Journal of Colloid and Interface Science 336 (2009) 599–606

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

Adsorption of polyamine on clay minerals

C. Blachier a,*, L. Michot a, I. Bihannic a, O. Barrès a, A. Jacquet b, M. Mosquet b
a
Laboratoire Environnement et Minéralurgie (LEM), Pôle de l’eau, UMR 7569, Nancy Université, CNRS15 Avenue du Charmois, 54 500 Vandoeuvre lès Nancy, France
b
LCR, Laboratoire Central de Recherche, Lafarge, 95 rue du Montmurier, BP 15, 38291 St. Quentin Fallavier Cedex, France

a r t i c l e i n f o a b s t r a c t

Article history: This work aims at a better understanding of the interaction between a polycationic quaternary amine
Received 16 January 2009 polymer (F25) and three different clay minerals: montmorillonite, illite, and kaolinite. For this, adsorption
Accepted 3 April 2009 isotherms of F25 on the clay surfaces were measured together with the evolution of the CEC along the
Available online 14 April 2009
isotherm, which revealed that cation exchange plays an important role in the adsorption process. These
first results were confirmed by XRD measurements on dried powders that are evidence in the case of
Keywords: montmorillonite of the presence of polymer in the interlayer spaces. In addition, the evolution of the
Clay
short range structure of clay minerals suspensions on polymer adsorption was followed by WAXS exper-
Polycation
Adsorption
iments. Polymer intercalation was observed while the structure of the resulting stacking appeared to
Flocculation change slightly along the polymer adsorption isotherm. Diffuse reflectance infrared measurements
Infrared spectroscopy revealed that significant conformational changes occur on polymer adsorption onto montmorillonite sur-
Zeta potential faces. Furthermore, adsorption above the CEC is observed which involves a charge reversal of clay mineral
SAXS-WAXS surfaces, the zero charge being reached for an adsorbed amount corresponding to the CEC. Finally, floc-
culation was discussed compared to adsorption amounts and zeta potential measurements, confirming
that optimum flocculation concentration is reached for noncharged particles.
Ó 2009 Elsevier Inc. All rights reserved.

1. Introduction strongly depends on the interactions between colloid surfaces

and chemical reagents [6,9].
Clay polymer interactions are of primary importance in numer- In the case of clay minerals that bear a structural negative
ous scientific fields ranging from materials science, where numer- charge, cationic products such as quaternary amine compounds
ous studies are devoted to the synthesis and characterization of are often used as flocculating agents of clay dispersions [10–13].
clay nanocomposites (e.g., [1,2]), to soil science and agronomy, The adsorption of cationic compounds onto clays has then been
where clay polymer interactions in soils control to a large extent investigated since the 1950s and most studies reveal that cationic
soil structure and plant nutrition (e.g., [3–5]). Clay polymer inter- exchange is the major driving force for adsorption (e.g., [14,15]).
actions are also particularly important in the numerous applica- Although numerous works have been devoted to such systems,
tions in which clay flocculation processes play a role, for there is still a relative lack of information about the correlation be-
instance, in mineral processing, oil drilling, paper making, or water tween adsorption mechanisms and optimum flocculation for many
clarification. In the case of mineral processing and oil drilling the polycationic compounds.
interest in such processes results from increasing environmental In the present paper we studied the adsorption of a quaternary
awareness that requires the reduction of the volume of clay tailings polyammonium cation on clay surfaces by varying the clay type
[6,7]. In the paper industry, flocculants called retention aids are and the nature of the exchangeable cation. The cationic compound
used for fines solid retention in papermaking processes [8]. In chosen for this study, referred to as F25, is often used as an organic
drinking water treatment, coagulation and flocculation are the ini- coagulant in water clarification processes and in mining applica-
tial steps that are used to reduce water turbidity before further fil- tions. The clay minerals used were montmorillonite, illite, and kao-
tration and adsorption treatments. Flocculation is a complex linite. The main aim of our study is to relate the adsorption
process, which involves the destabilization of a colloidal dispersion mechanism to the flocculation properties of the various systems
through the agglomeration of the fine particles by chemical re- in order to better understand the optimum of flocculation. To reach
agents. The flocs thus obtained settle under gravity and are then such a goal, adsorption mechanisms were investigated by combin-
eliminated from water. As a consequence, flocculation efficiency ing adsorption isotherms measurements with infrared spectros-
copy, X-ray diffraction, and wide angle X-ray scattering (WAXS).
In parallel, the flocculation behavior of the system was studied
* Corresponding author.
E-mail address: [email protected] (C. Blachier).
through the evolution of the surface charge of the materials.

0021-9797/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2009.04.021
600 C. Blachier et al. / Journal of Colloid and Interface Science 336 (2009) 599–606

2. Experimental 2.2. Adsorption studies

2.1. Materials Adsorption tests were carried out in a batch process by varying
adsorbate concentration using a constant clay amount of 10 g/L.
Three clay samples were used in this study: Montmorillonite The temperature was fixed at 28 °C and the pH was close to 8.5
Bleu de Sardaigne (supplied by Socodis), Illite du Puy (Socodis) (natural pH of the clay suspension). A weighed sample of clay
and Kaolinite BS3 (AGS). The raw samples were characterized by (mclay) was mixed with 40 ml (V) of F25 solution of known concen-
X-ray diffraction and infrared spectroscopy. Such analyses revealed tration in a centrifuge bottle. The mixture was shaken in a thermo-
that quartz, feldspath, dolomite, and calcite were the major impu- stated chamber overnight as preliminary kinetic tests had shown
rities accompanying raw clay. Prior to use, the samples were then that adsorption equilibrium was reached within 1 h. The mixture
purified and homoionized under their sodium and calcium forms was then centrifuged at 46250 rcf (19,000 rpm). The F25 equilib-
[16]. A 40 g/L clay suspension was first decarbonated for 2 h at rium concentration in the supernatants was determined by TOC
90 °C with 1 M sodium acetate solution at pH 5 by addition of ace- analysis while calcium and sodium concentrations were measured
tic acid. After centrifugation, the solid was exchanged three times by atomic absorption spectroscopy (Perkin Elmer A800). The ad-
in 1 M NaCl or CaCl2. The suspension was then dialyzed several sorbed polycation amount (Qe) was then obtained by difference
times in deionized water until the water was chloride free (AgNO3 between the initial (Ci) and the equilibrium (Ce) concentrations
and conductivity test). Finally, the fraction <2 lm was collected using the formula Qe = (Ci  Ce)*V/mclay.
after decantation of the impurities. After this purification proce- The solid fraction was washed three times in deionized water in
dure, XRD and FTIR patterns revealed only trace amounts of felds- order to remove the nonadsorbed polycation molecules. The
path and dolomite, showing the efficiency of the process. The organoclay thus obtained was dried and crushed and the powder
structural formulas of the purified samples were determined by was analyzed by diffuse reflectance Fourier transform infrared
ICP-AES determination of major element concentrations (CRPG- (DR-FTIR), DRX, and zeta potential measurements.
SARM) whereas the cation-exchange capacity (CEC) of the samples
was measured by exchange with hexaamminecobalt [17]. The re- 2.3. Flocculation test
sults are summarized in Table 1. It can be noted that the experi-
mental CEC value of the montmorillonite is a little bit lower than For each experiment, 0.2 g of clay was put in 100 ml solution at
the theoretical one predicted by the structural formula. The spe- the desired polymer concentration. The suspension was shaken for
cific surface areas measured by nitrogen adsorption at 77 K and 1 h. At 10 min after stirring was stopped, the top 20 ml of suspen-
application of the BET method were equal to 100, 96, and sion was transferred to a turbidimeter (Hach 2100P) for turbidity
24 m2 g1 for raw montmorillonite, illite, and kaolinite, respec- measurements.
tively. In the case of montmorillonite the specific surface area
was also determined by the EGME method [18] and yielded a value 2.4. Characterization techniques
of 624 m2 g1.
The water-soluble polymer was a quaternary polyammonium Diffuse reflectance Fourier transform infrared spectra were re-
cation F25½ðMe2 NCH2 CHOHCH2 Þþ

n ; nCl  used in water clarification
corded in the range 4000–600 cm1 using a Bruker IFS-55 spectro-
processes and mining applications as an organic coagulant. The photometer. The spectrum of the pure dried polymer was obtained
molecular weight of this polymer was about 5000 Da. The cationic in transmission mode using the microscope accessory. In all cases,
charge measured by titration of the polyamine with a solution of 200 scans were collected and averaged for each spectrum.
sodium dodecyl sulfate (SDS) was equal to 6.7 meq. This technique Electrophoretic mobility measurements were performed using
is based on the formation of a water-insoluble complex between an automated electrophoresis instrument, zetaphoremeter IV sup-
the anionic surfactant and the cationic reagent on interaction be- plied by CAD instrument. The setup is equipped with a microscope
tween SDS and polycationic. This method is derived from a two- and a CCD camera that allows measuring the electrophoretic
phase titration technique originally applied to anionic surfactants mobility of the particles that can be transformed into f-potential
[19–21]. Polymer concentrations were measured using a total or- data using Henry equation.
ganic carbon analyzer (Shimadzu TOC 5000). The carbon percent- The X-ray diffraction patterns of the solids were obtained using
age in pure F25 was determined as 42% by calibration. Total a D8 Advance Brüker AXS diffractometer working at 35 kV and
water solubility of this polymer was certified by the manufacturer. 45 mA, and employing CoKa filtered radiation (k = 0.1789 nm). Ori-
In addition, centrifuge and conductivity tests were carried out at ented preparations on glass slides were mostly used.
different concentrations in order to validate that no micellication Wide angle X-ray scattering experiments were carried out on
or precipitation occurred. suspensions or pastes placed in 1-mm-thick capillaries at beamline
A2 (Hasylab at DESY, Hamburg, Germany) using a wavelength of
0.150 nm and a sample to detector distance of 30 cm.
Table 1
Average structural formulas and measured cationic-exchange capacities of the three
clay samples. 3. Result and discussion
Counterion Structural formula CEC (meq/
100 g) 3.1. Adsorption studies
Montmorillonite Na — 94.36
Ca (Si7.81 Al0.19) (Al2.52 Fe0.69 96.54 Fig. 1a presents the adsorption curves of F25 on the six samples
Mg0.79)O20 (OH)4 Ca0.49 investigated plotted in adsorbed amount in milligram per gram as
Illite Na — 17.59 a function of polymer/clay ratios. Whatever the sample, the iso-
Ca (Si6.9 Al1.1) (Al2.5 Fe0.89 Mg0.54) K1.13 20.23 therms display a similar shape with a strong increase at low concen-
O20 (OH)4 Ca0.21
trations followed by a pseudo-plateau at higher concentrations. The
Kaolinite Na — 5.13 adsorbed amount appears to be independent of the nature of the
Ca (Si2) (Al1.89 Fe0.04 Mg0.02) O5 (OH)4 7.78
exchangeable cation but strongly depends on the nature of the clay
Ca0.01
used, montmorillonite displaying a higher adsorption capacity than
 C. Blachier et al. / Journal of Colloid and Interface Science 336 (2009) 599–606 601

Fig. 2. Adsorption of F25 in charge as a function of the initial polymer/clay ratio

(w/w).

with the adsorption of highly charged polycation onto negatively

charged particles have reported that a proportion of the polycat-
ionic charges remains unassociated in loops and trains for higher
polymer loading involving an adsorption beyond the CEC [30].
Such a mechanism seems to be applicable to explain our data
and will be further discussed in the following sections.
In charge scale, it can be observed that both the F25 adsorption
curve and that corresponding to the released sodium are superim-
posed (Fig. 3a). The same result was obtained for calcium (not
shown). The replacement of counterions by polymer therefore al-
most follows a 1:1 relationship for low polymer concentrations
as already observed for smaller quaternary amines [14]. It can then
be concluded that sodium and calcium counterions are replaced
Fig. 1. Adsorption of F25 polycation as a function of the initial polymer/clay ratio stoichiometrically by polymeric charges on adsorption. This is fur-
(w/w) in mg/g (a) and mg/m2 (b). ther confirmed by the CEC evolution (Fig. 3b), which shows that at
some point all the counterions have been replaced by F25 cations.
Further examination of Fig. 3b confirms that F25 adsorption con-
illite and kaolinite. However, when expressing the adsorbed amount tinues even when no more counterions are available for exchange.
per square meter (EGME method) (Fig. 1b), the curves do not super- By plotting the same isotherms in a semilogarithmic represen-
impose, which shows that adsorption sensu stricto may not be the tation (Fig. 4), it is possible to gain some further insight into the
only mechanism in this system. In order to better understand such adsorption mechanism. At low equilibrium concentrations, the
curves, and as charge is known to play an important role in the curve corresponding to Na montmorillonite is shifted toward lower
adsorption of cationic polymers, it appears relevant to replot these equilibrium concentrations compared to that of Ca-montmorillon-
isotherms using the amount of charge (in meq/100 g) adsorbed on ite whereas at higher concentrations the two curves merge into a
the samples (Fig. 2). Using such a representation, it is possible to di- single one. At low polymer concentrations, ionic exchange is then
rectly compare the adsorbed amount with the CEC of the sample. It the only operating mechanism. As exchanging sodium ions is eas-
appears that the isotherms correlate with the CEC of the samples, ier than exchanging calcium ions, the adsorption isotherm of Na-
and that adsorption curves at low polymer/clay ratios are almost montmorillonite is shifted toward lower equilibrium concentra-
overlayed with the 100% adsorbed theoretical curve revealing the tion. The shift between both curves that are nearly vertical remains
strong affinity of F25 for clay surfaces. In the case of montmorillonite constant until around 80% of the CEC is saturated. For higher val-
this also suggests that the polymer must have access to the internal ues, adsorption is mainly due to the rearrangement of the polymer
surfaces through intercalation in the interlayer regions. Such a chains onto clay surfaces resulting in the formation of a ‘‘loops and
tendency has already been observed in numerous studies (e.g., trains” conformation.
[22–26]).
In addition, it clearly appears for all samples that F25 adsorbs in 3.2. F25 intercalation
excess of the CEC of the samples. Similar observations have already
been described in numerous studies dealing with the interaction As suggested by the adsorption isotherms, as F25 molecules ex-
between cationic compounds and clay minerals. Adsorption in ex- change the interlayer cations, it is relevant in the case of montmo-
cess of the CEC could result from different mechanisms such as (i) rillonite to try to follow the possible intercalation of F25 molecules
adsorption involving alkyl chains as observed in the case of in the interlayer region and the textural and structural changes
HTDMA and other long alkyl chain amines [27,28], or (ii) intersala- associated with such intercalation. The evolution of the XRD pat-
tion (simultaneous interaction of anion and cation) which is often terns along the adsorption isotherm of F25 onto Na and Ca mont-
observed with cationic dyes [29]. In the case of F25, long alkyl morillonite is presented in Fig. 5a and b, respectively. In the case
chains are not present and no interactions with chloride anions of Na-montmorillonite, (Fig. 5a), the starting material displays a
are observed by conductivity measurements, which discard both classical XRD pattern with an average layer spacing of 12.5 Å cor-
aforementioned mechanisms. In contrast, several studies dealing responding to a dominant monolayer hydrate stage as observed
602 C. Blachier et al. / Journal of Colloid and Interface Science 336 (2009) 599–606

Fig. 3. Evolution of (a) sodium released and (b) CEC on adsorption of F25 onto Na-
montmorillonite.

Fig. 5. XRD pattern of montmorillonite as a function of polymer/clay ratio (a) Ca-

montmorillonite, (b) Na-montmorillonite.

For intermediate polymer loadings (around 7.5%), the diffrac-

tion patterns exhibit a significant decrease of crystallinity, as
judged from the width of the d001 signals. The peak width and
intensity increase again for high polymer loadings. For a large
range of polymer loadings, the patterns show two peaks, thus
revealing a coexistence of layers intercalated with sodium or cal-
cium cations, and layers intercalated with F25 molecules. This fea-
ture could also explain the decrease of crystallinity observed in the
intermediate polymer loading region, i.e., in the region of coexis-
tence of the two types of layers.
In order to gain some further insight into the intercalation
mechanism of F25 into the interlayer space of montmorillonite,
wide angle X-ray scattering experiments were carried out at beam-
Fig. 4. Adsorption isotherm of F25 polycation onto Ca and Na montmorillonite.
line A2 (Hasylab at DESY, Hambourg). In this case, the evolution of
the structure of the suspensions was followed directly in situ,
without any drying step as that needed for X-ray diffraction exper-
in numerous studies (e.g., [31,32]). The diffractogram of the start- iments. Evolution of WAXS patterns as a function of F25 adsorption
ing Ca-montmorillonite (Fig. 5b) displays a main d001 spacing is presented in Fig. 6. In these graphs, the intensity was multiplied
around 15 Å typical of a dominant bilayer state (e.g., [33,34]). by q (with q = 4*psin(h)/k where 2h is the scattering angle and k the
The addition of small amounts of F25 polymer immediately modi- wavelength) to the power of the slope of the scattering signal in
fies the patterns for both samples. For an initial content of 2.5% the low angle region in order to properly take into account the
weight of polymer a decrease of the peak corresponding to the form factor. In the case of Na-montmorillonite (Fig. 6a) the initial
starting clay is observed together with the development of a signal patterns observed in the absence of polymer do not display any
around 14.5 Å that can be assigned to F25 intercalated layers. In- correlation peaks. As in addition, the slope of the curve in the
deed, a similar d001 value was observed by Billingham and Breen low angle region is equal to 2 (not shown), this confirms that
using another cationic polymer with a nearly similar structure in dilute suspensions Na-montmorillonite is dispersed as single
[24]. layers [35,36]. In contrast for the same solid concentration, and
 C. Blachier et al. / Journal of Colloid and Interface Science 336 (2009) 599–606 603

in absence of polymer, Ca-montmorillonite (Fig. 6b) displays a where d is the particle size, k is the wavelength of the radiation
peak at 18.3 Å corresponding to a sample with three layers of (0.15 nm), h is the angle of the considered Bragg reflection, x is
water molecules. This shows that in the case of Ca-montmorillon- the width at the half of the intensity of the peak (2h scale), and K
ite due to ion correlations, the layers are associated in packs of a is a constant close to 0.9 depending on factors such as the geometry
few layers [35]. For polymer loadings as low as 2.5%, a peak at of the crystallites. Determination of pure and intercalated Ca mont-
14.6 Å appears characterizing the intercalation of F25 molecules morillonite crystallinity by this method shows that the number of
in the interlayer spaces. It must be pointed out that this value is ordered layers increases from 8 to 12 layers during the adsorption
the same as that obtained from XRD measurements showing that attesting that F25 adsorption is associated with a reconformation
F25 molecules totally prevent any water swelling. Furthermore, of the stacking of clay particles. In addition, these values are in good
in the case of Ca-montmorillonite, the WAXS patterns for interme- agreement with result obtained by high resolution transmission
diate loadings (5% and 10%) reveal the coexistence of Ca-montmo- electron microscopy (TEM).
rillonite with three water layers and polymer-intercalated
montmorillonite. It then appears that some layers are exchanged 3.3. Infrared spectroscopy
before others which may be linked to layer charge heterogeneities
that are nearly always present in natural swelling clays [37,38]. Samples obtained along the adsorption isotherm of F25 onto the
Further examination of the WAXS peaks using the Scherrer three clay minerals and for the two compensating cations were
equation allows us to calculate the number of ordered layers in analyzed by diffuse reflectance IR spectroscopy in order to gain
the elementary particles [39–41]. The Scherrer equation can be some further insight into the conformation of adsorbed polymer
written as molecules. We will here only focus on the results obtained for
Na-montmorillonite, which are the same as those obtained for
d ¼ Kk=x cos h;
Ca-montmorillonite. In the case of kaolinite, the adsorbed quanti-
ties are too low to enable a proper analysis of the spectra, whereas
for illite marginal differences are observed that will be discussed in
the text. For comparison purposes, the first presented spectrum is
that of pure dried F25 in the range 1000–4000 cm1 (Fig. 7). The
main observed signals are located at 3130, 3014, 2977, 2940,
2841, 1460–1480, and 1093 cm1. The signals at 3130 and
1093 cm1 correspond to the alcohol function (mOH and mCO,
respectively) whereas the other signals can be assigned to the
methyl and methylene groups (symmetric and antisymmetric
C–H stretching at 3014, 2977, 2940, and 2841) and methylene
(1477 cm1) and methyl (1469 cm1) groups scissoring modes
[41–43]. The fact that the stretching vibration of the OH group of
the alcohol is shifted toward lower wavenumber compared to
the literature data can be linked to the intense drying of the poly-
mer before analysis.
Fig. 8 shows the evolution of the infrared spectra with increas-
ing F25 adsorbed amounts. All these spectra are difference spectra
between samples after adsorption and the starting montmorillon-
ite. The subtraction was carried out after normalization at
3630 cm1 (structural OH stretching of montmorillonite). Bands
corresponding to adsorbed polymer molecules clearly appear for
all investigated loadings and increase in intensity for increasing
loadings. The signals assigned to the scissoring of methyl and
methylene groups do not seem to be affected by the adsorption
process and can then be used to check the quality of the subtrac-
tion procedure. Fig. 9 compares the integrated areas of these bands

Fig. 6. WAXS patterns of (a) Na-montmorillonite and (b) Ca montmorillonite as a

function of polymer/clay ratio. Fig. 7. FTIR spectrum of dried F25.
604 C. Blachier et al. / Journal of Colloid and Interface Science 336 (2009) 599–606

with the adsorbed amounts deduced from the adsorption iso-

therms (Fig. 1), and reveals a good correlation between both val-
ues. Whereas most signals in the subtracted spectra appear as
positive, there is a clear negative signal around 1635 cm1(Fig. 8).
This band corresponds to the bending of interlayer cation hydra-
tion water. In that context, the decrease of this signal can be inter-
preted as resulting from the exchange of interlayer cations by
adsorbed polymer molecules (Fig. 3). Fig. 10 displays the evolution
of the intensity (after baseline subtraction) of this band as a func-
tion of the initial polymer concentration. Two regimes are clearly
observed. For initial loadings 612.5%, a linear decrease is observed
followed by a plateau at higher values. This evolution is perfectly
correlated with the evolution of the released cations (Fig. 3).
The other bands assigned to polymer molecules are shifted on
adsorption. The most spectacular change is that of the mOH alcohol
band at 3130 cm1 which strongly decreases on adsorption. As no
band corresponding to C–O or C@O vibrations are observed in the Fig. 9. Integrated areas of the 1480 cm1 band compared to F25 adsorbed amount.

spectra, such an evolution cannot be due to polymer degradation.

Consequently, this feature may be interpreted as resulting from a 3.4. Zeta potential measurements
rigidification of the O–H bond resulting from interlamellar adsorp-
tion. Such an interpretation is partially confirmed by the spectra Fig. 11 presents the evolution of the zeta potential as a function
obtained on illite (not shown) that exhibit a less pronounced de- of the adsorbed amount (expressed as % of the CEC) for the three
crease in the alcohol absorption band that could then be linked investigated clay samples at natural pH. Independent of the nature
to the absence of interlamellar adsorption for this mineral. Parallel of the clay mineral, it appears that the zeta potential which is ini-
to the changes observed around 3000 cm1, significant modifica- tially negative for all samples reaches a value of 0 for an adsorbed
tions can be seen on the mC–O signal at 1093 cm1 which is strongly amount corresponding to the CEC. This confirms once again that
shifted to 1035 cm1. cation exchange is the main adsorption mechanism. For higher
Whereas the scissoring of the methyl and the methylene ap- loadings, charge reversal is observed for adsorbed values above
pears to be nearly unchanged with adsorption, small changes can the CEC. Similar results were obtained in the case of polycation
be observed in the range corresponding to C–H stretching, which adsorption on bentonite and were assigned to a mismatch between
are known to be very sensitive to conformational changes of the layer charge and size of the compensating polymer molecules. For
chains. This was recently evidenced in studies on the adsorption low polymer loadings all the charges can be compensated by flat-
of HDTMA and CTAB on montmorillonite [41–43], that showed that tened polymer molecules [30]. When the charge to be compen-
highly ordered chains of adsorbed molecules exhibited a peak at sated becomes below a threshold value, the size of the patches to
2916 cm1. Such a band position is also observed in the case of ad- be compensated can become smaller than the size of the polymer
sorbed F25 (Fig. 8) revealing significant conformational changes on molecule that will then leave charged segments toward the solu-
fixation onto montmorillonite surfaces. It must be pointed out that tion. Once this process is complete, no further adsorption can occur
the signal at 2916 cm1 appears for the lowest loadings in contrast which is confirmed by the plateau observed in the adsorption iso-
with what was observed in the case of CTAB where highly ordered therms at high loadings (Figs. 1–4).
chains were only present once an intercalated bilayer was present For the lowest adsorbed loadings, no variation in zeta potential
[41]. This suggests that due to the polymeric nature of F25, multi- can be observed in the case of montmorillonite. This may be due to
ple adsorption sites force the molecule to adopt a flat configuration the fact that for such loadings, both polymer-exchanged particles
on the surface with a low d-spacing of 14.6 Å. and calcium-exchanged particles are present, calcium-exchanged

Fig. 8. FTIR spectra of increasing loading of F25 intercalated montmorillonite.

 C. Blachier et al. / Journal of Colloid and Interface Science 336 (2009) 599–606 605

Fig. 12. Flocculation of 2 g/L clay suspensions as a function of initial F25/clay ratio
1 (a) and F25 adsorbed loading (b).
Fig. 10. Evolution of the IR water band at 1635 cm (a) and it intensity (b) as a
function of the polymer/clay ratio.

ure of the compensating cation is observed. The residual turbidity

reaches a value close to 0 for a relatively narrow window of poly-
mer loadings. On both sides of this window the turbidity increases.
According to these experiments the optimal flocculation conditions
can be defined as polymer/clay ratio of about 12.5% in the case of
montmorillonite, 5% for illite, and 0.75% for kaolinite. In order to
rationalize these results, it is more relevant to plot the same curves
as a function of the adsorbed amount expressed as a percentage of
the CEC (Fig. 12b). Using such a representation, all the curves are
superimposed and the optimal flocculation window corresponds
to exchange percentages between 70% and 110%. These results
are perfectly correlated to the evolution of zeta potential as this
range corresponds to zeta potential values between 10 and
+ 10 mV, i.e., the range classically assigned to the flocculation zone
in colloid science and water treatment operations.

4. Conclusions and perspectives

Fig. 11. Dependence of zeta potential and F25 adsorbed amount.
Using various structural and spectroscopic methods appears as
a very fruitful approach to understand the adsorption of charged
particles being dominant (Fig. 6b). Zeta potential measurements polymeric molecules at the surface of clay minerals. In addition,
are then likely sensitive to calcium-exchanged particles only. this can also provide relevant operational parameters to achieve
efficient flocculation procedures. According to the results obtained
3.5. Flocculation tests in the present study, when using charged polymers adsorption in
slight excess of the CEC occurs due to a mismatch between charge
Fig. 12a displays the evolution of residual turbidity for the six and size of the polymers, that ends up leaving mobile charged
samples investigated in the present study. The general shape of chains toward the solution for high polymer loadings. In that con-
these curves is similar for all samples and no influence of the nat- text, it would certainly be interesting to follow the evolution of
606 C. Blachier et al. / Journal of Colloid and Interface Science 336 (2009) 599–606

chain mobilities using 13C NMR along the adsorption isotherms. It [18] D.L. Carter, M.M. Mortland, W.D. Kemper, Specific Surface: Methods of Soil
Analysis, Physical and Mineralogical Methods, second ed., vol. 1, American
may also be relevant to study in more details, the influence of layer
Society of Agronomy, Soil Science Society of America, Inc., Madison, WI, 1986,
charge and layer charge heterogeneity of the clay minerals on the pp. 413–423.
adsorption features. [19] V.W. Reid, G.F. Longman, E. Heinerth, Tenside 4 (1967) 292.
[20] G.W. Power, The Volumetric Determination of Organic Sulfates or Sulfonates
by the Double Indicator Method, Communication C-225, Amoco Production
Acknowledgment Co., Tulsa, OK, 1970.
[21] E. Sabah, Water Res. 36 (2002) 3957–3964.
[22] C. Breen, H. Loughin, Clay Miner. 29 (1996) 775–783.
Support of beamtime at Hasylab by the European Community— [23] J. Billingham, C. Breen, J.O. Rawson, J. Yarwood, B.E. Mann, J. Colloid Interface
Reseach Infrastructure Action under the FP6 ‘‘Structuring the Euro- Sci. 193 (1997) 183–189.
pean Research Area” Programme (through the Integrated Infra- [24] J. Billingham, C. Breen, Vibr. Spectrosc. 14 (1997) 19–34.
[25] Y. Xi, Z. Ding, J. Colloid Interface Sci. 277 (2004) 116–120.
structure Initiative ‘‘Integrated Activity on Synchrotron and Free [26] A. Perez-Santano, J. Colloid Interface Sci. 284 (2005) 239–244.
Electron Laser Science”). Contract RII3-CT-2004-506008 is grate- [27] C. Wang, J. Colloid Interface Sci. 280 (2004) 27–35.
fully acknowledged. We thank Sergio Funari for his help during [28] Z. Zhong Zhang, Environ. Sci. 27 (1993) 1625–1631.
[29] Q. Li, Q.-Y. Yue, Y. Su, B.-Y. Gao, L. Fu, J. Hazard. Mater. 147 (2007) 370–380.
the SAXS-WAXS experiments on beamline A2. [30] C. Breen, Appl. Clay Sci. (15) (1999) 187–219.
[31] I. Berend, Clays Clay Miner. 43 (3) (1995) 324–336.
[32] E. Ferrage, B. Lanson, B.A. Sakharov, V.A. Drits, Am. Miner. 90 (8–9) (2005)
References 1358–1374.
[33] J.M. Cases, I. Berend, M. Francois, J.P. Uriot, L.J. Michot, F. Thomas, Clays Clay
[1] A. Usuki, N. Hasegawa, M. Kato, Adv. Polym. Sci. 179 (2005) 135–195. Miner. 45 (1) (1997) 8–22.
[2] J. Liu, W.-J. Boo, A. Clearfield, H.-J. Sue, Mater. Manufact. Process. 21 (2) (2006) [34] E. Ferrage, C. Tournassat, E. Rinnert, B. Lanson, Geochim. Cosmochim. Acta 69
143–151. (11) (2005) 2797–2812.
[3] C.A. Vacher, R.J. Loch, S.R. Raine, Austr. J. Soil Res. 41 (8) (2003) 1509–1520. [35] R. Hight Jr., W.T. Higdon, H.C.H. Darley, P.W. Schmidt, J. Chem. Phys. 37 (3)
[4] Awad M. Helalia, J. Letey, Soil Sci. Soc. Am. J. 52 (1) (1988) 243–246. (1962) 502–510.
[5] A. Nadler, Y. Steinberger, Soil Sci. 155 (2) (1993) 114–122. [36] L.J. Michot, I. Bihannic, K. Porsch, S. Maddi, C. Baravian, J. Mougel, P. Levitz,
[6] R. Hogg, Int. J. Miner. Process. 58 (2000) 223–236. Langmuir 20 (25) (2004) 10829–10837.
[7] P. Mpofu, J. Addai-Mensah, Miner. Eng. 17 (2004) 411–423. [37] G. Lagaly, Clays Clay Miner. 30 (3) (1982) 215–222.
[8] C. Ovenden, H. Xiao, Colloids Surf. 197 (2002) 225–234. [38] F. Villieras, L.J. Michot, J.M. Cases, I. Berend, F. Bardot, M. François, G. Gerard,
[9] J. Gregory, J. Colloid Interface Sci. 42 (1973) 448–456. J. Yvon, Stud. Surf. Sci. Catal. 104 (1997) 573–623.
[10] T.R. Jones, Clay Miner. 18 (1983) 399–410. [39] B. Kübler, M.C.R. Jaboyedoff, Acad. Sci. Paris, Earth Planetary Sci. 331 (2000)
[11] R.I.S. Gill, T.M. Herrington, Colloids Surf. 22 (1986) 51–76. 75–89.
[12] M. Janek, G. Lagaly, Colloid Polym. Sci. 281 (2003) 293–301. [40] H. Borchert, E.V. Schevchenko, Langmuir 21 (2005) 1931–1936.
[13] A. Vanerek, B. Alince, T.G.M. Van der Ven, Colloids Surf. 273 (2006) 193–201. [41] N.V. Venkataraman, S. Vasudevan, J. Phys. Chem. B 105 (2001)
[14] J.L. Mc Atee, Am. Miner. 44 (1959) 1230–1236. 1805–1812.
[15] R.D. Laura, Clays Clay Miner. 23 (1975) 61–69. [42] Y. Li, Langmuir 19 (2003) 2479–2484.
[16] D. Vantelon, Thesis Institut National Polytechniques de Lorraine, 2001. [43] H. Hongping, Spectrochim. Acta Part A 60 (2004) 2853–2859.
[17] J.C. Remy, L. Orsini, Sci. Sol. 4 (1976) 269–275.