Available online at www.sciencedirect.

com

Applied Clay Science 40 (2008) 27 – 37

www.elsevier.com/locate/clay

Influence of M II /M III ratio in surface-charging behavior of

Zn–Al layered double hydroxides
R. Rojas Delgado a,1 , C.P. De Pauli a,⁎, C. Barriga Carrasco b,2 , M.J. Avena c,3
a
INFQC, Departamento de Fisicoquímica, Facultad de Ciencias Químicas, Universidad Nacional de Córdoba,
Ciudad Universitaria, 5000 Córdoba, Argentina
b
Departamento de Química Inorgánica e Ingeniería Química, Campus de Rabanales, Universidad de Córdoba,
Edificio C-3, 14071, Córdoba, Spain
c
Departamento de Química, Universidad Nacional del Sur, Avenida Alem 1253, 8000 Bahía Blanca, Argentina

Received 31 January 2007; received in revised form 20 June 2007; accepted 22 June 2007
Available online 10 July 2007

Abstract

Zn–Al layered double hydroxides with different MII/MIII ratios were synthesized by the coprecipitation method and the
obtained solids were characterized by powder X-ray diffraction, infrared spectroscopy, acid–base potentiometric titrations,
electrophoretic mobility and modeling of the electrical double layer. As general behavior, increasing hydroxyl adsorption with
increasing pH and decreasing supporting electrolyte concentration takes place between pH 7 and 12. The electrophoretic mobility
of the samples are in all cases positive and it decreases when the pH becomes higher than 9. A net positive charge of the particles,
with a point of zero charge (PZC) around 11.8, is deduced from electrophoretic data. A modified MUSIC model, where OH−/Cl−
exchange at structural sites and proton desorption from surface hydroxyl groups are taken into account, is used to estimate the
deprotonation constants, the number of surface groups, the number of sites with structural charges an the affinity constants for Cl−
and OH− of the samples with different MII/MIII ratios. The modeling of potentiometric titrations and electrophoresis data indicates
that anions between the layers and negatively charged groups in the particle surface balance most of the structural charge at the
particles surface. It shows an important variation of the number of structural sites with the MII/MIII ratio and an important role of
surface protonation–deprotonation sites to the electrical behavior and exchange properties of LDHs.
© 2007 Elsevier B.V. All rights reserved.

Keywords: Layered double hydroxides; Acid–base potentiometric titrations; Electrophoretic mobilities; Electric double layer

1. Introduction catalysts, exchangers, adsorbents, etc. These materials can

be classified by the charge excess in the layers of the host
Layered solids with intercalation properties cover a material (Wells, 1984) into three groups: solids with
wide variety of materials with important applications as neutral layers, such as graphite; solids with negatively
charged layers, such as phyllosilicate clays (Sposito,
⁎ Corresponding author. Fax: +54 351 4334188. 1984) and solids with positively charged layers, repre-
E-mail addresses: rrojas@mail.fcq.unc.edu.ar (R. Rojas Delgado), sented by layered double hydroxides (LDHs) or hydro-
depauli@mail.fcq.unc.edu.ar (C.P. De Pauli), iq1bacac@uco.es talcites (Bruce and O'Hare, 1997; Newman and Jones,
(C.B. Carrasco), mavena@uns.edu.ar (M.J. Avena).
1
Fax: +54 351 4334188.
1999; Rives, 2001).
2
Fax: +34 957 418 648. In solid phase the positively charged layers of LDHs
3
Fax: +54 291 4595160. are balanced by anions, which can be inorganic (from
0169-1317/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.clay.2007.06.010
28 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37

chloride or nitrate (Khan and O'Hare, 2002) to poliox- An− represent the divalent cation, the trivalent cation and
ometalates (Ulibarri et al., 1994)), coordination com- the anion that neutralizes the positive structural charge,
pounds (Rives and Ulibarri, 1999), organic anions respectively. One positive charge per substituting trivalent
(Carlino, 1997) or even biomolecules (Choy et al., cation is introduced in the structure, so MII/MIII will affect
2000). Thus, interlayer ions are responsible for the solid the exchange capabilities of the intercalation compounds
electroneutrality in solid phase, but when immersed in and will determine their surface-charging behavior.
aqueous solutions some of the interlayer anions move to In this work we study the influence of MII/MIII ratio
the aqueous solution and the solid develops an electric in the charging properties of LDHs. With this aim we
double layer at the surface, allowing exchange with ions studied the proton adsorption and electrokinetic behav-
from the solution. ior of Zn–Al LDHs with different Zn2+/Al3+ ratios at
Not only electrostatic forces (which rule ion exchange) different pHs and electrolyte concentrations. A model
are involved in the processes that take place when layered including anionic exchange and protonation–deproto-
solids are immersed in an aqueous solution. The surface of nation processes at the surface of the particle was
these solids may establish chemical bonds (sometimes applied to these data in order to explain them.
called specific bonds) with species in the aqueous phase.
For example, the proton and ion binding properties of 2. Experimental
groups located at the edges of phyllosilicate layers are
especially reactive (Baeyens and Bradbury, 1999; Avena, 2.1. Samples preparation
2002). The charging behavior of phyllosilicate particles
The samples were prepared by coprecipitation at controlled
can be theoretically described by assuming that the
pH (Cavani et al., 1991), as described previously for Zn and Al
particles carry two kind of electrical charges: “permanent” containing LDHs (Badreddine et al., 1998; Ennadi et al.,
structural charges, that are the main responsible for the 2000). 200 ml-mixtures of ZnCl2 and AlCl3 (both Anedra)
cation exchange properties and “variable” (pH dependent) aqueous solutions, with different [Zn2+]/[Al3+] = R = 2.4, 3.6
charges resulting from proton adsorption–desorption and 4.8 ratios and the same [Zn2+] + [Al3+] = 0.1 M were added
reactions at the edge groups. Whereas the structural drop wise in a flask containing a vigorously stirred 0.1 M NaCl
charge is negative, the variable charge can be positive, (Merck) solution. The reaction pH was set constant to 7 by the
neutral or negative depending of the degree of protonation simultaneous addition of a KOH solution. Once the addition of
of the edge surface. Since the amount of edge groups is reactants was finished, the resulting slurry was left under
usually lower than the amount of permanent charges these stirring for 72 h. The precipitate was then separated from the
supernatant by centrifugation and then washed several times
solids carry a net negative charge. Recent theoretical
with water. The wet solid obtained was dispersed in water and
studies have succeeded in explaining proton adsorption–
kept at pH 7 as a stock suspension until it was used. The
desorption processes, electrokinetic data, ion (others than samples will be named ZnAl2, ZnAl3, ZnAl4 according to the
proton) adsorption measurements and electrical double different [Zn2+]/[Al3+] = R ratio in the starting solutions.
layer properties of phyllosilicates (Kraepiel et al., 1998; All solutions were prepared with purified water (Milli-Q
Avena and De Pauli, 1998). system) boiled and purged with N2. All samples were prepared
The charging and binding properties of surface groups at room temperature and under a stream of N2 in order to avoid,
of LDHs were poorly studied, although some reports or at least minimize, the contamination by atmospheric CO2.
about chemical bonds with protons and ions exist in the
literature (How et al., 2001; Li et al. 2003). In a previous 2.2. Experimental techniques
work (Rojas Delgado et al., 2004) a model assuming the
presence of these two types of charges in a Zn–Cr LDH 2.2.1. Structural characterization
was used and the charging behavior of this solid was Elemental chemical analysis was carried out in an AA-3100
successfully explained. In contrast to the behavior instrument from Perkin Elmer; samples were dissolved in
observed with phyllosilicates, the development of charges minimum necessary volume of HCl 35% (w/w) and then diluted
to the appropriate concentration. To obtain the water content of
at variable charge groups in LDHs seems to be important
the LDHs samples, an aliquot of the stock suspension was
enough to produce a significant decrease in the zeta
centrifuged and dried at 70 °C, weighted and then heated at
potential and to reach a PZC at high pH values. 150 °C to quantify the weight loss between both temperatures.
The crystal structure of LDHs is based on a stacking Powder X-Ray diffraction (PXRD) patterns of a previously dried
of brucite-like layers where a fraction x of divalent sample were recorded in a Rigaku Miniflex instrument, using a
cations is substituted by trivalent cations (De Roy et al., Cu-kα (λ = 1.5418 Å). FT-IR spectra were recorded following

1992). The general n
formula of these compounds is the KBr pellet technique (1% weight sample in KBr) in a FT-IR
x  ðOHÞ2 Ax=n  mH2 O, where M , M
II III
MII1
x MIII and Bruker IFS28 instrument.
 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37 29

2.2.2. Surface studies

In order to measure the proton adsorption–desorption of
LDHs, series of acid–base potentiometric titrations at different
salt concentrations were performed for every LDH sample.
The method is similar to that used for metal oxides and clay
minerals (Avena, 2002) and for LDHs in a previous work
(Rojas Delgado et al., 2004). A 60 ml dispersion containing 1 g
of the corresponding LDH was prepared by diluting the stock
dispersion with a NaCl solution of the desired concentration.
The suspension was allowed to equilibrate for 1 hour under
stirring and N2 bubbling, and then titrated with a standard
KOH solution. The titrant volume added was set in every
addition to achieve a pH variation lower than 0.2 units. After
each titrant addition, the slurry was stirred until equilibration,
the readings being accepted when the drift was less than Fig. 1. PXRD diagram of ZnAl2 sample.
0.2 mV/min. The titration was stopped when the pH was
around 12, and then, to check reversibility, the same BN 9101 glass electrode and an Orion 900200 double junction
suspension was back-titrated with a standard HCl solution. Ag/AgCl/Cl reference electrode.
No data could be obtained at higher pH values because of the
extremely high buffer capacity of aqueous solutions under
these conditions, which results in unreliable blank corrections. 3. Results
The proton and hydroxyl adsorption were calculated as the
difference between total amounts of H+ or OH− added to the 3.1. Structural characterization
dispersion and those required to bring a blank solution of the
same volume and electrolyte concentration to the same pH. The chemical compositions of the different LDHs are
This procedure was performed for all samples studied at three included in Table 1, which shows that the [Zn2+]/[Al3+]
different NaCl concentrations. The measurements were done at ratio in the solids is close to that in the starting solutions.
a constant temperature (30 °C) and with N2 bubbling to avoid The formula in the table considers that chloride is the
CO2 contamination.
only compensating anion and does not take into account
Electrophoretic mobilities were measured using a Rank
Brothers electrophoresis apparatus, equipped with a 2 mm-
the possible presence of carbonate impurities in the
cylindrical cell. An aliquot of the corresponding stock interlayer space.
suspension was added to 200 ml of a 0.01 M NaCl solution The PXRD patterns were quite similar for the three
and was ultrasonically dispersed. The pH of this dispersion compounds prepared. Fig. 1 shows the pattern for the
was decreased to approximately 6 with an HCl solution and sample ZnAl2, as an example. The patterns were indexed
afterwards increased adding NaOH until pH = 11.5 in 0.5–1 in a rhombohedral lattice and reveal the layered structure
pH-units steps. In every step the suspension was stirred for of these solids. The cell parameters of the three samples
5 min and the pH and μ were then measured. All dispersions are included in Table 2. The c parameter (which represents
were prepared and used in a N2 atmosphere to avoid the thickness of three layers plus the interlayer space
contamination with carbon dioxide. Electrophoretic mobilities between them) was calculated from the average of (00l)
were converted to zeta potentials (ζ) with the Smoluchowski
reflections. The c parameter decreases as the MII/MIII
equation in order to use these data in the surface-charging
model, but it must be considered that ζ values obtained by this
ratio decreases, what has been assigned (Rives, 2001) to
equation are only estimations as it can present deviations at stronger interactions between the layers and the interlayer
high ionic strength for this kind of particles. anions. The a dimension (which represents the shortest
In all cases, titrants additions were done with a 665 distance between two cations of the layer) was calculated
Dosimat (Metrohm) automatic dispenser and pH was mea- as twice the position of the (110) reflections. This
sured with an Orion EA 940 pH meter equipped with an Orion parameter decreases as the Al3+ content increases, as

Table 1
Elemental chemical analysis data and interlayer water content of Zn,Al-LDH samples and their calculated formulae
Sample MII/MIIIteor % Zn % Al MII/MIIIexp % H2O Chemical formulae
ZnAl2 2.4 43.0 7.7 2.3 6.8 Zn0.70Al0.30(OH)2Cl0.3·0.4H2O
ZnAl3 3.6 48.0 5.6 3.5 6.5 Zn0.78Al0.22(OH)2·Cl0.22·0.38H2O
ZnAl4 4.8 50.7 4.3 4.9 7.1 Zn0.83Al0.17(OH)2·Cl0.17·0.42H2O
30 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37

Table 2 tration is obtained at pH around 11 and at higher pH

Crystallographic parameters, molecular weight, total layer specific area values the curves approach each other.
and calculated density of structural charges of Zn,Al-LDH samples
Zeta potential vs. pH curves are also shown in Fig. 3.
Sample a(Å) c(Å) M(g) S(m2/g) FNstr (C/m2) All samples show a positive ζ in the studied pH range,
ZnAl2 3.06 22.8 105.7 924 0.30 although the decrease of ζ as the pH increases suggests
ZnAl3 3.08 23.2 105.6 936 0.21 that the isoelectric point should be at pH 12 or somewhat
ZnAl4 3.10 23.3 106.5 941 0.16
higher. However, no reliable measurements could be
performed at these high pH values because of the
increase of the ionic strength under these conditions.
observed in other LDHs (López-Salinas et al., 1997), The general behavior presented in Fig. 3 was
because this cation has a lower ionic radius than Zn2+. previously observed for a Zn–Cr LDH (Rojas Delgado
Table 2 contains values of total layer specific area, S, et al., 2004). Besides this general behavior, some
calculated (S. P. Li et al., 2003) from lattice parameters and interesting differences can be found comparing the
chemical composition with the equation S = 31/2a2N /M data of the three samples studied here. The maximum
where a is the lattice parameter, N is the Avogadro's hydroxyl adsorption increases with increasing MII/MIII
constant and M is the chemical formula weight. The total ratio in the samples, from almost 0.15 C/m2 for ZnAl4 to
specific area of the layers of all samples is similar. N2 around 0.30 C/m2 for ZnAl2, which can be related to the
adsorption was not used to measure the specific surface larger anion exchange capacity of the sample with
area because in layered solids as phyllosilicate and LDHs higher Al3+ content. On the other hand, ζ vs. pH curves
N2 cannot enter the interlayer space, so it cannot give a show little variation when structural charge is varied,
measure of the total area of the layers. indicating that zeta potentials and isoelectric points are
The FT-IR spectra of the three Zn–Al–Cl LDH rather independent on the density of structural charges.
samples were also similar. The spectrum of ZnAl2 is
given in Fig. 2 as an example. They present a very intense 4. Discussion
band around 3440 cm− 1, which is due to the stretching
vibration of hydroxyl groups of the sheets and water The crystal structure of LDHs is based on M(OH)6
molecules of the interlayer domain. The deformation octahedral units sharing edges, building a sequence of
vibration of water molecules is responsible for the broad brucite-like layers where isomorphic substitution of a
band recorded at 1618 cm− 1. Lattice vibrations appear in fraction x of divalent cations by trivalent cations occurs.
the 414–613 cm− 1 range. The weak band at 1354 cm− 1 The layers become therefore charged, as one positive
indicates the presence of small amounts of carbonate in charge per substituting trivalent cation is introduced into
the galleries. The spectra agree with the more exhaustive the structure. Therefore, the structural charge density
description given by Kloprogge et al. (2004). results to be proportional to the trivalent metal ratio
x = MIII/(MII + MIII). In the case of Zn–Al LDHs Zn(OH)6
3.2. Surface studies and Al(OH)6 octahedra are arranged in the layers on a
hexagonal structure of parameter a0. For particular values
Potentiometric titration data for Zn–Al LDHs sam-
ples are shown in Fig. 3a, b and c as hydroxyl
consumption (σOH) vs pH curves. Symbols represent
experimental data and lines represent model predictions
(see Modeling section). Titration data include points
obtained by titration from low pH to high pH with a KOH
solution together with points obtained by titration from
high pH to low pH with a HCl solution. The coincidence
of data for a given NaCl concentration indicates good
reversibility in the titration curves. Fig. 3 shows that all
samples have a net OH− consumption in the pH range
measured. OH− consumption starts at around pH = 7 and
increases as pH increases. It can be also observed that at a
given pH an increase in the electrolyte concentration
produces a decrease in OH− consumption. The greatest
separation of the curves at different electrolyte concen- Fig. 2. FT-IR spectra of ZnAl2 sample.
 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37 31

Fig. 3. Experimental (symbols) and calculated (lines) σOH vs. pH and ζ vs pH curves. Circles and solid lines: 0.01 M NaCl; triangles and dashed
lines: 0.1 M NaCl; 1 M NaCl. (a and b) ZnAl2 sample; (c and d) ZnAl3 sample; (e and f) ZnAl4 sample. The ionic strength in the calculations was
obtained from NaCl concentrations using the Davies equation.

of x, ordering of divalent and trivalent cations can form In this work, integer values of this relation have been
superstructures, although the existence of such super- avoided in order to stay away from superstructures, so
structures has been demonstrated in only a few cases. we can suppose a random distribution of the metallic
LDHs can be synthesized with MII /M III ratios cations in the layers. Since neither of the samples has
varying from 2 to 4 and even more. The upper limit is superstructures it can be supposed that all of them have a
generally attributed to electrostatic repulsion between similar layer structure. Differences can be found
adjacent trivalent metals in the layers. Too high values between the LDHs when the composition of their
of MII/MIII would result in collapse of the interlamellar interlayer space is considered. Depending on the
domain as they are less populated by charge compen- synthesis conditions and the Zn/Al ratio they can differ
sating anions. For Zn–Al-LDHs it has been described in the water content and/or the type and concentration of
that at neutral pH, LDH phases can be obtained with exchangeable anions that neutralize the positive charge.
MII/MIII varying from 5 to 1. It is well known that chloride is easily exchanged by
32 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37

and XOH− groups. OH− and Cl− represents in the

equation the anionic species in the solution bulk. It is a
simplification of the actual procedure of structural
charge balancing as balancing anions are situated in
the particle surface as well as in the bulk of the solid,
placed between the layers. Anyway, the process clearly
indicates that increasing the pH value or decreasing the
chloride anion concentration increases OH− consump-
tion. This agrees, at least qualitatively, with trends
shown in Fig. 3.
The other OH− consuming process is given by the
reactivity of surface hydroxyls attached to metal ions in
the layers. In Zn–Al-LDHs, hydroxyls are bonded to
three cations of the same layer, so we can find two types
of surface hydroxyl groups: OH− attached to two Zn2+
and one Al3+ cations (Zn2AlOH groups) and others
linked only to Zn2+ cations (Zn3OH groups). Hydroxyl
groups linked to more than one Al3+ cations would be
forbidden by the electrical repulsions between adjacent
Al3+ cations. If these hydroxyls are in contact with the
Fig. 4. Schematic drawing of the LDH/solution interface according to aqueous solution, their constituting oxygen atoms are
the proposed model. potentially reactive and can be protonated or deproto-
nated. The OH− consumption by the groups can be
other anions when dispersed in aqueous solution. written in a general formula as
Carbonate anions, on the contrary, are strongly attached
to brucite-like layers and they only can be exchanged at SOH þ OH− ¼ SO− þ H2 O ð2Þ
very low pHs (Ulibarri et al., 1995). Then, avoiding the where SOH and SO − represents protonated and
presence of carbonate anions during synthesis and deprotonated surface groups (Zn3OH or Zn2AlOH
characterization is a fundamental issue to study the groups). As pH increases this reaction will be favored,
surface-charging behavior of LDHs. increasing OH− consumption as shown in Fig. 3.
PXRD and FT_IR data in Figs. 1 and 2 are typical of
LDHs and similar to the reported in the literature for
pure chloride intercalated Zn–Al-LDHs phases 4.1. Modeling
(Badreddine et al., 1998; Legrouri et al. 2005). These
facts, together with the chemical analyses and water According to the two reactions proposed above, the
content measurements, allow us to propose chemical overall OH− consumption results from two processes:
formulae for the prepared samples (Table 1). It has been Cl−/OH− anionic exchange (Eq. (1)) and deprotonation
considered that the interlayer space is occupied only by of surface hydroxyl groups (Eq. (2)). A model of the
chloride anions, which where present at relatively high LDHs-water interface can separate the contribution of
concentration during the synthesis of the solids. both processes. In a previous work we have found that
Fig. 3 shows that LDHs consume a considerable both processes are equally important in order to explain
quantity of hydroxyl anions as pH is increased. OH− the surface-charging behavior of LDHs, here we will
consumption by Zn–Al-LDHs will be considered to take study the effect of MII/MIII ratio in both anionic
place by two mechanisms: one of them is anionic exchange and protonation–deprotonation of surface
exchange of chloride anions in the interlayer by hydroxyl groups.
hydroxyl anions of the solution. This process can be The proposed model can be divided into two parts:
written as: one representing the surface of the LDH particle, the
type and properties of surface sites, the charge
XCl− þ OH− ¼ XOH− þ Cl− ð1Þ distribution in the particle, etc., and other representing
the electrical double layer (EDL) that describes the
where X represents groups located in the surface, charge distribution and potential decay on the aqueous
capable of attaching Cl− or OH− anions, forming XCl− side of the interface.
 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37 33

The scheme of a LDH layer particle according to the surfaces densities (mol/m2) of SOH and SO− groups,
proposed model is shown in Fig. 4. Structural charges respectively, ψ0 is the potential at the surface of the
are assumed to be internal and placed parallel to the particle and F, R and T have their usual meaning. The
surface of the particle. The charge density in this plane is surface density of protonation–deprotonation groups
designated as σstr (C/m2) and the density of structural will be denoted Nv (Nv = ΓSOH + ΓSO).
charges (mol/m2) in the internal plane as Nstr. The It must be noted that close to the SOH groups exists
relation between them is an excess of positive charge due to the presence of Al3+.
The positive charge is not explicitly written in the
rstr ¼ FNstr ð3Þ groups as it is considered as a structural charge forming
Although the exchange sites are placed in the surface part of the lattice. The presence of structural charges in
and also in the bulk of the particle, in this model they are the surrounding area of the group may increase the
considered to express as discrete sites X that can bind affinity for deprotonation of surface groups because H+
and exchange the anions of the aqueous solution (Cl− is repelled by the positive structural charges. In Zn3OH
and OH−), thus compensating the positive charge excess groups we will suppose that the existence of Al3+
of the particle. In other words, Nstr will have three cations in the lattice does not affect the reactivity of the
components: NX, NXCl and NXOH, that stand for the group, as it is not directly coordinated to the central
density of unscreened charges and of those balanced by OH− , so these groups will have less affinity for
chloride and hydroxyl anions, respectively. deprotonation than Zn2AlOH ones, as will be discussed
The adsorption reactions of Cl− and OH− on X sites later.
can be written as follows The charge density and the electric potential at the
surface plane are denoted as σ0 and ψ0, respectively.
X þ Cl− ¼ XCl− ð4Þ The charge density at the surface plane can be calculated
from the densities of charged sites at this plane; thus
X þ OH− ¼ XOH− ð5Þ
r0 ¼ F ð
CXCl
CXOH
CSO Þ ð10Þ
The intrinsic constants for these processes are
CXCl It must be considered that the net charge of the
int
KCl ¼ expð
Fw0 =RT Þ ð6Þ
CX ½Cl
 particle will be given by the sum of positive structural
charges inside the particle and the negative surface
CXOH charges, so
int
KOH ¼ expð
Fw0 =RT Þ ð7Þ
CX ½OH
 rstr þ r0 ¼ F ðCX
CSO Þ ð11Þ
where ΓXCl and ΓXOH denote the surface densities (mol/ which means that the net charge of the particle results
m2) of XCl− and XOH− , respectively. ΓXCl and ΓXOH from the non-screened X sites minus the deprotonated
represent the positive structural charges canceled by the surface sites.
interlayer anions and ΓX represents the non-screened The net OH− consumption (σOH), as previously said
structural charges. [Cl−] and [OH−] are the activities of is a combination of hydroxyl used by exchange and
the anion in the solution bulk. deprotonation processes. It can be expressed as
The surface also presents SOH groups, which can
produce hydroxyl consumption according to Eq. (2), rOH ¼ F ðCXOH þ CSO Þ ð12Þ
producing also the compensation of a portion of the
This model of the solid particle must be combined
structural charge. This reaction is equivalent to the
with an electrostatic one in order to describe charge
deprotonation process
distribution and potential drop across the solid–liquid
SOH ¼ SO− þ Hþ ð8Þ interface. The Stern–Gouy–Chapman model is used

which has the intrinsic deprotonation constant:

1 CSO ½Hþ  Table 3
Kaint ¼ ¼ expð
Fw0 =RT Þ ð9Þ Refined parameters of the proposed model for Zn,Al-LDH samples
int
KH CSOH
Sample pKint
a pKint
Cl pKint 2 2 2
OH FNv (C/m ) FNstr (C/m ) C (F/m )

where Kaint is the intrinsic deprotonation constant of ZnAl2 12.05 0.20 − 1.83 0.28 0.30 0.08
SOH. KH int
is the intrinsic constant of the reverse reaction ZnAl3 11.80 − 0.15 − 1.67 0.25 0.20 0.07
11.85 − 0.44 − 1.85 0.22
(protonation of SO− groups), ΓSOH and ΓSO are the ZnAl4 0.14 0.08
34 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37

here, which assumes that the surface plane and the solution (Hiemstra et al., 1996), and Σsj gives the
diffuse layer are separated by a region of constant neutralization of the oxygen valence (V = − 2) produced
capacitance called the Stern layer and that the potential by bonds with lattice cations, or protons and water
drop across the diffuse layer is described by the Gouy– molecules from solution.
Chapman equation. The set of equations, confirming In the case of Zn–Al-LDH samples we have
this electrostatic model are considered, as previously said, two kinds of surface
hydroxyl groups (Zn2AlOH and Zn3OH): the Σsj will be
w0
wd ¼ ðrstr þ r0 Þ=C ð13Þ affected by the presence or absence of Al3+ cations in

rd ¼
0:1174I1=2 sinhð Fwd =2RT Þ ð14Þ

rstr þ r0 þ rd ¼ 0 ð15Þ
where σd and ψd are the charge and the potential in the
diffuse layer and I is the ionic strength in the solution,
calculated with the Davies equation.
In this model we have proposed only one type of
reactive surface group for protonation–deprotonation.
This is the classical treatment of the metal (hydr)oxide
surface, but it can result physically unrealistic because
there is often more than one surface reactive group in
real samples (Parks, 1990). To overcome this limitation
Hiemstra et al (1989) have developed a method to
estimate the intrinsic protonation constant of surface
groups of metal oxides and hydroxides. This method has
evolved into the termed MUSIC model (Bleam, 1993;
Hiemstra et al., 1996), that has been applied to explain,
describe and predict the surface behavior of iron,
titanium, and other metal oxides as well as phyllosilicate
clays (Avena, 2002). The modified MUSIC model
relates the affinity of surface groups for protons to the
Lewis basicity of surface oxygen atoms, which are the
surface groups capable of protonation–deprotonation
reactions. It states that bonds with the nearest atoms
neutralize the valence V of an oxygen atom in the
particle bulk. Each surrounding atom will neutralize part
of V with their bond valence (s), estimated as s = z / CN,
where z is the valence and CN is the coordination
number of the corresponding adjacent atom. At the
surface, part of the bonds with the surrounding cations
will be broken and new bonds will be formed with
hydrogen cations and water molecules. This usually
leads to overcompensation or undercompensation of the
valence V of surface oxygen, which will react to return
to an exact compensation of his valence. Mathemati-
cally, it can be written as:
X 
logKHint ¼
19:8 sj þ V ð16Þ
Fig. 5. Model predictions for site distribution at NaCl concentration
0.1 M with parameters of Table 3. Surface sites densities are multiplied
where the parameter − 19.8 results from the calibration by the Faraday constant to obtain C/m2 units. Triangles, XCl− groups;
of the model with the experimental protonation constant solid circles, XOH− groups; squares, X groups; open circles, Zn2AlO−
of several oxo- and hydroxo-metal complexes in groups. a) ZnAl2 sample; b) ZnAl3 sample; c) ZnAl4 sample.
 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37 35

the surroundings of the oxygen. For Zn2AlOH and surface area. Therefore, the number of edge groups will
Zn2AlO− groups it can be calculated that: be negligible when compared to the number of basal
X surface groups and the former will only be important if
sj ¼ 2sZn þ sAl þ msH þ ið1
sH Þ ð17Þ the latter are nonreactive. This is the case of phyllosi-
licate clays as montmorillonite and illite, that possess
where sZn is the bond valence of Zn (sZn = 2/6), sAl is the
inert SiO2 groups at the basal surfaces, and ten edge
bond valence of Al (sAl = 3/6), and sH is the bond
groups control their acid–base behavior (Avena, 2002).
valence of a proton donating bond (sH = 0.8). The
However, surface groups at the basal planes are rather
integers m and i denote the number of proton donating
reactive in LDHs (Rojas Delgado et al., 2004), and thus
and proton accepting hydrogen bonds formed with the
reactivity of edges can be ignored.
surrounding water molecules, and the number 2
Although up to six parameters (Nstr, Ns, Kaint, KCl int
,
indicates that two Zn2+ are bonded to the surface int
KOH and C) must be evaluated in order to solve the
oxygen in the group.
equations of the proposed model, some of them can be
In the case of Zn2AlO− groups, m = 0 and i could be
estimated a priori from structural data or using the
either 1 or 2 (there is no information about the exact
modified MUSIC model, and used as initial parameters
number of water molecules bounded through hydrogen
in a fitting procedure. Nstr, whose values are shown in
bond to each surface groups). Calculations with this
Table 2, can be obtained from the total layer specific
model assign log KH int
(Zn2AlO−) = 12.5 for i = 1 and 8.6
area and the Al3+ content. Nv values can be obtained in a
for i = 2. As we do not know the number of water
similar way, it can be estimated around 2 × 10− 5 mol/m2
molecules bonded to the oxygen atom, only a range for
int (F Ns = 1.9 C/m2). This value may not be very accurate
the value of log KH can be estimated. This suggests that
because not all the hydroxyl groups in the solid are
the surface density of Zn2AlOH and Zn2AlO− will vary
accessible by the OH− in solution, being thus non-
notably through this pH range: at pH values lower than
reactive. The initial value for log Kaint parameter was
8.6 it could be mainly found ZnAlOH groups and at pH
10.5, the intermediate value between the calculated with
values higher than 12.5 principally ZnAlO− would exist.
the modified MUSIC model for Zn2AlOH groups
A similar estimation can be made for Zn3OH and
bounded to 1 and 2 water molecules. No values have
Zn3O− groups. The absence of Al3+ bonded to the int int
been found for log KCl and log KOH ; therefore their
oxygen atom will reduce the surface group affinity for
initial values in the fitting procedure have been
deprotonation, as its valence is less oversaturated. It can
considered equal to 1, although it seems that LDHs
be calculated that:
X have more affinity for OH− than for Cl− (Li et al., 2003).
sj ¼ 3sZn þ msH þ ið1
sH Þ ð18Þ Optimization of the parameter values was performed
using the simplex method (Avena and De Pauli, 1998)
For Zn3O− groups, i and m can take the same values on the σOH vs. pH and ζ vs. pH curves, assuming that
than for Zn2AlO− groups, and the values of KH int
ζ = ψd. This last assumption is usually not valid at
calculated for i = 1 and 2 are 15.8 and 11.9, respectively. relatively high (higher than 0.01 M) electrolyte concen-
Thus, at pH values below 11.9 these groups will be all trations because the potential drop in the interface is
protonated and no deprotonation process would be rather pronounced. This is why we have worked in
registered. This means that the OH− consumption 0.01 M NaCl for electrophoresis, although we recognize
registered in the potentiometric titrations shown in that calculated ψd may overestimate somewhat the zeta
Fig. 3 at pH values between 7 and 11 are due mainly to potentials. In spite of this, calculated ψd are compared to
Cl−/OH− anionic exchange and surface Zn2AlOH group ζ just to show that the model predicts the shape of the
deprotonation process. ζ vs pH curves and the isoelectric points.
Other oxygen-containing groups could be located at The calculated σOH vs. pH and ζ vs. pH curves for the
the surface of Zn–Al LDHs. If the ordering of cations in three prepared samples are shown in Fig. 3, whereas
the layers is not perfect, hydroxyls groups coordinated Table 3 list the best-fit parameters for all modeled
to two Al3+ cations, although in low concentration, samples. The model predicts quite well the shape of the
could be present. The edge surface may also contain curves and their shift with variations of electrolyte
groups different from that of the bulk of the surface, and concentration and zeta potential data. The best-fit
hydroxyl groups bonded to less than three metallic parameters are quite in agreement with the initial
centers can be found. However, since the thickness of estimates in all samples. log Kaint is in the range of 8.5–
the particles is much lower than their length and width, 12.5 calculated by the MUSIC model and values of Nstr
the edge surface area is very small compared to the basal are quite similar to that calculated by structural data. Only
36 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37

for ZnAl4 sample σstr slightly deviates to lower values, exchange properties of LDHs because Zn2AlO− negative
which could be assigned to the carbonate presence in the groups produce screening of structural charge, decreasing
sample. The parameters also indicate that affinity for OH− the number of exchange sites available and reducing the
is higher than for Cl−(pKCl int
N pKOH
int
), as stated previously. surface potential, thus decreasing the electrostatic affinity
The Nv value is quite lower than the initial estimate, as of the surface for chloride anions. As MII/MIII increases this
suggested previously, indicating that only 10% of the OH effect becomes more important, as the model predicts that
groups are accessible by the solution and thus capable of the density of exchange sites (Nstr) decreases more rapidly
deprotonation. than protonation–deprotonation sites density (Nv). Thus, at
As can be seen from optimization values, decreasing pH = 12 and 0.1 electrolyte concentration, for ZnAl2
MII/MIII ratios leads to an increase in Nstr values. As we sample 46% of X sites is available for exchange, whereas
have seen before, increasing Al3+ content are supposed for ZnAl4 sample only 25 % of the sites is available.
to produce increasing positive charge excess in the layer,
which means an increasing number of exchange sites. 5. Conclusions
On the other hand, Nv value decreases as MII/MIII
increases because only Zn2AlOH/ Zn2AlO− are consid- A model for the reactivity of Zn,Al-LDH surface has
ered to be reactive, and their surface concentration is been applied in order to explain the hydroxyl adsorption
thus proportional to the content of Al3+. pKaint parameter and electrokinetic properties as a function of pH and
shows little variation with increasing MII/MIII ratio, NaCl concentration and to explain the effects of the
which seems to support the idea that the same type of sample MII/MIII ratio. The adsorption is explained as the
protonation–deprotonation surface groups are reacting sum of two contributions: OH−/Cl− anionic exchange
int
in all cases. Little variations in KCl values also shown in because of the presence of structural charges and
Table 3, but we have found no clear explanation for this. protonation–deprotonation of surface hydroxyl groups.
Fig. 5 shows the predicted LDHs site distribution in The reactive oxygen atoms in the surface of LDH
0.1 M electrolyte as a function of pH. The site distribution particles are bonded to 2 Zn2+ cations and one Al3+
at other electrolyte concentrations is not shown here but cations (Zn2AlOH groups), while oxygen atoms bonded
follows similar trends. For all samples studied the model to 3 Zn2+ cations are predicted to be non-reactive.
predicts that at pH around 7 the particle surface presents As MII/MIII increases the number of structural charges
mainly exchange sites attached to chloride anions and decreases, producing a lower exchange capacity of the
protonated Zn2AlOH groups. It also predicts a small but solid, and a lower OH− consumption. On the other hand
noticeably density of unoccupied exchange sites, which the lower content of Al3+ reduces the number of Zn2AlOH
would be responsible of the positive charge of the groups, although the extent of this reduction is lower than
particles, resulting in a positive zeta potential. At pH N 7 for exchange sites.
Zn2AlOH groups start to consume OH− from solution and According to the model the structural charge is
Zn2AlO− groups appear. Their surface concentration compensated at the surface by anions attached to
reaches a plateau at pH above 12. The model predicts that exchange sites and by negatively charged Zn2AlO−
acid–base surface groups are not totally protonated, being groups. The remaining non-screened structural charges
the deprotonated groups percentage around 50% for all generate a net positive surface potential in a wide range
samples. Above pH = 9 there is also a significant OH− of pH. Increasing variable charges with increasing pH
consumption due to the attachment of hydroxyl anions to produces a reduction in the exchange properties of the
exchange sites X, forming XOH− groups. It is noteworthy LDH and finally a reduction in the positive zeta
that Cl− anions start to detach from X sites at pH above 7 potential, reaching an PZC at pH around 11.8.
because of an electrical effect, caused by the presence of The Al3+ content reduction increases the variable
deprotonated Zn2AlO− groups, that compensates part of charges contribution to the electrical behavior of LDHs
the positive structural charge, thus decreasing the surface surface, reducing the LDHs exchange capacity at high
potential. pH values.
The model predicts that most of the positive structural
charge of LDHs is screened by XCl− (at low pHs) or Acknowledgements
XOH− and Zn2AlO− (at high pH) surface groups. Thus,
the net surface charge concentration is very low and even This study was financed by CONICET, SECYT and
decreases at high pHs, becoming zero at a pH around 11.8. AECI (PCI program). R. Rojas Delgado thanks AECI
The model also predicts that Zn2AlOH groups play an (MAE program) and CONICET (post-doctoral grant)
important role in electrical surface behavior and anionic for economic support.
 R. Rojas Delgado et al. / Applied Clay Science 40 (2008) 27–37 37

References Kloprogge, J.T., Frost, R.L., Hickey, L., 2004. FT-Raman and FT-IR
spectroscopic study of synthetic Mg/Zn/Al-hydrotalcites. Journal
Avena, M.J., 2002. Acid–base behavior of clay surfaces in aqueous of Raman Spectroscopy 35, 967–974.
media. In: Hubbard, A. (Ed.), Encyclopedia of surface and colloid Kraepiel, A.M.L., Keller, K., Morel, F.M.M., 1998. On the acid–base
chemistry of permanently charged minerals. Environmental
science. Marcel Dekker, New York, pp. 17–46.
Avena, M.J., De Pauli, C.P., 1998. Proton adsorption and electroki- Science and Technology 32, 2829–2838.
netics of an Argentinean montmorillonite. Journal of Colloid and Legrouri, A., Lakraimi, M., Barroug, A., De Roy, A., Besse, J.P., 2005.
Removal of the herbicide 2,4-dichlorophenoxyacetate from water
Interface Science 202, 195–204.
Badreddine, M., Barroug, A., Khaldi, M., Legrouri, A., De Roy, A., to zinc–aluminium–chloride layered double hydroxides. Water
Besse, J.P., 1998. Chloride-hydrogenophosphate ion exchange into Research 39, 3441–3448.
the zinc–aluminium–chloride layered double hydroxide. Materials Li, S.-P., How, W.-G., Han, S.-H., Li, L.-F., Shao, W.-A., 2003. Studies
on intrinsic ionization constants of Fe–Al–Mg hydrotalcite-like
Chemistry and Physics 52, 235–239.
Baeyens, B., Bradbury, M.H., 1999. A mechanistic description of Ni compounds. Journal of Colloid and Interface Science 257,
and Zn sorption on Na-montmorillonite Part I: titration and 244–249.
sorption measurements. Journal of Contaminant Hydrology 27, López-Salinas, E., García-Sánchez, M., Montoya, J.A., Acosta, D.R.,
Abasolo, J.A., Schifter, I., 1997. Structural characterization of
199–222.
Bleam, W.F., 1993. On modeling proton affinity at the oxide/water. synthetic hydrotalcite-like [Mg1 − xGax(OH)2](CO3)x / 2×·mH2O.
Journal of Colloid and Interface Science 159, 312–318. Langmuir 13, 4748–4753.
Newman, S.P., Jones, W., 1999. Comparative study of some layered
Bruce, D.W., O'Hare, D., 1997. Inorganic Materials. Wiley, Chichester.
Carlino, S., 1997. The intercalation of carboxylic-acids into layered hydroxide salts containing exchangeable interlayer anions. Journal
double hydroxides— a critical evaluation and review of the of Solid State Chemistry 148, 26–40.
different methods. Solid State Ionics 98, 73–84. Parks, G.A., 1990. Surface energy and adsorption at mineral/water
interfaces: an introduction. In: Hochella, M.F., White, A.F.
Cavani, F., Trifiro, F., Vaccari, A., 1991. Hydrotalcite-type anionic
clay: preparation and application. Catalysis Today 11, 173–301. (Eds.), Mineral–Water Interface Geochemistry. Reviews in Miner-
Choy, J.-H., Kwak, S.-Y., Jeong, Y.-J., Park, J.-S., 2000. Inorganic alogy, vol. 23. Mineralogical Society of America, Washington D. C.,
layered double hydroxides as nonviral vectors. Angewandte pp. 133–175.
Rives, V., 2001. Layered Double Hydroxides: Present and Future.
Chemie 11, 1671–1674.
De Roy, A., Forano, C., El Malki, K., Besse, J.P., 1992. Trends in Nova Sci. Pub. Inc., New York.
pillaring chemistry. In: Ocelli, M.L., Robson, H.E. (Eds.), Rives, V., Ulibarri, M.A., 1999. Layered double hydroxides (LDH)
Expanded Clays and other Microporous Solids. Van Nostrand intercalated with metal coordination compounds and oxometalates.
Coordination Chemistry Reviews 181, 61–120.
Reinhold, New York, pp. 108–169.
Ennadi, A., Legrouri, A., De Roy, A., Besse, J.P., 2000. X-ray Rojas Delgado, R., Arandigoyen Vidaurre, M., De Pauli, C.P., Ulibarri,
diffraction pattern simulation for thermally treated [Zn–Al–Cl] M.A., Avena, M.J., 2004. Surface-charging behavior of Zn–Cr
layered double hydroxide. Journal of Colloid and Interface Science
layered double hydroxide. Journal of Solid State Chemistry 152,
568–572. 280, 431–441.
Hiemstra, T., De Witt, J.C.M., Van Riemsdijk, W.H., 1989. Multisite Sposito, G., 1984. The Surface Chemistry of Soils. Oxford University
proton adsorption modeling at the solid/solution interface of (hydr) Press, New York.
Ulibarri, M.A., Labajos, F.M., Rives, V., Trujillano, R., Kagunya, W.,
oxides: a new approach. Journal of Colloid and Interface Science
133, 91–117. Jones, W., 1994. Comparative study of the synthesis and properties
Hiemstra, T., De Witt, J.C.M., Van Riemsdijk, W.H., 1996. Intrinsic of vanadate-exchanged layered double hydroxides. Inorganic
proton affinity of reactive surface groups of metal (hydr)oxides: Chemistry 33, 2592–2599.
Ulibarri, M.A., Pavlovic, I., Hermosín, M.C., Cornejo, J., 1995.
the bond valence principle. Journal of Colloid and Interface
Science 184, 680–692. Hydrotalcite-like compounds as potential sorbents of phenols from
How, W.-G., Su, Y.-L., Sun, D.-J., Zhang, C.-G., 2001. Studies on zero water. Applied Clay Science 10, 131–145.
Wells, A.F., 1984. Structural Inorganic Chemistry, 5th edition. Oxford
point of charge and permanent charge density of Mg–Fe
hydrotalcite-like compounds. Langmuir 17 (6), 1885–1888. Sci. Pub., Oxford.
Khan, A.I., O'Hare, D., 2002. Intercalation chemistry of layered
double hydroxides: recent developments and applications. Journal
of Materials Chemistry 12, 3191–3198.