Journal of Colloid and Interface Science 281 (2005) 316324

www.elsevier.com/locate/jcis

Aggregation of hydrophobically modified polysaccharides in solution

and at the airwater interface
Widad Henni a , Michel Deyme a, , Michel Stchakovsky b , Didier LeCerf c , Luc Picton c ,
Vronique Rosilio a
a Physico-Chimie des Surfaces, UMR CNRS 8612, Universit Paris-Sud, 5 Rue Jean-Baptiste Clment, 92296 Chtenay-Malabry Cedex, France
b Jobin Yvon S.A.S., Z.A. de la Vigne-aux-loups, 5 avenue Arago, 91380 Chilly-Mazarin Cedex, France
c Polymres-Biopolymres-Membranes, UMR CNRS 6522, Universit de Rouen, F76821 Mont-Saint-Aignan Cedex, France

Received 17 March 2004; accepted 11 August 2004

Available online 7 October 2004

Abstract
The present article focuses on the comparative study of physicochemical properties of two ionic pullulan derivatives modified by 10 or
35 C8 chains per 100 anhydroglucose units, named CMP10 C8 and CMP35 C8 , respectively. In aqueous solutions, these derivatives exhibited
an associative behavior as evidenced by pyrene fluorescence spectroscopy. This phenomenon, which stems from intra- and/or intermolecular
interactions between the hydrophobic groups grafted on the polymer backbone, results in the formation of more or less condensed aggregates
depending on the C8 ratio and the ionic strength of the media. The hydrophobically modified pullulans also displayed surface properties.
Their adsorption at the airsolution interface was assessed from surface tension measurements. The results showed that both hydrophobized
polymers adsorb in a coil conformation occupying a large interfacial molecular area. The comparison of these molecular areas indicated that
CMP35 C8 adopts a more shrunken conformation at the interface than CMP10 C8 , due to stronger intramolecular interactions. The stability
of the adsorbed monolayer under bulk dilution was investigated by ellipsometric measurements. Whereas bulk dilution had no effect on the
stability of the adsorbed CMP35 C8 film, it provoked significant changes in the adsorbed CMP10 C8 monolayer. The stability of the CMP35 C8
monolayers was attributed to the existence of intermolecular associations between the adsorbed coils.
2004 Elsevier Inc. All rights reserved.
Keywords: Hydrophobized carboxymethylpullulan; Monolayers; Pyrene fluorescence; Surface tension; Ellipsometry

1. Introduction
Hydrophobically modified polymers (HM polymers) are
amphiphilic water-soluble macromolecules mainly constituted of a hydrophilic backbone and hydrophobic side
groups. Interest in this class of macromolecules has grown
since their conception about 50 years ago [1], because of
their unique associative behavior and their potential applications in pharmaceuticals, cosmetics, paints and coatings, oil
recovery, etc. [2].
* Corresponding author. Fax: +33(0)146835312.

E-mail address: michel.deyme@cep.u-psud.fr (M. Deyme).

0021-9797/$ see front matter 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcis.2004.08.094

The physicochemical properties of HM polymers depend not only on the structural parameters of the polymer
such as the nature of the macromolecular backbone and
the length and rate of hydrophobic moieties but also on
environmental parameters (pH, salinity, temperature, etc.).
There are many synthetic materials based on charged backbones: polyacrylate derivatives (HMPA) [3,4], or neutral
skeletons, such as ethylene oxide urethane (HEUR), polyethylene oxide (POE) [5], and polyvinyl alcohols (PVA) [6].
Among these water-soluble polymers, biopolymers are particularly attractive because of their biocompatibility and low
toxicity. Landoll [7] was among the first authors to report
on the viscosimetric and surface-active properties of a series of nonionic polysaccharide derivatives with different

W. Henni et al. / Journal of Colloid and Interface Science 281 (2005) 316324

hydrophilic backbones (methylcellulose, hydroxyethylcellulose, and hydroxypropylcellulose) grafted by various alkyl

groups ranging from C12 to C24 . Other biopolymers have
been hydrophobically modified, such as carboxymethylcellulose [8], hyaluronan [9], dextran [10], pullulan [11], alginate [12], carrageenan [13], starch [14], or chitosan [15].
Interesting physicochemical characteristics of the hydrophobically modified polymers are related to both intra- and/or
intermolecular associations between their hydrophobic segments in aqueous solution, which would lead to the formation of more or less aggregated structures containing hydrophobic microdomains. The conformation of these polymers in solution may be studied by fluorescence techniques
using probes such as pyrene, which has the ability to interact with hydrophobic domains in aqueous solutions. In
1976, Kalayanasundaram and Thomas used the I3 /I1 ratio
(ratio of the intensities of the third and first peaks in the
pyrene emission spectrum) to monitor the effects of pyrene
microenvironment on its fluorescence and to determine the
critical micellar concentration of various surfactants [16]. In
aqueous solutions of HM polymers, pyrene is preferentially
localized in the hydrophobic microdomains. The changes in
its fluorescence spectrum provide information on the conformation of the macromolecules in solution [1719]. This
is of particular interest because hydrophobic microdomains
in a soluble polymer may possess properties similar to those
of surfactant micelles, such as the solubilization of waterinsoluble enzymes or drugs [20].
Due to their amphiphilic structure, HM polymers also
have potential high surface and interfacial properties. They
diffuse through the bulk phase and adsorb at the interface,
inducing a sharp reduction in the surface or interfacial tension of a polymer solution [21,22]. Surface tension studies
have evidenced various structures at the airwater interface
depending on the concentration and the characteristics of
these polymers. Indeed, adsorption at the airwater interface strongly depends on macromolecular architecture such
as the stiffness of the backbones and the charge density
for polyelectrolytes. For example, the high stiffness of a
macromolecular skeleton would limit the number of contacts of hydrophobic moieties with the surface [23]. Also, the
chemical nature and the rate and length of the hydrophobic
groups would affect the surface activity of these surfactants
[2430]. HM polymers can adopt variable conformations
at the airsolution interface: (i) Extended: the hydrophobic
groups are exposed to the air phase while the hydrophilic
moiety is completely immersed in the aqueous phase. They
form brushlike structures at the airsolution interface, especially at high surface coverage. De Gennes observed that the
adsorbed end-capped poly(ethylene oxide) possess a mushroom conformation at low surface coverage and transform
to brushes at high coverages [31]. (ii) Shrunken: the hydrophobic groups are buried inside micelle-like aggregates
and are thus unavailable for polymer adsorption at the interface [32].

317

In this work, we focus on hydrophobically modified carboxymethylpullulan derivatives. Pullulan is a flexible and
linear polysaccharide. Its structure is widely described as
-1,4 linked glucose units, which are included in -1,6 maltotriose units [33]. Nevertheless, a small proportion of -1,3
linkages has also been evidenced [34]. Akiyoshi et al. have
reported on the properties of hydrophobized nonionic pullulan that bears cholesteryl moieties. They have shown that
those polymers have the ability to form stable nanoparticles in water by self-association [35,36]. An interesting feature of the cholesterylpullulan nanoparticles is their ability to form complexes with hydrophilic [3739] as well as
with hydrophobic substances [40]. In previous work, we
have studied the interfacial properties of neutral cholesteryl
pullulan [4143]. Surface tension and surface potential measurements revealed that this modified polysaccharide adopts
an extended conformation at the interface in which its sugar
moieties are completely immersed in the aqueous phase,
whereas the hydrophobic cholesteryl groups are exposed to
the air phase. The aim of the present work was to study the
physicochemical properties of two new ionic pullulan derivatives, obtained by carboxymethylation followed by grafting
of C8 alkyl chains onto their backbones. In an attempt to
formulate an efficacious drug delivery system, the ability
of these HM polymers to form self-assembling structures,
allowing solubilization of hydrophobic molecules such as
pyrene was determined. The interfacial behavior of the polymers and the stability of the resulting monolayers were also
studied.

2. Experimental
2.1. Materials
The hydrophobically modified carboxymethylpullulan
derivatives were synthesized in two steps according to the
previously described procedures [44,45], starting from pullulan (produced by Aureobasidium pullulans, Hayashibara
Biochemical Laboratory, Japan). First, the sodium salt of
carboxymethypullulan (CMP, Na+ ) was synthesized in water/isopropanol medium by reacting the hydroxyl groups of
pullulan with sodium chloracetate in the presence of sodium
hydroxide. This synthesis has been described in detail by
Bataille et al. [44]. The degree of substitution (DS) in carboxymethyl groups of CMP was determined by conductometric titration according to the Eyler method [46]. The DS
is defined as the number of carboxymethyl groups per anhydroglucose unit (AGU) and may vary from 0 to 3, since
there are three sites for chemical modification on each AGU.
For the studied CMPs, DS of 0.92 has been found. NMR
studies of CMPs (Glinel et al. [47]) have shown that the
substitution at C2 carbon is predominant and decreases according to the order C2 > C3 > C6 > C4. The synthesis
of the hydrophobically modified CMPs was adapted from
the procedure described by Della Valle [48] and Fisher

318

W. Henni et al. / Journal of Colloid and Interface Science 281 (2005) 316324

Table 1
Characteristics of the hydrophobically modified carboxymethylpullulan
derivatives
Pullulana
CMP
CMP10 C8
CMP35 C8

M n (g/mol) 10%

M w (g/mol) 10%

170,000
160,000
140,000
140,000

345,000
300,000
400,000
570,000

a From Ref. [44].

et al. [49] for other HM polymers and it was detailed in

a previous work (Duval-Terrie et al. [11]). The modified
polymers are referred to as CMP -C8 , with being the
hydrophobic chain substitution rate (%) and 8 the number
of carbons in the grafted moiety. The octyl pendant groups
are covalently bound to the carboxymethyl groups via ester links. Their percentage ( ) is expressed as the number
of hydrophobic moieties for 100 AGU. has been determined by both 13 C NMR measurements and quantification
of alcohols release after total alkaline hydrolysis of hydrophobically modified carboxymethylpullulans (HMCMP)
using gas phase chromatography according to the method
described by Duval-Terrie et al. [11]. Table 1 summarizes the average molar masses of the pullulan precursor,
CMP, and the studied HMCMP derivatives. Average molar masses have been determined by SEC/MALLS (pullulan
and CMP) and F4/MALLS (HMCMP) measurements. Description and relevance of these physicochemical techniques
have been previously discussed [11,50]. One should note
that Bataille et al. [44] modified the CMP synthesis in order to avoid the macromolecule degradation observed in
previous works [51]. This is evidenced in Table 1: the average molar masses of pullulan and of CMP are very similar.
HMCMPs present number-average molar masses (M n ) close
to that of the CMP precursor, whereas the weight-average
molar masses (M w ) of the hydrophobically modified samples are larger than that of CMP. This is due to the sensitivity
of M w to small proportions of large species. Consequently,
the observed increase in the M w value accounts for the aggregation of the amphiphilic polymer in aqueous solutions
due to hydrophobic interactions. This phenomenon has been
largely reported and discussed in previous papers [39,50].
Polymer solution samples were prepared by dissolving
the appropriate amount of polymer in phosphate buffer
(KH2 PO4 , KOH) or in water. Buffer pH was chosen equal
to 7.5, corresponding to the full ionization of carboxylate
groups [52]. The pH of polymer solutions prepared in phosphate buffer remained constant as the polymer concentration
increased. The polymer solutions were then stirred at room
temperature for at least 48 h. Before the experiments, the
polymer solutions were diluted to obtain samples with concentrations ranging from 14 to 1.4 104 g/L. The water
was de-ionized using a Milli-Ro 6+ system (Millipore) and
then doubly distilled on acid KMnO4 before experiments.
This water had a surface tension of 71.8 0.1 mN/m at
25 C. KH2 PO4 and KOH were purchased from Merck and

KMnO4 from Aldrich. Highly purified pyrene (>99%) was

obtained from Fluka and used as received. The glassware
was cleaned in a freshly prepared sulfochromic solution and
abundantly rinsed with the ultrapure water.
2.2. Methods
2.2.1. Fluorescence spectroscopy
The experiments were performed using a Spex-Fluorog
1681-0.22m spectrometer. In these experiments, a pyrene
stock solution (103 M) was prepared in acetone. A sample of 10 l of this solution was introduced into empty
vials and the solvent was evaporated under vacuum. After
evaporation, the vials were filled with 10 ml of polymer
solution and gently stirred for 18 h to ensure the incorporation of the molecular probe into polymer hydrophobic microdomains. The final pyrene concentration was 106 M. At
this low concentration, no excimer band due to the interaction of an excited state pyrene with a ground state pyrene
was observed [17,19,53]. Such a low concentration was chosen to minimize the influence of pyrene on the formation
and/or the stability of hydrophobic microdomains. All samples were excited at 335 nm and the emission spectra of
pyrene showed vibronic peaks at 1 = 372 nm (intensity I1 )
and 3 = 382 nm (intensity I3 ).
2.2.2. Surface tension measurements
Surface tension was measured by the Wilhelmy plate
method using a K10 tensiometer (Krss, Germany) or a
Sanborn tensiometer associated with an electronic Hewlett
Packard unit. The measurements were made without detaching the plate from the interface and the data were continuously plotted on a chart recorder for 24 h. In order to maintain a constant level of the liquid and avoid any drift in the
measured surface tensions, all experiments were performed
at 251 C under satured vapor pressure to limit water evaporation during the experiments. The reported surface tension
values are mean values of at least two measurements. The
surface pressures values were deduced from = 0 ,
where 0 is the surface tension of the pure water and is the
surface tension of a solution.
2.2.3. Ellipsometric measurements
Ellipsometric measurements were performed to compare
the stability of the HMCMPs monolayers under bulk dilution. Such experiments could not be achieved by the Wilhelmy plate technique, which requires a constant level of
the liquid. Ellipsometry measures optical properties of interfacial thin layers by measuring the change of polarization
between incident and reflected beams. The two ellipsometric
angles (psi) and  (delta) are linked to the complex ellipticity, , and to the ratio Fresnel reflection coefficients, rp
and rs , where rp is the ratio of the electric field amplitudes
parallel to the incident plane after and before reflection of
light, whereas rs refers to the ratio for light with the electric

W. Henni et al. / Journal of Colloid and Interface Science 281 (2005) 316324

field perpendicular to the plane of incidence:

= tan ei .
At the Brewster angle, the complex ellipticity is very sensitive to any structure present at the interface [5456].
Ellipsometric measurements were performed using the
commercial instrument UVISEL, manufactured by JobinYvon (France). It is a spectroscopic phase-modulated ellipsometer. The spectral range was 250800 nm, the spot size
1 mm2, and the angle of incidence 53.26. All measurements
were made at the airliquid interface.
In our experiments, HMCMP solutions (polymer concentration C = 9.2 g/L and ionic strength I = 0.06 M) were
poured into measurement cells and left to stabilize for 24 h
at room temperature, allowing the adsorption of the polymer. Then a given volume of buffer was slowly injected into
the subphase to dilute the solution to two-thirds of its initial
concentration. Following bulk dilution, the ellipsometric parameters were taken three times, at 5-min intervals, in order
to compare the stability of the HMCMP monolayers under
bulk dilution.
3. Results
3.1. Incorporation of pyrene in polymer aggregates
in solution
3.1.1. Effect of the alkyl substitution rate (CMP10 C8
and CMP35 C8 )
The fluorescence of pyrene at a constant concentration
(106 M) was measured in aqueous solutions of CMP10 C8
and CMP35 C8 . Fig. 1 illustrates the variation of the I1 /I3
ratio with the logarithm of polymer concentration. At low
polymer concentrations, the I1 /I3 ratio was about 1.67,
a value similar to that measured in an aqueous environment (106 M of pyrene in water or in phosphate buffer).
As polymer concentration increased, a decrease in I1 /I3

Fig. 1. Variation of I1 /I3 versus log C for CMP35 C8 and CMP10 C8 in

0.1 M phosphate buffer.

319

was observed, revealing an increasingly apolar local microenvironment of the pyrene. This phenomenon suggests
that hydrophobic intra- and/or intermolecular interactions
between the HMCMP alkyl groups induce the formation
of hydrophobic microdomains in which pyrene is solubilized. Apparently, as the polymer concentration increases,
the pyrene is preferentially incorporated into the hydrophobic microdomains until free pyrene concentration in the
surrounding water becomes negligible. A similar partition
of the molecule has already been reported by Petit-Agnely
et al. [53].
Fig. 1 also shows the effect of the alkyl chain content
on the I1 /I3 ratio, which decreases to a higher extent with
CMP35 C8 than with CMP10 C8 . The probability of finding
two alkyl chains next to one another becomes greater as
they grow more numerous, favoring hydrophobic interactions and thus, the formation of dense polymer aggregates.
The higher I1 /I3 ratio observed for CMP10 C8 compared to
CMP35 C8 may indicate that the hydrophobic core of the former is less compact than that of the latter. This could be due
to water penetration into these aggregates extending up their
cores [17]. It is noticeable that the effect of the alkyl substitution rate, reported here for solutions in 0.1 M phosphate
buffer, was also observed for the other ionic strengths reported thereafter.
3.1.2. Effect of the ionic strength
Fig. 2 displays the variation of the I1 /I3 ratio with the
HMCMP concentration in water and in phosphate buffer at
three ionic strengths (0.01, 0.1, and 0.5 M). In the very dilute
regime (C = 1.4 104 g/L), both systems exhibit I1 /I3
values close to 1.67, similar to that of pyrene in an aqueous environment. Apparently at this polymer concentration,
there is no aggregate formation.
For CMP10 C8 , a slight decrease in the I1 /I3 ratio is observed at a polymer concentration of 1.4 g/L (Fig. 2a).
This suggests that pyrene molecules are transferred from
the aqueous phase to a less polar environment (i.e., the hydrophobic cores of micelle-like aggregates). As the ionic
strength increases, the I1 /I3 ratio decreases, revealing the
contraction of the macromolecular chain due to the progressive screening of the charged carboxylic groups located
along the polymer chain, which electrostatic repulsions favored an extended conformation. However, the moderate
I1 /I3 decrease suggests that although CMP10 C8 is able to
form micelle-like structures in solution, pyrene solubilization is weak. This is probably due to the low substitution
rate of this polymer, insufficient to induce strong hydrophobic interactions, and thus leading to the formation of coils
with large cores.
For CMP35 C8 , the decrease in the I1 /I3 ratio appears
at lower polymer concentrations than for CMP10 C8 (C =
1.4 102 g/L) (Fig. 2b). This is in agreement with the
results reported in Fig. 1, which show that CMP35 C8 has
a more pronounced hydrophobic character than CMP10 C8 .
This result is typical of HM polymers: the higher the num-

320

W. Henni et al. / Journal of Colloid and Interface Science 281 (2005) 316324

Fig. 2. Variation of I1 /I3 versus log C for (a) CMP10 C8 and (b) CMP35 C8 in 0.01, 0.1, and 0.5 M phosphate buffer.

(a)

(b)

(a)

(b)

Fig. 3. (a) Surface pressure versus time for various CMP10 C8 concentrations in 0.1 M phosphate buffer; (b) surface tension versus log C for
CMP10 C8 in 0.1 M phosphate buffer.

Fig. 4. (a) Surface pressure versus time for various CMP35 C8 concentrations in 0.1 M phosphate buffer; (b) surface tension versus log C for
CMP35 C8 in 0.1 M phosphate buffer.

ber of hydrophobic groups grafted to a macromolecule, the

greater its ability to self-associate in solution. From the results plotted in Fig. 2b, it appears that (i) the I1 /I3 ratio
decreases when the salt concentration increases from 0.01
to 0.1 M. The addition of salt seems to favor alkyl chain interactions by screening the negative repulsive charges along
the polymer backbone; (ii) conversely, from 0.1 to 0.5 M, the
I1 /I3 ratio increases, revealing a higher polarity of pyrene
environment.

3.2. Interfacial behavior of the hydrophobized polymers

3.2.1. Effect of the alkyl chain substitution rate (CMP10 C8
and CMP35 C8 ) on the surface pressure
Figs. 3 and 4 show the evolution of surface pressure versus time and polymer concentration of CMP10 C8 (Fig. 3a)
and CMP35 C8 (Fig. 4a) in 0.1 M phosphate buffer. Both
components exhibited surface properties and lengthy periods were necessary to attain steady state values of the surface

W. Henni et al. / Journal of Colloid and Interface Science 281 (2005) 316324

321

Fig. 5. Surface pressure versus time for 1.4 g/L solutions of (a) CMP35 C8 and (b) CMP10 C8 in water and in 0.01, 0.1, and 0.5 M phosphate buffer.

pressure. However, it seems that the increase in the substitution rate of a polymer contributed to its faster adsorption at
the interface. The CMP35 C8 adsorption at the interface appeared to be a fast process. The surface pressure increased
with increasing polymer concentration and this was slightly
more pronounced for CMP35 C8 than for CMP10 C8 . It is interesting to note that CMP35 C8 , despite its higher hydrophobicity, showed only a slightly higher ability to increase the
surface pressure than CMP10 C8 . At the highest studied concentration (14 g/L), the surface pressure increase was about
26 mN/m for CMP10 C8 and did not exceed 29 mN/m for
CMP35 C8 . The data also show that, for CMP10 C8 , the maximal surface pressure value was reached at a polymer concentration of 1.4 g/L. It may thus be assumed that at this
concentration the interface was already saturated.
3.2.2. Effect of the ionic strength on the surface pressure
The effect of the ionic strength on the surface properties
of the HMCMP was studied at a single polymer concentration (1.4 g/L).
For CMP35 C8 , the data reported in Fig. 5a show an effect of the ionic strength on the surface pressure, since
an increase in is observed when salt concentration increases from 0 (salt-free solution) to 0.1 M. Whereas the
surface pressure in water is about 17 mN/m, it increased to
22 mN/m at the ionic strength of 0.1 M. However, a further
addition of salt (0.5 M) did not modify the surface activity.
For CMP10 C8 , a strong dependence of the surface pressure on the addition of salt was observed (Fig. 5b). Whereas
the maximum surface pressure in water was 10 mN/m, it
reached 26 mN/m in the presence of salts. The higher number of free carboxylic groups in CMP10 C8 compared to
CMP35 C8 would account for the higher sensitivity of the
surface pressure of the former polymer solution to the addition of salts. It is well known that a neutral polymer has
a greater ability to adsorb at the airwater interface than a
charged one [57].
3.2.3. Effect of the dilution of the subphase on the stability
of a formed HMCMP monolayer
Ellipsometric measurements were performed to evaluate
the stability of the HMCMP adsorbed monolayers. The el-

(a)

(b)
Fig. 6. Ellipsometric data for (a) CMP35 C8 and (b) CMP10 C8 solutions in
0.066 M phosphate buffer, C = 9.2 g/L (1, 2, 3: first, second, and third
measurements).

lipsometric parameters ( and ) are presented versus the

wavelengths in Fig. 6.
The results reveal that, for CMP35 C8 solutions, and 
remained unchanged under solution dilution (Fig. 6a). The
dilution process did not cause any significant variation in the
thickness nor the density of the CMP35 C8 adsorbed layer.
This would indicate that this polymer is strongly anchored
at the airsolution interface and forms a stable layer.
Conversely, for CMP10 C8 , the ellipsometric parameters
and  appeared strongly modified indicating that changes

322

W. Henni et al. / Journal of Colloid and Interface Science 281 (2005) 316324

in the organization of the film at the interface took place under dilution. A decrease in and change in the slope was
observed for  after each measurement (Fig. 6b), revealing
a thinner and less dense adsorbed layer.

4. Discussion
Independent of the composition of the polymer solution,
the surface tension measurements showed that the maximum
surface pressure did not exceed 30 mN/m and that these values were very similar for the two studied HM polymers. This
was rather surprising because one would expect the surface
pressure increase to be proportional to the alkyl chain substitution rate [58]. Zhang and Marchant [28] have shown
that the adsorption of a maltose substituted by a C8 chain
can lead to a maximum surface pressure of 40.8 mN/m.
This surface pressure value is much higher than those measured in our experiments. This suggests that the presence of
a macromolecule in the subphase limits the surface tension
lowering induced by C8 chains. This is probably related to
the impossibility of the adsorbed alkyl side chains forming
dense packing at the interface because of the steric hindrance
originating from the local stiffness of the macromolecular
backbone.
By plotting the equilibrium surface tension of the HMCMP samples (in 0.1 M phosphate buffer) versus their logarithmic concentration (Figs. 3b and 4b), we have calculated
the apparent interfacial excess concentration using the
simplified Gibbs equation. The Gibbs equation was applied
in the concentration range in which the surface tension decreases linearly, with the slope equal to d /d ln C:



d
1
.
=
(1)
RT d ln C
In these calculations, it was assumed that, at low HMCMP
concentrations, the activities could be replaced by their concentrations, since the second virial coefficient as determined
by F4/MALLS measurements was negative [59]. The corresponding molecular areas were then deduced from the equation
A = N/,

(2)

where N is Avogadros number.

In Fig. 3b for CMP10 C8 , the inflection point corresponding to surface saturation is clearly evident. Conversely, it is
not observed for CMP35 C8 (Fig. 4b). Indeed, for the latter,
the surface tension corresponding to surface saturation could
not be measured accurately due to the high solution viscosity.
The calculated molecular areas (from the equation =
0.43/RT (d /d log C)) were respectively 97 and 107 2 for
CMP35 C8 and CMP10 C8 . Compared to the value described
by Zhang and Marchant [28] for maltose surfactants (40 2 ),
the molecular area of the studied polymers was 2.3 to 2.7
times larger for CMP35 C8 and CMP10 C8 , respectively. It is

clear that the molecular areas of the studied polymers do not

correspond solely to the molecular area of the hydrophobic
anchors and that they are related to the conformation of the
polysaccharide in the subphase. In our previous studies of
cholesterylpullulans, we have shown that when the substitution rate is low, the polymer adsorbs at the airwater
interface, in an extended conformation with the cholesteryl
groups oriented toward the air phase and the polysaccharide moiety forming long loops immersed in the solution. In
such a situation a brushlike conformation is observed at the
airsolution interface and the interfacial molecular area corresponds to that of a single cholesterol molecule [42]. As the
substitution rate increases, a greater number of cholesteryl
moieties is adsorbed at the airwater interface [43]. Unpublished data from our laboratory show that carboxymethylpullulan derivatives substituted with 4 and 7% of hexadecyl and
tetradecyl chains, respectively, adopt the same interfacial
conformation than the cholesteryl pullulans. This is not surprising since when the hydrophobic substitution rate is low,
the distance between the neighboring alkyl chains anchored
at the interface is large enough to enable the formation of
so-called loops in the aqueous phase, and thus a dense
packing of the alkyl chains at the interface.
Contrary to those polymers, CMP35 C8 and CMP10 C8 exhibited large molecular areas compared to that of a single
maltose surfactant [28]. The obtained results may be explained if one considers that for these CMPs with short pending chains, the substitution rate is high enough to prevent
formation of polymer loops. Such an interfacial behavior
has been described previously [32,23]. This would confer
a local stiffness on the macromolecular backbone and limit
the ability of all the C8 side chains to reach the interface.
Consequently, the adsorbed polymers adopt a shrunken conformation, which is more or less condensed depending on
the C8 substitution rate (Fig. 7). This phenomenon leads
to the enlargement of the interfacial molecular area, compared to C8 chains bound to one free sugar moiety [28].
Using the approach described by Babak et al. [23,60], standard adsorption free energies (G) were calculated. They
amount to 29.2 and 24.2 kJ mol1 for CMP10 C8 and
CMP35 C8 , respectively. These results are similar to those obtained by Ros et al. [29] for a polyelectrolyte substituted by
C8 chains (23 kJ mol1 ). They may be interpreted in terms
of a deeper immersion of the CMP10 C8 alkyl chains into the
air phase compared to those of CMP35 C8 . This could be related to two main factors: (i) the high stiffness of CMP35 C8
backbone due to the important substitution rate and (ii) the
involvement of a greater number of CMP35 C8 alkyl chains
in the formation hydrophobic microdomains. The adsorption
at the interface of CMP35 C8 is hindered compared to that of
CMP10 C8 .
In solution, hydrophobic interactions induce the formation of polymer aggregates as evidenced by pyrene fluorescence spectroscopy. The formation and structure of these
aggregates depend to a large extent on the substitution rate of
carboxylic groups [61] and are thus influenced by the ionic

W. Henni et al. / Journal of Colloid and Interface Science 281 (2005) 316324

323

Fig. 7. Schematic representation and mean molecular areas of (a) CMP10 C8 and (b) CMP35 C8 adsorbed at the airwater interface.

strength of the media. From the data reported in Fig. 2b,

it may be assumed that the CMP35 C8 charges are progressively screened upon addition of electrolytes until the salt
concentration 0.1 M is reached. This concentration would
correspond to a total screening of the electrostatic charges.
On further salt addition, the observed increase in I1 /I3 suggests that the CMP35 C8 molecules conformation becomes
more expanded. Although still unexplained, such an increase
of macromolecular dimensions at high ionic strength has
been previously observed by viscosimetric and surface tension measurements for other hydrophobically modified carboxymethylcellulose derivatives [62,63]. The comparison of
CMP10 C8 and CMP35 C8 shows that the medium influences
the associative behavior differently, depending on the C8
chain substitution rate of the HM polymer. The high substitution rate of CMP35 C8 accounts for the reduction of the
number of free ionic groups by formation of ester bonds.
Therefore, its charges are fully neutralized at a lower ionic
strength than for CMP10 C8 . It is clear that there is a balance
between two effects: (i) the electrostatic repulsions between
the charges of the polymer chain, which would contribute to
the formation of expanded aggregates, and (ii) the attractive
hydrophobic interactions between alkyl groups that would
account for the formation of dense aggregates with well defined hydrophobic microdomains. Similar behavior of HM
polymers has been previously reported [44,50]. Moreover,
for a given ionic strength of the solution, an increase in the
substitution degree of the polymer would lead to a denser
structure of the aggregates. The kinetics of surface pressure
(Figs. 3a and 4a) are in good agreement with this hypothesis
as, for equal number average molecular masses, CMP35 C8
adsorbs faster than CMP10 C8 .
The evolution of a monolayer after bulk dilution gives a
complementary insight into the interfacial polymer organization. Bulk dilution was aimed to destabilize the polymer
monolayer and thus, to evaluate the strength of lateral interactions between adsorbed coils. Whereas for CMP10 C8 ,
the ellipsometric angles varied, indicating either conformational changes (coil expansion) or polymer desorption, for
CMP35 C8 , they remained unchanged, accounting for the stability of the monolayer under bulk dilution.
It is reasonable to think that for CMP35 C8 , as the substitution rate is high, the C8 chains are not only anchored at the

interface and engaged into intramolecular hydrophobic interactions, but also involved into interaggregates associations
contributing to the stabilization of the CMP35 C8 monolayer
(Fig. 7b). The formation of such interactions between aggregates has also been evidenced in concentrated CMP35 C8
solutions. The F4/MALLS experiments clearly showed the
existence of two species in the CMP35 C8 aqueous solutions:
compact isolated macromolecular coils resulting from intramolecular interactions between the alkyl groups, and large
species due to intermolecular associations [50]. This phenomenon was not observed for CMP10 C8 at these concentrations.
In conclusion, the studied hydrophobically modified carboxymethylpullulan derivatives exhibit good associative and
interfacial properties, which strongly depend on their structural parameters, especially on the C8 substitution rate. From
all reported experiments, it appears clearly that increasing
the substitution rate from 10 to 35% induces a reinforcement of the hydrophobic interactions and thus, the formation of more compact and stable aggregates. The use of the
CMP35 C8 could thus be envisioned in systems where a hydrophobic compound has to be solubilized in an aqueous
phase.

Acknowledgments
The authors are grateful to J.-Y. Dugast (UMR CNRS
6522, France) for the synthesis of the HMCMPs, to Dr.
M. Ollivon (UMR CNRS 8612, France) for his help in carrying the fluorescence measurements, and to J.-P. Gaston
(Jobin-Yvon, France) for providing access to the ellipsometer.

References
[1] P. Strauss, E.J. Jackson, Polym. Sci. 6 (1951) 649.
[2] J.E. Glass (Ed.), Associative Polymers in Aqueous Media, ACS Symposium Series, vol. 765, American Chemical Society, Washington,
DC, 2000.
[3] F. Petit, R. Illiopoulos, R. Audebert, S. Szonyi, Langmuir 13 (1997)
4229.
[4] F. Petit-Agnely, R. Illiopoulos, J. Phys. Chem. 103 (1999) 4803.

324

W. Henni et al. / Journal of Colloid and Interface Science 281 (2005) 316324

[5] F. Liu, Y. Frere, J. Franois, Polymer 42 (2001) 2969.

[6] G.O. Yahia, S.K. Asrof Ali, M.A. Alnafaa, E.Z. Hamad, J. Appl.
Polym. Sci. 57 (1995) 343.
[7] L.M. Landoll, J. Polym. Sci. Polym. Chem. 20 (1982) 443.
[8] P.M. Van der Horst, J. G. Batelaan, B. Reuvers, NL patent PCTWO9424169, 1994.
[9] R. Geciova, A. Flarbani, F. Delben, G. Liut, R. Urnani, A. Cesaro,
Macromol. Chem. Phys. 196 (9) (1995) 2891.
[10] C. Rouzes, A. Durand, M. Leonard, E. Dellacherie, J. Colloid Interface
Sci. 253 (2002) 217.
[11] C. Duval-Terrie, J. Huguet, G. Muller, Colloids Surf. A Physicochem.
Eng. Aspects 220 (2003) 105.
[12] M. Rastello-De-Boisseson, M. Leonard, P. Hubert, P. Marchal, A. Stequert, C. Castel, E. Favre, E. Dellacherie, J. Colloid Interface Sci.
273 (1) (2004) 131.
[13] J.C. Jacquier, P. Macartain, K.A. Dawson, Prog. Colloid Polym.
Sci. 118 (2001) 127.
[14] K.B. Wesslen, B. Wesslen, Carbohydrate Polym. 4 (2002) 303.
[15] C. Esqueret, E. Bulher, Macromolecules 34 (15) (2001) 5287.
[16] K. Kalyanasundaram, J.K. Thomas, J. Am. Chem. Soc. 99 (1977)
2039.
[17] M.M. Amiji, Carbohydrate Polym. 26 (1995) 211.
[18] C. Damas, A. Brembilla, P. Lochon, F. Baros, M.L. Viriot, Polymer 10
(1995) 2095.
[19] A. Fisher, M.C. Houzelle, P. Hubert, M.A.V. Axelos, C. GeoffroyChapotot, M.C. Carr, M.L. Viriot, E. Dellacherie, Langmuir 14
(1998) 4482.
[20] V.P. Torchilin, J. Controlled Rel. 73 (2001) 137.
[21] A. Baszkin, W. Norde (Eds.), Physical Chemistry in Biological Interfaces, Dekker, New York, 2000, p. 155.
[22] S. Magdassi, A. Kamyshny, A. Baszkin, J. Dispers. Sci. Technol. 22
(2001) 313.
[23] V. Babak, I. Lukina, G. Vikhoreva, J. Desbrires, M. Rinaudo, Colloids
Surf. A Physicochem. Eng. Aspects 147 (1999) 139.
[24] H.E. Ros, J.S. Rojas, I.C. Gamboa, R.G. Barraza, J. Colloid Interface
Sci. 156 (1993) 388.
[25] A.F. Olea, C. Gamboa, B. Acevedo, F. Martinez, Langmuir 16 (2000)
6884.
[26] R.G. Barraza, A.F. Olea, F. Martinez, I. Ruiz-Tagle, J. Colloid Interface Sci. 261 (2003) 559.
[27] Z. Su, T.J. McCarthy, S.L. Hsu, H.D. Stidham, Z. Fan, D. Wu, Polymer 39 (1998) 4655.
[28] T. Zhang, R. Marchant, J. Colloid Interface Sci. 177 (1996) 419.
[29] H.E. Rios, J.S. Rojas, I.C. Gamboa, R.G. Barraza, J. Colloid Interface
Sci. 156 (1992) 388.
[30] G.O. Yahya, K.A. Asrof, K.Z. Hama, Polymer 37 (1996) 1183.
[31] P.G. De Gennes, Adv. Colloid Interface Sci. 27 (1987) 189.
[32] V.G. Babak, E.A. Skotnikova, I.G. Lukina, S. Pelletier, P. Hubert,
E. Dellacherie, J. Colloid Interface Sci. 225 (2000) 505.
[33] T. Kato, T. Okamoto, T. Takuda, A. Takahashi, Biopolymers 21 (1982)
1623.
[34] W. Sowa, A.C. Blackwood, G.A. Adams, Can. J. Chem. 41 (1963)
2314.

[35] K. Akiyoshi, S. Deguchi, N. Moriguchi, S. Yamaguchi, J. Sunamoto,

Macromolecules 26 (1993) 3062.
[36] K. Akiyoshi, J. Sunamoto, Supramol. Sci. 3 (1996) 157.
[37] K. Akiyoshi, T. Nishikawa, Y. Mitsui, T. Miyata, M. Kodama,
J. Sunamoto, Colloids Surf. A Physicochem. Eng. Aspects 112 (1996)
91.
[38] K. Akiyoshi, I. Taniguchi, H. Fukui, J. Sunamoto, Eur. J. Pharmacol.
Biopharmacol. 42 (1996) 286.
[39] K. Akiyoshi, S. Kobayashi, S. Shichibe, D. Mix, M. Baudys, S.W.
Kim, J. Sunamoto, J. Controlled Rel. 54 (1998) 313.
[40] K. Akiyoshi, S. Deguchi, H. Tajima, T. Nishikawa, J. Sunamoto, Proc.
Jpn. Acad. 71 (1995) 15.
[41] A. Baszkin, V. Rosilio, G. Albrecht, J. Sunamoto, J. Colloid Interface
Sci. 145 (1991) 502.
[42] B. Deme, V. Rosilio, A. Baszkin, Colloids Surf. B Biointerface 4
(1995) 357.
[43] B. Deme, V. Rosilio, A. Baszkin, Colloids Surf. B Biointerface 4
(1995) 367.
[44] I. Bataille, J. Huguet, G. Muller, G. Mocanu, A. Carpov, Int. J. Biol.
Macromol. 20 (1997) 179.
[45] L. Picton, G. Mocanu, D. Mihai, A. Carpov, G. Muller, Carbohydrate
Polym. 28 (1995) 131.
[46] R.W. Eyler, T.S. Klug, F. Siephuis, Anal. Chem. 19 (1947) 24.
[47] K. Glinel, J.P. Sauvade, H. Oulyadi, J. Huguet, Carbohydr. Res. 328
(2000) 343.
[48] F. Della Valle, European Patent 94400352-8 (1992).
[49] A. Fisher, M.C. Houzelle, P. Hubert, M.A.V. Axelos, C. GeoffroyChapotot, M.C. Carre, M.L. Viriot, E. Dellacherie, Langmuir 14
(1998) 4482.
[50] S. Simon, J.Y. Dugast, D. Le Cerf, L. Picton, G. Muller, Polymer 44
(2003) 7917.
[51] L. Picton, G. Mocanu, D. Mihai, A. Carpov, G. Muller, Carbohydr.
Polym. 28 (1995) 131.
[52] I. Bataille, Ph.D. thesis, Universit de Rouen, 1998.
[53] F. Petit-Agnely, I. Illiopoulos, R. Zana, Langmuir 16 (2000) 9921.
[54] R.M.A. Azzam, N.M. Bashara, Ellipsometry and Polarized Light,
North-Holland, Amsterdam, 1977.
[55] A. Hambardzumyan, V. Agui-Bghin, M. Daoud, R. Douillard, Langmuir 20 (2004) 756.
[56] V. Aroulmoji, V. Agui-Bghin, M. Mathlouthi, R. Douillard, J. Colloid Interface Sci. 276 (2004) 269.
[57] R.R. Netz, D. Adelman, Phys. Rep. 380 (2003) 1.
[58] C. Rouzes, A. Durand, M. Leonard, E. Dellacherie, J. Colloid Interface
Sci. 253 (2002) 217.
[59] L. Picton, L. Merle, G. Muller, Intern. J. Polym. Anal. Charact. 2
(1996) 103.
[60] V.G. Babak, G.A. Vikhoreva, I.G. Lukina, Colloids Surf. A Physicochem. Eng. Aspects 128 (1997) 75.
[61] F. Millet, M. Nedyalkov, B. Renard, P. Perrin, F. Lafuma, J.J. Benattar,
Langmuir 15 (1999) 2112.
[62] V. Rosilio, G. Albrecht, A. Baszkin, L. Merle, Colloids Surf. B Biointerface 19 (2000) 163.
[63] D. Charpentier, L. Merle, G. Muller, Cah. Rhol. 15 (1997) 251.