An Assessment of the Surface Properties of Milled Attapulgite Using Inverse Gas Chromatography

Abstract

The most common means of reducing the particle size of solids is by grinding, a process which can affect the surface properties and the behavior of the solid in later stages (granulation, compaction, etc.), and which can influence the end-use properties of the final product. Inverse gas chromatography (IGC) measurements were used here to evaluate the influence of grinding, in a ball mill, on attapulgite. The milling experiments were performed in dry media for various periods. After 30 min of grinding, significant decreases in the particle size and specific surface areas were observed when calculated using different probes. No noticeable variation in the surface properties was observed by IGC either at infinite dilution or at finite concentration, however. In particular, the distribution functions of the adsorption energies (DFAE), giving information about the surface heterogeneity for both an apolar probe(octane) and a polar probe (isopropanol), remained unchanged, regardless of the grinding time. The stability of the surface energy with respect to the grinding process was seen to be related to the particular fibrous structure of the attapulgiteclay.

Full text

AN ASSESSMENT OF THE SURFACE PROPERTIES OF MILLED ATTAPULGITE
USING INVERSE GAS CHROMATOGRAPHY
L. BOUDRICHE
1,2
,B.HAMDI
2
,Z.KESSAI
¨SSIA
2
,R.CALVET
3,
*, A. CHAMAYOU
3
,J.A.DODDS
3
,AND
H. BALARD
4
1
Centre de Recherche Scientifique et Technique en Analyse Physico-Chimique (C.R.A.P.C.), BP 248, Alger Rp, 16004, Alger,
Algeria
2
Laboratiore d’Etude Physico-Chimique des Mate´riaux et Application a` l’Environnement, Faculte´deChimie,USTHB,BP32El
Alia, Bab Ezzouar, 16111, Alger, Algeria
3
Ecole des Mines d’Albi Carmaux, Campus Jarlard, 81013 Albi CEDEX 09, France
4
Laboratoire de Chimie Physique, ENSISA-W, 11, rue Werner, 68093 Mulhouse CEDEX, France
Abstract—The most common means of reducing the particle size of solids is by grinding, a process which
can affect the surface properties and the behavior of the solid in later stages (granulation, compaction, etc.),
and which can influence the end-use properties of the final product. Inverse gas chromatography (IGC)
measurements were used here to evaluate the influence of grinding, in a ball mill, on attapulgite. The
milling experiments were performed in dry media for various periods. After 30 min of grinding, significant
decreases in the particle size and specific surface areas were observed when calculated using different
probes. No noticeable variation in the surface properties was observed by IGC either at infinite dilution or
at finite concentration, however. In particular, the distribution functions of the adsorption energies
(DFAE), giving information about the surface heterogeneity for both an apolar probe (octane) and a polar
probe (isopropanol), remained unchanged, regardless of the grinding time. The stability of the surface
energy with respect to the grinding process was seen to be related to the particular fibrous structure of the
attapulgite clay.
Key Words—Adsorption Energy Distribution Functions, Attapulgite, Grinding, Inverse Gas
Chromatography, Palygorskite, Surface Properties.
INTRODUCTION
Grinding processes are important industrial opera-
tions, currently used to reduce the particle sizes of
various materials, and in particular to produce powders
from the rocks extracted from mines or quarries, as is the
case for attapulgite.
Different studies have been carried out on the effect
of grinding on clay minerals, such as the effect of
micronization on the crystalline structure of kaolinite
(Suraj et al., 1997). Ball milling was found to be a
relatively low-impact process where the structural
changes in the mineral were minimal, whereas use of
an oscillatory mill could lead to an almost complete
destruction of the clay structure, including amorphiza-
tion. Using the ball mill, a greater effect was observed
(by X-ray diffraction, XRD) for particles after dry
grinding than after wet grinding. A study of mont-
morillonite (Hrachova et al., 2007) revealed a complete
breakdown of its layers after 3 h of dry grinding in a
vibration mill. Papirer et al. (1986) monitored the
surface free-energy characteristics of mica (muscovite)
samples after grinding in water and organic solvents
(methanol and toluene) and found that in methanol the
dispersive component of the surface energy (g
s
d
)
increased significantly after very short grinding times.
The same observation was made regarding the evolution
of the specific surface area. The increase in these two
parameters could be related to the creation of a lateral
surface area of mica flakes. On the other hand, in toluene
g
s
d
decreased after prolonged grinding as a consequence
of particle agglomeration.
Grinding generally induces a fragmentation of the
original particles, creating new surfaces which may
exhibit new surface properties, in particular, under the
effect of the mechanical stresses imposed by the
grinding tool. Understanding the effects of a grinding
process is, therefore, important in order to characterize
the ground solids, from the point of view of their
surface-interaction potential. Analytical methods based
on the adsorption phenomenon are particularly well
suited to such a characterization, among them inverse
gas chromatography, which has many advantages in
providing information on the interactivity, morphology,
and heterogeneity of the solid surface.
The aim of the present work was to use these
analytical techniques on a crystalline hydrated magne-
sium aluminum silicate (Murray, 2000; Gala´n, 1996)
with a fibrous morphology, to see if its surface proper-
ties changed when it was subjected to a grinding process.
Attapulgite, or palygorskite, is a fibrous Mg-clay
* E-mail address of corresponding author:
rachel.calvet@mines-albi.fr
DOI: 10.1346/CCMN.2010.0580201
Clays and Clay Minerals, Vol. 58, No. 2, 143–153, 2010.

generally containing large amounts of Al(III) and Fe(III)
cations. Their precise crystal chemistries (site occu-
pancy, distribution of cations, di- or trioctahedral
character) have been the subject of many
studies (Gala´n and Carretero, 1999; Chahi et al., 2002;
Suarez and Garcı´a-Romero, 2006). Attapulgite belongs
to the class of 2:1 phyllosilicates (Garcı´a-Romero et al.,
2004) in which the sheets of silica tetrahedra are
periodically inverted with respect to the tetrahedral
bases. Consequently, the octahedral sheets are periodi-
cally interrupted and terminal cations must complete
their coordination sphere with water molecules. The
crystal structure was proposed for the first time by
Bradley in 1940 (Figure 1). Because of this inversion,
the structure of attapulgite exhibits a fibrous habit.
THEORETICAL CONSIDERATIONS
IGC theory
Unlike analytical gas chromatography (GC) where a
known stationary phase is used to separate and identify
the unknown injected product, Inverse Gas
Chromatography (IGC) uses well characterized mole-
cules, known as ‘probes,’ to determine the interaction
capacities of the unknown material packed into the
column. The technique can be carried out in two ways: at
infinite dilution (IGC-ID) (Balard et al., 2000) and at
finite concentration (IGC-FC) (Conder and Young,
1979; Balard, 1997).
Inverse gas chromatography at infinite dilution (IGC-ID)
IGC-ID consists of the injection of minute amounts of
vapor of the probe molecules, the properties (length,
polarity, acidity) of which are known at low concentra-
tions in such a way that the interactions between
adsorbed molecules can be considered as negligible.
Depending on the chemical nature of the probe
molecule, IGC-ID provides information about several
parameters: (1) The dispersive component of the surface
energy (g
s
d
) is obtained by injections of linear alkanes.
The parameter indicates the ability of the solid surface to
have non-specific interactions with probe molecules
(Dorris and Gray, 1979). (2) The nanomorphological
index I
M
(w
t
) can be calculated by injections of branched
or cyclic alkanes and the value gives information about
the regularity of the solid surface at the molecular scale
(Brendle´ and Papirer, 1997). (3) The specific component
of the surface energy (I
sp
) can be determined by
injections of polar probes, and gives information about
polar interactions (Balard et al., 2000). The principles,
advantages, and limitations of IGC-ID were reviewed by
Balard (2000) and will not be discussed further here.
Inverse gas chromatography at finite concentration
(IGC-FC)
In IGC-FC, a few mL of liquid probe are injected into
a column containing the solid to be analyzed, in order to
provide a covering that is approximately mono-layered
on the surface of the solid. A highly asymmetric
chromatographic peak results when ideal, nonlinear
conditions are fulfilled. Due to the presence of very
high-energy adsorption sites, a non-negligible part of the
injected probe is usually not eluted and the signal returns
to the initial base line. In order to assess this irreversibly
adsorbed amount, the temperature of the chromatograph
oven is increased to conditioning temperature, leading to
the appearance of a secondary small peak. This gives an
assessment of the irreversible part of the adsorption
phenomena from the ratio of the area of the thermo-
desorption peak with respect to the total area of the
chromatogram, allowing an irreversibility index (I
irr
), as
previously defined (Balard et al., 2008), to be computed
according to equation 1:
I
irr
=S
th
/(S
rv
+S
th
)(1)
where S
rv
is the surface of the main chromatographic
peak and S
th
is the surface corresponding to the
thermodesorption peak.
By applying ‘the elution characteristic point method’
(ECP) to the reversible isothermal part of the experi-
mental chromatogram, the desorption isotherm of the
probe molecule can be obtained from a single chromato-
Figure 1. (a) Crystal structure of attapulgite (Bradley, 1940) and (b) its polyhedral representation (Mckeown et al., 2002).
144 L. Boudriche et al.Clays and Clay Minerals

graphic peak (Conder and Young, 1979). The first
derivative of the isotherm is related directly to the net
retention time of each point of the diffuse front of the
chromatogram by equation 2:
@N
@P
8
>
:
9
>
;L;t
¼JDcðTt0Þ
mRTð2Þ
where, for a given characteristic point, Nis the number of
adsorbed probe molecules; P, the partial pressure of the
probe at the output of the column (directly related to the
height of the signal); t
r
, the retention time; t
0
, the retention
time of methane; D
C
, the corrected flow rate; m,themass
of the solid in the column; and L, the column length. R is
the universal gas constant and Tthe temperature.
Integration of equation 2 provides the desorption isotherm
which allows Henry’s constant and the BET constants to
be determined, as well as the specific surface area
(Figure 2). The surface heterogeneity influences the
shape of the desorption isotherm. Intuitively, the resi-
dence times of the adsorbed molecules depend on the
energy of the adsorption site. Analysis of the shape of the
isotherm therefore permits calculation of the distribution
function of the adsorption energies
Distribution function of the adsorption energies (DFAE)
Estimation of the surface heterogeneity of a solid
through the adsorption energy distribution function is
generally based on a physical adsorption model which
notes that the global isotherm may be considered to be
the sum of local isotherms of adsorption on iso-energetic
domains (patchwork model) (Balard, 1997; Rudzinski et
al., 1982; Balard et al., 1998). The distribution function
of the relative abundance of each type of domain having
the same characteristic energy of interaction (e)isthus
given by the following integral equation:
N(T,P)=NÐ
O
y(e,T,P)w(e)de(3)
where N(T,P) is the number of molecules adsorbed at
pressure Pand temperature Tof the measurement; N
0
,
the number of molecules corresponding to the formation
of a monolayer; y(e,T,P), the local isotherm correspond-
ing to adsorption sites having the same characteristic
adsorption energy, e;w(e), the so-called distribution
function of the adsorption energies (DFAE) describing
the energies which exist at the gas-solid interface; and O
the physical domain of the adsorption energy.
Solution of the integral in equation 3, which is a
convolution product, with Langmuir’s isotherm as
Kernel is not trivial because computing the DFAE
from the isotherm data corresponds to a deconvolution
operation. Different models have been applied, particu-
larly the well known condensation approximation and
the extended Rudzinski et al. method (Balard, 1997;
Rudzinski et al., 1982). A method has also been
developed allowing the lateral energy of interaction to
be taken into account, while also delivering a DFCA
corrected from this interaction so that a Langmuir local
equation can be used (Balard et al., 2008). This
approach, based on decomposition of the extrapolated
DFCA into two components (homogeneous and hetero-
geneous), delivers an index of heterogeneity of the solid
surface of interest for the probe used, which is equal to
the ratio of the area of the homogeneous component
(A
Hm
)overthetotalareaoftheexperimentalDFCA
(A
T
), so that I
hete
is given by equation 4:
Ihete ¼AHm
AT
ð4Þ
Starting from the DFCA corrected from the lateral
energy of interaction, Rudzinski-Jagiello approximations
canbecomputedupto4
th
order, designed as DFRJ,
Figure 2. Exploitation of the desorption isotherm.
Vol. 58, No. 2, 2010 Surface properties of milled attapulgite 145

using a computing method based on Fourier’s transforms
as described elsewhere (Balard, 1997).
MATERIALS AND METHODS
Attapulgite
The attapulgite used in this study came from the
region of Ghoufi, near Biskra in the northeast of Algeria.
The crude attapulgite was first coarsely crushed and
sieved to obtain a <1 mm size fraction allowing easier
analysis, measurements, and treatments to be made in
laboratory-scale processes.
Grinding conditions
Ball-milling experiments were carried out for differ-
ent processing times (15, 30, 60, 120, and 180 min), with
aMARNE1devicesuppliedbytheFAUREEquipment
Company (Limoges, France). One-liter jars were filled
with 1 kg of ceramic grinding media (balls with different
diameters: 20% of 24 mm, 20% of 19 mm, and 60% of
12 mm). The rotation speed of the ball mill was fixed at
70 rpm. A volume of 200 cm
3
of attapulgite powder was
placed in the jars with the grinding media.
Characterization of ground attapulgite
After each grinding experiment, particle-size distri-
bution measurements were performed by a laser diffrac-
tion method using a Malvern Mastersizer 2000 device,
fed using a dry-way feeder with a dispersion pressure of
3.5 bar.
The specific surface areas, BET constants, and meso-
and micro-pore volumes were determined using nitrogen
adsorption at liquid nitrogen temperature (ASAP 2010,
Micromeritics) by applying the BET, BJH, and HK
methods, respectively.
Phase identification was performed on a Panalytical
X’PERT Pro diffractometer (Philips) (CuKaradiation)
with a step size of 0.017º for 2yangles ranging from 5º
to 80º. The 2yvalues were compared with the JCPDS
database. The particle morphology was investigated
using a scanning electron microscope (Philips XL 30
model ESEM-FEG) operating at 3 kV. The elementary
composition of attapulgite was established by X-ray
Fluorescence (XRF) using a Philips MagiX model
(Mulhouse, France).
Inverse gas chromatography
The IGC measurements were performed using two
(commercial) GC devices (Agilent 7890 A and 6890), fitted
with sensitive flame ionization detectors (FID). Helium was
used as the carrier gas with a flow rate of 30 mL/min
measured with an electronic flowmeter (Flow 500-Agilent).
The temperatures of the injector and detector were 130ºC
and 200ºC, respectively. The columns were filled with a
mixture of attapulgite powder and glass beads, 90 mm in
diameter, in order to avoid fitting leaks by reducing the
pressure drop (<1 bar) through the columns.
The IGC-ID study was performed with an oven
temperature of 130ºC, using stainless steel columns
20 cm long and 0.6 cm in diameter. The columns were
conditioned overnight at 150ºC. The probes used were
linear alkanes (hexane C6, heptane C7, and octane C8)
(Fluka); a branched alkane, 2,3,4-trimethylpentane
(2,3,4-TMP) (Aldrich); a cyclic alkane, cycloctane
(cyclo8) (Fluka); and polar probes (chloroform (Acros
organics) and dichloromethane (Fisher)). All probes
used were of GC purity.
For IGC-FC, the chromatographic columns were
10 cm long with a diameter of either 0.6 cm or 0.3 cm.
The conditioning temperature was 130ºC and the
analysis temperature depended on the probe used,
according to the Conder (1979) criterion which stated
that the contribution of probe vapor to the flow of carrier
gas should not exceed 5% of the initial flow at the
maximum of the chromatographic peak. Two probes
were used in IGC-FC, octane (C8) at 53ºC and
isopropanol (IP) (Fluka) at 43ºC. The chromatograms
were analysed using software created by Balard.
RESULTS AND DISCUSSION
Before grinding
Observation by SEM (Figure 3) revealed the fibrous
structure of the attapulgite, which explains its large
specific surface area of ~125 m
2
/g and its significant
adsorption capacity, which in turn explains its applica-
tion in environmental remediation and in anti-diarrhoeal
medications.
The reflections identified in the XRD pattern
(Figure 4) indicated that the material consisted of
attapulgite but also contained some significant amounts
of dolomite (CaMg(CO
3
)
2
) and ankerite
(Ca(Fe,Mg,Mn)(CO
3
)
2
). The elementary composition,
established by XRF, was compared to that reported by
Bradley (1940) (Table 1). The two attapulgite samples
had similar compositions, and the differences could be
attributed to the geological origins of the materials, one
coming from Georgia and the other from Algeria.
Influence of grinding
Particle-size analysis. The first effect of the grinding
process was to reduce the particle size of the solid. The
values of the main characteristic diameters of the ground
attapulgite samples were measured (Table 2). The
particle diameters of the ground attapulgite decreased
quickly during the first few minutes and reached a limit
within2hofgrinding,withaweaktrendtore-
agglomeration for the longer processing times. The
evolution of the particle-size distribution followed the
grinding time (Figure 5).
Up to a 30 min grinding time, the particle-size
distribution curves were bi-modal, exhibiting an increas-
ing component centered around 30 mm, whereas the
146 L. Boudriche et al.Clays and Clay Minerals

coarse particle population decreased concomitantly.
Beyond this critical grinding time, the distribution
functions became mono-modal and independent of the
grinding time, indicating that a limit had been reached in
the grinding process.
BET surface analysis. To check whether this process
influenced the specific surface area and the surface
morphology of the attapulgite, the samples were
submitted to BET surface analysis (Table 3). During
the first hour of grinding, a decrease of 30% in the BET
specific surface area was observed, with an increase
occurring between 120 and 180 min. The initial
decrease, in apparent contradiction to particle-size
reduction, could be explained by changes in the particle
structure. Taking into account the measured particle size
and the external specific surface area (calculated with
the assumption of spherical particles), the BET specific
surface area of attapulgite mainly results from an
internal porosity and not from the particle size.
According to this interpretation, after 2 h of grinding
the increase in the BET specific surface area could be
attributed to the appearance of a new internal porosity.
The variations in the BET specific surface area agreed
with the variations in the mesoporosity. A significant
change in the BET constant was observed during the first
30 min, indicating large changes in the surface interac-
tion capacity. After this initial 30 min, it became stable.
XRD analysis. The grinding process in a ball mill for a
period of 3 h did not influence the mineral composition,
which remained constant within the resolution of XRD
(Figure 4).
In order to gain a better insight into the surface
properties of the attapulgite, the solid was examined
using IGC analysis in infinite dilution (IGC-ID) and
finite concentration (IGC-FC) conditions.
Characterization of samples by IGC-ID
The IGC parameters presented in Tables 46include
a relative standard deviation which takes into account
the variability due to measurements with the chromato-
graph and to the heterogeneity of the powder.
The value of the dispersive component of the surface
energy (g
s
d
) of attapulgite before grinding was found to
be 164 mJ/m
2
(Table 4). Similar values of g
s
d
were
Table 1. Chemical composition (XRF) of raw attapulgite.
Elements (%) A B
SiO
2
55.03 52.61
MgO 10.49 13.34
Al
2
O
3
10.24 10.67
CaO 9.55
Fe
2
O
3
3.53 4.66
K
2
O 0.47 1.11
(A) Chemical composition of attapulgite from Attapulgus
(Georgia), freed of impurities (Bradley, 1940).
(B) Chemical analysis of the clay fraction used in the present
study.
Figure 3. SEM image of attapulgite fibers.
Figure 4. XRD patterns of attapulgite from Algeria (Att: attapulgite, Ak: ankerite, D: dolomite).
Vol. 58, No. 2, 2010 Surface properties of milled attapulgite 147

observedbySaadaet al. (1995) on two other clays, an
illite and a kaolinite, which those authors attributed to
the insertion of the n-alkane probes into the slot-like
insertion sites present on the lateral surfaces of the
crystals, corresponding to a greater force field than that
resulting from a simple adsorption on a flat basal layer.
In the present study, the probe was assumed to be
inserted into the fibrous channels of the attapulgite or
into structure defects between fibers.
The g
s
d
values measured on initial and ground
attapulgite samples were compared (Table 4). Taking
into account the relative standard deviation (4%), the
values of g
s
d
decreased slightly when the grinding time
increased. The reduction by mechanical stress of some
structural defects of the attapulgite crystal, which are
able to receive the linear alkane probes, may be
responsible.
To support this hypothesis, the adsorption behaviors
of cyclic alkane probes were examined. The I
M
(w
t
)
parameters were measured on initial and ground
attapulgite samples (Table 4). The I
M
(w
t
) parameters
did not change significantly with grinding time, demon-
strating that the surface morphology was not signifi-
cantly affected by the grinding process.
By injecting acidic probes, such as chloroform and
dichloromethane, the basic character of the attapulgite
surface was assessed. The corresponding specific inter-
action parameters, I
sp
, were also evaluated (Table 4).
Again no significant change in these parameters was
observed with grinding time. One should note, however,
that the measured values of the I
sp
of attapulgite were
greater than those determined by Saada (1995), Saada et
al. (1995), and Aouadj (1997) on illite, kaolinite, and
muscovite samples. Several other basic probes, such as
dioxan, THF, ether, and pyridine were injected, but due
to their strong interaction with the acidic hydroxyl
groups on the attapulgite surface, they could not be
Table 2. Variation with dry grinding time (in a ball mill) of
the equivalent diameters of attapulgite particles, measured by
laser light diffusion.
Grinding time d
10
(mm)
d
50
(mm)
d
90
(mm)
Initial 31 510 1159
30 min 8 37 185
60 min 7 33 105
120 min 7 24 50
180 min 9 25 58
Figure 5. Evolution of the particle-size distributions of ground attapulgite samples, with increasing grinding times.
Table 3. Variation with dry grinding time (in a ball mill) of
the specific surface area (S
BET
), BET constant (C
BET
),
mesopore (V
meso
) volume, and micropore (V
micro
) volume of
ground attapulgite samples, determined by nitrogen adsorp-
tion at 77 K.
Grinding time S
BET
(m
2
/g)
0.3
C
BET
V
meso
(cm
3
/g)
V
micro
(cm
3
/g)
Initial 125.2 437 0.30 0.05
30 min 99.3 187 0.31 0.04
60 min 85.6 187 0.24 
120 min 85.8 209 0.24 
180 min 100.3 182 0.34 0.04
148 L. Boudriche et al.Clays and Clay Minerals

eluted through the chromatographic column even at an
oven temperature of 200ºC, testifying to the strongly
acidic character of the attapulgite surface.
The results obtained by IGC-ID analysis, for grinding
times up to 3 h, were therefore in agreement with the
evolution of the BET specific surface area. The decrease
in internal porosity led to the disappearance of adsorp-
tion sites detected by the alkane probes, which was
emphasized by the decrease in the dispersive component
of the surface energy g
s
d
.
The IGC-ID parameters are clearly related to the
adsorption sites having the greatest energy of interaction
(Balard et al., 2000). The IGC-FC analysis was then
used to provide information on the whole surface of the
samples under study.
Characterization of samples by IGC-FC
Determination of specific surface areas. The initial and
ground samples were submitted to IGC-FC analysis using
n-octane, an apolar probe, which is mainly sensitive to the
surface morphology, and a polar probe, isopropanol,
which is more sensitive to the surface functionality,
especially the presence of silanol groups. The specific
surface areas and the corresponding BET constants were
measured by IGC-FC with the aforementioned probes on
initial and milled attapulgite samples (Table 5). The
specific surface areas varied in the same way regardless
of the probe used: they decreased for the first hour of
grinding and increased after 2 h. The specific surface
areas obtained using n-octane as the probe were close to
those measured with nitrogen at 77 K. On the contrary,
isopropanol led to smaller specific surface area values,
equal to ~2/3 of those obtained by nitrogen and n-octane.
This discrepancy between the specific surface areas,
which was not observed for fumed silica with a water
probe (Brendle´et al., 2005), could be attributed to the fact
that the polar isopropanol probe was adsorbed only on the
most polar part of the surface. It could also be explained
by the more irreversible character of isopropanol adsorp-
tioncomparedtothatofn-octane,astestifiedbytheir
respective irreversible indexes (I
irr
) (Table 5). Of course,
this irreversible adsorption contributed to the decrease in
the measured specific surface areas computed from the
reversible part of the adsorption phenomenon. On the
other hand, the BET constant did not vary for n-octane
(C
C8
) and only very weakly for isopropanol (C
IP
).
Finally, the evolution of the specific surface areas,
independent of the probe, showed a decrease in the
internal porosity.
To corroborate these results, a study of the influence
of the grinding process on the heterogeneous character
of the studied solid was carried out by determining the
distribution functions of the adsorption energies of both
n-octane and isopropanol probes.
Study of surface heterogeneity.Asdescribedinthe
theoretical section above, the asymmetry of the distribu-
tion function of the adsorption energies (DFAE) can
describe the surface heterogeneity by means of an index
of heterogeneity (I
hete
). The values were determined for
Table 5. Comparison of specific surface area (S
C8,
S
IP
), BET constant (C
C8
,C
IP
), and heterogeneity (I
hete
) and irreversibility (I
irr
)
index values, measured by IGC-FC for n-octane and isopropanol probes, on attapulgite samples and for different grinding times.
N
2
———— n-octane ———— ———— Isopropanol ————
Grinding time
(min)
S
BET
(m
2
/g)
0.3
S
C8
(m
2
/g) 2.8
C
C8
I
irr
(%) 0.2
S
IP
(m
2
/g) 4.0
C
IP
I
irr
(%) 0.6
0 125.2 114.5 9 0.1 70.9 22 8.3
30 99.3 88.9 9 1.1 69.4 19 9.1
60 85.6 84.4 8.8 1.6 61.6 23 11.2
120 85.8 84.2 8.4 1.8 65.7 23 10.8
180 100.3 92.9 9 0.7 68.3 24 12.3
Table 4. Variation with dry grinding time (in a ball mill) of the surface energy dispersive component (g
s
d
), of the
nanomorphological index (I
M
), and of the specific parameter (I
sp
) values of attapulgite samples, determined by IGC-ID at
130ºC.
g
s
d
————— I
M
————— ——— I
sp
(kJ/mole) ———
Grinding time (mJ/m
2
) 4% 2,3,4 -TMP 0.03% Cyclo 8 0.01% CHCl
3
1% CH
2
Cl
2
1%
Initial 164 0.69 0.22 14 17
30 min 154 0.63 0.20 13 16
60 min 143 0.64 0.22 13 16
120 min 147 0.67 0.24 13 17
180 min 148 0.68 0.23 13 16
Vol. 58, No. 2, 2010 Surface properties of milled attapulgite 149

both n-octane and isopropanol probes (Table 6). As
expected, the index of heterogeneity, I
hete
,ofisopropa-
nol was much greater than that of n-octane, because the
former interacted strongly with the surface functional
groups through hydrogen bonds, i.e. with silanol groups,
whereas the latter exchanged only non-specific interac-
tions with the surface and was quite insensitive to the
surface functionality. The indices obtained with octane
or with isopropanol were quite independent of the
grinding times, and their variations were erratic.
The distribution functions of the adsorption energies
of n-octane and isopropanol probes on the ground
attapulgite samples were considered for increasing
milling times (Figure 6). For each probe used, n-octane
and isopropanol, the DFAE exhibited a similar shape
regardless of the milling time. Only one slightly
increased tailing in the domain of the high energies,
>28 kJ/mole for n-octane and >25 kJ/mole for iso-
propanol. For isopropanol, a weak shoulder at 28 kJ/
mole appeared progressively. Hence, the study of the
evolution of the surface heterogeneity showed only a
very slight change in the surface properties.
Computing the DFAE is based on a deconvolution
operation, which never leads to a single solution, due to
its ill-posed character. In fact, some changes in the shape
ofFDAEmaybeobservedfromoneoperatortoanother.
In order to test the operator influence, or, in other words,
to test the stability and robustness of the chosen
computing method, the DFAE were computed indepen-
dently by three of the co-authors. The main IGC-FC
characteristic parameters of both chromatograms (irrever-
sibility index, computed isotherm, specific surface area,
BET constant, DFAE, and index of heterogeneity),
obtained by the three different operators, were compared
(Table 7). The most important relative standard deviations
were observed for the index of heterogeneity, ~10%; less
for the specific surface areas, <4%; and for the BET
constants and irreversibility indexes with only one
exception on the relative standard deviation obtained on
the irreversibility index at 180 min of grinding time. The
values confirmed the good reproducibility given by the
computing method used to process the experimental
chromatograms for obtaining the main IGC-FC para-
meters. With respect to the reproducibility of the
computed DFAE, the distribution functions obtained by
the three different operators were compared (Figures 7
and 8). Again no significant differences were observed
between the FDRJ calculated by the three operators, either
for the n-octane probe or the isopropanol probe, testifying
again to the good stability of the deconvolution computing
process. The slight evolutions observed for IGC-CF
Table 6. Heterogeneity (I
hete
) index values, measured by
IGC-FC for n-octane and isopropanol probes, on attapulgite
samples for different grinding times.
Grinding time
(min)
I
hete
(C8)
(%) 1.2
I
hete
(IP)
(%)  4.9
0 13.6 41.2
30 14.6 34.5
60 12.3 37.4
120 16.1 38.0
180 14.3 34.5
Figure 6. Distribution functions of the adsorption energies of n-octane (left) and isopropanol (right) probes measured on a ttapulgite
samples for increasing grinding times.
150 L. Boudriche et al.Clays and Clay Minerals

parameters and DFAE with grinding were, therefore, not
significant. The IGC-FC analysis supported the fact that
the grinding process did not significantly modify the
surface heterogeneity of attapulgite.
CONCLUSIONS
Particle-size analysis showed that 30 min of grinding
time was sufficient to reach a limiting value in particle
size of ~25 mm. During the first hour of grinding, the
specific surface area measured with different probes,
nitrogen, octane, and isopropanol, decreased, in accord
with the disappearance of the internal structure, and the
interpretation was supported by a slight decrease in the
dispersive component of the surface energy determined
by IGC-ID, due to the disappearance of adsorption sites.
The IGC-FC method provided information on the
influence of grinding on the heterogeneous character of
the attapulgite surface. The advantage of IGC-FC over
the other adsorption techniques, such as nitrogen
adsorption, is its ability to test a solid surface with
different organic probes, apolar or polar, basic or acid, in
order to quantify the influence of a treatment (e.g.
grinding) on the surface functionality. The modifications
observed with grinding were weak from the point of
view of surface heterogeneity and the interaction
capacity of attapulgite, as shown by the stability of the
distribution functions for both octane and isopropanol.
The weak sensitivity of all the IGC measurements to
grinding was shown to be related to the particular
Table 7. Main IGC-FC parameters of both n-octane and isopropanol probes, computed by three different operators for two
attapulgite samples, initial and after grinding for 3 h.
————— Octane ————— ————— Isopropanol —————
Grinding time
(min)
Operator S
C8
(m
2
/g)
C
C8
I
hete
(%)
I
irr
(%)
S
IP
(m
2
/g)
C
IP
I
hete
(%)
I
irr
(%)
0
1 114 9 14 0.6 71 22 41 8
2 107 9 16 0.6 69 22 49 8
3 114 9 13 0.6 73 21 41 8
Mean value 111.7 9.0 14.3 0.6 71.0 21.7 43.7 8.0
Relative standard deviation 4% 0% 11% 0% 3% 3% 11% 0%
180
1 93 9 15 0.7 68 17 34 12
2 92 9 12 0.7 67 16 39 13
3 94 9 14 0.7 70 17 35 10
Mean value 93.0 9.0 13.7 0.7 68.3 16.7 36.0 11.7
Relative standard deviation 1% 0% 11% 0% 2% 3% 7% 13%
Figure 7. Distribution functions of the adsorption energies of the n-octane probe computed by three different operators for two
attapulgite samples, initial (left) and after grinding for 3 h (right).
Vol. 58, No. 2, 2010 Surface properties of milled attapulgite 151

structure of attapulgite clay, which is made of micro-
porous fibers. New surfaces, resulting from the fiber-
breaking process, could not contribute to modification of
the surface heterogeneity.
REFERENCES
Balard, H. (1997) Estimation of the surface energetic hetero-
geneity of a solid by inverse gas chromatography. Langmuir,
13, 12601269.
Balard, H., Saada, A., Hartmann, J., Aouadj, O., and Papirer,
E. (1996) Estimation of the surface energetic heterogeneity
of fillers by inverse gas chromatography. Makromolekulare
Symposium,108,6380.
Balard, H., Aouadj, O., and Papirer, E. (1997) Monitoring, by
inverse gas chromatography, of the variation of the surface
energetic heterogeneity of ground muscovite samples.
Langmuir,13, 12511255.
Balard, H., Brendle´ , E., and Papirer, E. (2000) Determination
of the acid-base properties of solid surfaces using inverse
gas chromatography: Advantages and limitations. Pp.
299316 in: Acid-Base Interactions: Relevance to
Adhesion Science and Technology (K. Mittal, editor). VSP,
Utrecht, The Netherlands.
Balard, H., Maafa, D., Santini, A., and Donnet, J.B. (2008)
Study by inverse gas chromatography of the surface
properties of milled graphites. Journal of Chromatography
A,1198, 173180.
Bradley, W.F. (1940) The structural scheme of attapulgite.
American Mineralogist,25,405410.
Brendle´ , E. and Papirer, E. (1997) A new topological index for
molecular probes used in inverse gas chromatography for
the surface nanorugosity evaluation. Journal of Colloid and
Interface Science,194, 207216.
Brendle´ , E., Ozil, F., Balard, H., and Barthel, H. (2005)
Adsorption of water on fumed silica. Pp. 888894 in:
Organosilicon Chemistry VI From Molecules to Materials
(N. Auner and J. Weiss, editors). Wiley, Weinheim,
Germany.
Chahi, A., Petit, S., and Decarreau, A. (2002) Infrared
evidence of dioctahedral-trioctahedral site occupancy in
palygorskite. Clays and Clay Minerals,50,306313.
Conder, J.R. and Young, C.L. (1979) Physicochemical
Measurements by Gas Chromatography. Wiley
Interscience, New York, pp. 385390.
Dorris, G.M. and Gray, D.G. (1979) Adsorption, spreading
pressure and London force interactions of hydrocarbons on
cellulose and wood fiber surfaces. Journal of Colloid and
Interface Science,71,93104.
Gala´ n, E. (1996) Properties and applications of palygorskite-
sepiolite clays. Clay Minerals,31,443453.
Gala´ n, E. and Carretero, I. (1999) A new approach to
compositional limits for sepiolite and palygorskite. Clays
and Clay Minerals,47,399409.
Garcı´a-Romero, E., Suarez Barrios, M., and Bustillo Revuelta,
M.A. (2004) Characteristics of Aa Mg-palygorskite in
Miocene rocks, Madrid Basin (Spain). Clays and Clay
Minerals,52,484494.
Hrachova, J., Komadel, P., and Fajnor, V.S. (2007) The effect
of mechanical treatment on the structure of montmorillonite.
Materials Letters,61, 33613365.
Mckeown, D.A., Post, J.E., and Etz, E.S. (2002) Vibrational
analysis of palygorskite and sepiolite. Clays and Clay
Minerals,50,667680.
Metraux, C., Grobety, B., and Ulmer, P. (2002) Filling of
chrysotile nanotubes with metals. Journal of Materials
Research,17,11291135.
Murray, H.H. (2000) Traditional and new applications for
kaolin, smectite, and palygorskite: A general overview.
Applied Clay Science,17, 207221.
Papirer, E., Roland, P., Nardin, M., and Balard, H. (1986)
Variation of the surface energy characteristics of mica
(muscovite) upon grinding. Journal of Colloid and Interface
Science,113,6266.
Roland, P. (1986) Contribution a`l’e´tude du broyage des
amiantes et du mica muscovite. Ph.D. thesis, Universite´de
Haute Alsace, France, 55 pp.
Rudzinski, W., Jagiello, J., and Grillet, Y. (1982) Physical
adsorption of gases on heterogeneous solid surfaces:
Figure 8. Distribution functions of the adsorption energies of the isopropanol probe computed by three different operators for two
attapulgite samples, initial (left) and after grinding for 3 h (right).
152 L. Boudriche et al.Clays and Clay Minerals

evaluation of the adsorption energy distribution from
adsorption isotherms and heats of adsorption. Journal of
Colloid and Interface Science,87,478491.
Saada, A. (1995) Origine des diffe´rences de proprie´te´s de
surface responsables des contrastes de mouillabilite´des
mine´ raux argileux des gisements pe´troliers. PhD thesis,
Universite´ de Haute-Alsace, 1995, France, 143 pp.
Saada, A., Papirer, E., Balard, H., and Siffert, B. (1995)
Determination of the surface properties of illites and
kaolinites by inverse gas chromatography. Journal of
Colloid and Interface Science,175,212218.
Suarez, M. and Garcı´a-Romero, E. (2006) FTIR spectroscopic
study of palygorskite: influence of the composition of the
octahedral sheet. Applied Clay Science,31, 154163.
Suraj, G., Iyer, C.S.P., Rugmini, S., and Lalithambika, M.
(1997) The effect of micronization on kaolinites and their
sorption behaviour. Applied Clay Science,12,111130.
(Received 11 April 2009; revised 29 January 2010;
Ms. 305; A.E. P. Malla)
Vol. 58, No. 2, 2010 Surface properties of milled attapulgite 153