTXRD and DTA/
TG. Solid State Ionics 1998, 106 (3-4), 279.
(23) Kanezaki, E. Thermally induced metastable solid phase
of Mg/Al layered double hydroxides by means of in-situ hightemperature powder X-ray diffraction. J. Mater. Sci. Lett. 1998,
17 (5), 371.
(24) Kloprogge, J. T.; Frost, R. L. Fourier transform infrared
and Raman spectroscopic study of the local structure of Mg-, Ni-,
and Co-hydrotalcites. J. Solid State Chem. 1999, 146, 506.
(25) Kloprogge, J. T.; Frost, R. L. Infrared emission spectroscopic study of the dehydroxylation of synthetic Mg/Al and
Mg/Zn/Al-hydrotalcites. Phys. Chem. Chem. Phys. 1999, 1, 1641.
(26) Kloprogge, J. T.; Frost, R. L. Infrared emission spectroscopic study of the thermal transformation of Mg-, Ni-, and
Co-hydrotalcite catalysts. Appl. Catal., A 1999, 184, 61.
(27) Cavini, F.; Trifiro, E.; Vaccari, A. Hydrotalcite-type anionic
clays-preparation, properties and applications. Catal. Today 1991,
11, 173.
(28) Constantino, V. R. L.; Pinnavaia, T. J. Basic properties of
Mg1-x2+Alx3+ layered double hydroxides intercalated by carbonate,
hydroxide, chloride, and sulfate anions. Inorg. Chem. 1995, 34,
883.
(29) Rhee, S. W.; Kang M. J. Kinetics on dehydration reaction
during thermal treatment of Al-CO3-LDHs. Korean J. Chem. Eng.
2002, 19, 653.
(30) Baumgarten, E.; Zachos, A. Infrared spectroscopic investigations on the adsorption of CO2 on aluminas and silica aluminas
at elevated temperatures. Spectrochim. Acta 1981, 37A, 93.
Received for review November 5, 2003
Revised manuscript received February 6, 2004
Accepted February 10, 2004
IE0308036