8106 Langmuir 2005, 21, 8106-8113

Articles
Adsorption of a Cationic Surfactant onto Cellulosic Fibers
I. Surface Charge Effects
Sabrine Alila,† Sami Boufi,*,† Mohamed Naceur Belgacem,‡ and
Davide Beneventi‡
LMSE, Faculté des Sciences de Sfax, BP 802-3018, Sfax, Tunisia, and LGP2-UMR5518,
Ecole Française de Papeterie et des Industries Graphiques, INPG, BP 65,
F-38402 St. Martin d’Hères, France

Received February 9, 2005. In Final Form: April 5, 2005

The adsorption of four cationic surfactants with different alkyl chain lengths on cellulose substrates was
investigated. Cellulose fibers were used as model substrates, and primary alcohol groups of cellulose
glycosyl units were oxidized into carboxylic groups to obtain substrates with different surface charges. The
amount of surfactant adsorbed on the fiber surface, the fiber ζ-potential, and the amount of surfactant
counterions (Cl-) released into solution were measured as a function of the surfactant bulk concentration,
its molecular structure, the substrate surface charge, and the ionic strength. The contribution of each of
these parameters to the shape of the adsorption isotherms was used to verify if surfactant adsorption and
self-assembly models usually used to describe the behavior of surfactant/oxide systems can be applied, and
with which limitations, to describe cationic surfactant adsorption onto oppositely charged cellulose substrates.

Introduction surfactant self-assembly at the solid-liquid interface has

been well admitted since it was first proposed by Gaudin
The adsorption of surfactants at the solid-liquid
and Fuerestenau,14,15 the exact shape of these structures
interface has been a subject of considerable interest and
is still a subject of debate.
plays an important role in controlling a variety of
interfacial processes such as flotation,1,2 paper manufac- During the past decade, various kinds of experimental
turing and recycling,3 flocculation and dispersion,1,2 soil methods such as AFM,16,17 NMR,18,19 ellipsometry,20
remediation,4 and adsorption of contaminants in water fluorescence,11,21 and FT-IR22,23 have been used to give
treatment.4,5 evidence of the adsorbed domains of ionic surfactants at
Although the behavior of surfactants and their self- a hydrophilic oxide surface. Using pyrene as a fluorescence
association in bulk solution are well understood phenom- probe, the micelle-like aggregation of the adsorbed surf-
ena,4 the behavior of amphiphilic surfactant molecules at actant molecules has been confirmed.11 Likewise, the
the solid-liquid interface is a complex phenomenon which aggregation number of the adsorbed surfactant aggregates
depends on the properties of the solid surface, the ionic at the surface has been determined by using time-resolved
strength of the solution, the surfactant structure, the pH, florescence quenching.21 Electron spin resonance (ESR)
and the temperature.6-13 Even though the concept of spectroscopy suggests that the adsorbed surfactant mono-
mer is oriented vertically at the surface,24 and a structural
* To whom correspondence should be addressed. Fax: 00 33 4
transition of the adsorbed cationic surfactant from random
76 82 69 33. E-mail: sami.boufi@fss.rnu.tn. adsorption to a more organized structure has been shown
† Faculté des Sciences de Sfax. by combining ATR polarization spectroscopy with contact
‡ INPG. angle and ζ-potential measurements at silica surfaces.25
(1) Rosen, M. Surfactants in Emerging Technologies; Marcel Dek- On behalf of these recent investigations, several morphol-
ker: New York, 1987.
(2) Myers, D. Surfactants Science and Technology; VCH: New York,
1992. (13) Böhmer, M. R.; Koopal, L. K. Langmuir 1992, 8, 2660.
(3) Beneventi, D.; Carré, B.; Gandini, A. Colloids Surf., A 2001, 189, (14) Somasundaran, P.; Fuerstenau, D. W. J. Phys. Chem. 1966, 70,
65. 90.
(4) Christian, S. D.; Scamehorn, J. F. Solubilization in Surfactant (15) Somasundaran, P.; Fuerstenau, D. W. Trans. AIME 1972, 252,
Aggregates; Surfactant Sciences Series 38; Marcel Dekker: New York, 275.
1995. (16) Heather, N. P.; Gregory, G. W.; Srinivas, M.; Ilhan, A. A.
(5) Wang, Y.; Banziger, J.; Dubin, P. L.; Fillippelli, G.; Nuraje, N. Langmuir 1999, 15, 1685.
Environ. Sci. Technol. 2001, 35, 2608. (17) Sharma, B. G.; Basu, S.; Sharma, M. M. Langmuir 1996, 12,
(6) Atkin, R.; Craig, V. S. J.; Wanless, E. J.; Biggs, S. Adv. Colloid 6506.
Interface Sci. 2003, 103, 219-304. (18) Söderlind, E.; Stilbs, P. Langmuir 1993, 9, 1678.
(7) Goloub, T. P.; Koopal, L. K. Langmuir 1997, 13, 673. (19) Quist, P. Q.; Soderlin, E. J. Colloid Interface Sci. 1995, 172, 510.
(8) Fuerstenau, D. W. J. Colloid Interface Sci. 2002, 256, 79. (20) Tiberg, F.; Jönsson, B.; Lindman, B. Langmuir 1994, 10, 79.
(9) Scamehorn, J. F.; Schechter, R. S.; Wade, W. H. J. Colloid Interface (21) Ström, C.; Hansson, P.; Jönsson, B.; Söderman, O. Langmuir
Sci. 1982, 85, 463. 2000, 16, 2469.
(10) Chandar, P.; Somasundaran, P.; Turro, N. J. J. Colloid Interface (22) Li, H.; Tripp, C. P. Langmuir 2002, 18, 9441.
Sci. 1987, 117, 31. (23) Singh, P. K.; Adler, J. J.; Rabinovich, Y. I.; Moudgil, B. M.
(11) Fan, A.; Somasundaran, P.; Turro, N. J. Langmuir 1997, 13, Langmuir 2001, 17, 468.
506. (24) Esumi, K.; Matoba, M.; Yamanaka, Y. Langmuir 1996, 12, 2130.
(12) Böhmer, M. R.; Koopal, L. K. Langmuir 1992, 8, 2649. (25) Huang, L.; Somasundaran, P. Colloids Surf., A 1996, 117, 235.

10.1021/la050367n CCC: $30.25 © 2005 American Chemical Society

Published on Web 07/26/2005
Surfactant Adsorption onto Cellulosic Fibers Langmuir, Vol. 21, No. 18, 2005 8107

ogy patterns of the adsorbed aggregates have been

proposed: monolayer14,15,26 (hemimicelle), bilayer27 (ad-
micelle), spherical,28 and cylindrical17 surface aggregates.
These organized structures can be continuous along the
surface, or they may form patches of aggregated domains.
Adsorption isotherms usually provide information con-
cerning the onset of surfactant self-assembly occurring
when the solution concentration and the surface excess
are altered.
The common feature of experimental isotherms with
surface aggregation is their shape, which can be divided
into three distinct regions,8,15,26,27-30 corresponding to
different states of the adsorbed phase. In region I, located
in the low concentration domain, ionic surfactants are
adsorbed individually on a localized site with opposite
sign through an ion exchange mechanism. Adsorption is
sparse with no interaction among adsorbed surfactant
molecules. Up to a critical surfactant concentration, which Figure 1. Five-region general adsorption isotherm foreseen
corresponds to the boundary between regions I and II, by theoretical models for the surfactant adsorption on an
adsorbed monomers begin to associate through lateral oppositely charged surface. Dotted and thin lines represent the
interactions, inducing the formation of monolayered partitioning of surfactant molecules in the three possible
adsorbed states, viz., monomer, hemimicelle, and admicelle.
(hemimicelle) and bilayered (admicelle) aggregates. This
concentration, called the hemimicelle concentration (hmc),
occurs well below the surfactant cmc. The generation of corresponding isotherms exhibited three to five charac-
surfactant aggregates boosts monomer adsorption and teristic regions, depending on the surfactant structure,
explains the abrupt rise in the S-shaped isotherm. A which were interpreted in terms of surfactant aggregation
further increase in the surfactant concentration leads to at the solid-liquid interface. The purpose of this work
the progressive stabilization of surfactant adsorption up was to analyze the effect of the cellulose charge on the
to the saturation of the interface near the cmc. The driving adsorption behavior of cationic surfactants. The surface
force behind surfactant aggregation is the transfer of charge of cellulose fibers was modified by a soft oxidation
surfactant alkyl chains from an aqueous to a hydrocarbon treatment37,38 which allowed the selective conversion of
medium (hemimicelle and admicelle core) during self- the primary alcohol groups of the glycosyl unit into
association. This attractive interaction is counteracted carboxylic groups which remain fully ionized at pH higher
by electrostatic repulsion forces between charged head- than 6.
groups within surfactant aggregates, and when dispersive
and electrostatic-repulsive forces are balanced, a satura- Materials and Methods
tion plateau is attained, viz., region III. Materials. Cellulose fibers used in this study were a bleached
Although their interpretation is simple, S-shaped soda pulp from the Tunisian annual plant esparto (alfa, Stipa
isotherms do not allow clear identification of monolayer tenacissima). They were highly porous and had a specific surface
to bilayer phase transitions; indeed, the detailed descrip- area in a dry state of 3 m2 g-1. Their morphology was determined
tion of surfactant self-assembly onto oppositely charged by averaging the length and width of 7380 fibers characterized
with optical microscopy. The mean fiber length and width were
surfaces provided by theoretical models31-33 predicts a
found to be 0.75 mm and 14.2 µm, respectively, thus corresponding
five-region adsorption isotherm, Figure 1, where surface to an aspect ratio of 52. The carboxylic content of the original
saturation occurs via a hemimicelle and admicelle forma- fibers was 45 µmol g-1. Chemicals used to oxidize cellulose fibers,
tion sequence. The general five-region adsorption isotherm namely, TEMPO (2,2,6,6-tetramethyl-1-piperidinyloxy radical),
has a three-region shape when monomer-to-monolayer sodium bromide (NaBr), and 12% sodium hypochlorite solution,
and monolayer-to-bilayer transitions occur in a narrow were of laboratory grade and used without further purification.
concentration range, which seems to be the case for most Four commercial, analytical grade, cationic surfactants,
surfactant/oxide systems. namely, octadecyltrimethylammonium bromide (C18), hexa-
In this work, we pursue our investigation on the decylpyridinium chloride (C16), tetradecyltrimethylammonium
interactions of cationic surfactants with a cellulose bromide (C14), and dodecylpyridinium chloride (C12), were
dissolved in distilled water to obtain high-concentration mother
substrate. In a previous work we showed that, even if
solutions. Their critical micelle concentration (cmc) in distilled
cellulose is a poorly charged substrate, cationic surfactants water (25 °C) determined by conductometry was 2.1 × 10-4, 7.5
are adsorbed in a relatively high amount.34-36 Their × 10-4, 3.4 × 10-3, and 1.2 × 10-2 mol L-1 for C18, C16, C14,
and C12, respectively.
(26) Koopal, L. K.; Lee, M. E.; Böhmer, M. R. J. Colloid Interface Sci. Oxidation Procedure of Cellulosic Fibers. Oxidation of
1995, 170, 85. cellulose fibers was carried out in the following conditions. Fibers
(27) Harwell, J. H.; Hoskin, J.; Schechter, R. S.; Wade, W. H.
Langmuir 1985, 1, 251. (10 g) were dispersed in distilled water (500 mL). TEMPO (25
(28) Adler, J. J.; Singh, P. K.; Patist, A.; Rabinovich, Y. I.; Shah, D. mg) and NaBr (250 mg) were added to the suspension, and the
O.; Moudgil, B. M. Langmuir 2000, 16, 7255. pH was adjusted to 10 by addition of a 0.5 mol L-1 NaOH solution.
(29) Gouloub, T. P.; Koopal, L. K.; Bijsterbosch, B. H.; Sidorova, M. A 12% sodium hypochlorite solution was added to the suspension
P. Langmuir 1996, 12, 3188. over 30 min, and the resulting suspension was then stirred for
(30) Fuerestenau, D. W.; Herrara-Urbina, R. In Surfactant-Based 2 h. The pH of the suspension was continuously adjusted to 10
Separation Process; Scamehorn, J. F., Harwell, J. H., Eds.; Surfactant
Sciences Series 33; Marcel Dekker: New York, 1989; pp 259-320.
(31) Łajtar, L.; Narkiewicz-Michałek, J.; Rudzinski, W. Langmuir (35) Aloulou, F.; Boufi, S.; Belgacem, N.; Gandini, A. Colloid Polym
1994, 10, 3764. Sci. 2004, 283, 344.
(32) Li, B.; Ruckenstein, E. Langmuir 1996, 12, 5052. (36) Aloulou, F.; Boufi, S.; Beneventi, D. J. Colloid Interface Sci.
(33) Drach, M.; Łajtar, L.; Narkiewicz-Michałek, J.; Rudzinski, W.; 2004, 280, 350.
Zaj_c, J. Colloids Surf., A 1998, 145, 243. (37) De Nooy, A. E.; Besemer, A. C.; Van Bekkum, H. Carbohydr.
(34) Aloulou, F.; Boufi, S.; Chakchouk, M. Colloid Polym Sci. 2004, Res. 1995, 269, 89.
282, 699. (38) Chang, P. S.; Robyt, J. F. J. Carbohydr. Chem. 1996, 15, 819.
8108 Langmuir, Vol. 21, No. 18, 2005 Alila et al.

Table 1. Carboxylic Content and Crystallization Degree

of Oxidized Cellulose Samples
carboxyl carboxyl
contentb contentc crystallization
abbreviation Va (µmol g-1) (µmol g-1) degreed (%)
cellulose-0 0 39 45 65
cellulose-150 15 145 155 70
cellulose-300 32 294 304 79
cellulose-600 65 567 585 82
a Volume of 12% hypochlorite solution added to 10 g of fiber

suspension. b Carboxyl content determined by methylene blue

sorption. c Carboxyl content determined by the conductometric
titration method d Crystallization degree determined by X-ray
powder diffraction.

chloride) (PVC) plasticized by bis(2-ethylhexyl phthalate) (20/

80, w/w) containing 1% surfactant carrier complex. The latter
was prepared by stoichiometric mixing of the anionic surfactant
and hexadecyltrimethylammonium chloride in water, followed
Figure 2. TEMPO-mediated oxidation of cellulose primary by its recrystallization in a diethyl ether-dichlormethane
hydroxyl groups to carboxyl groups. mixture (50/50, v/v). The membrane potential was measured with
the following concentration cell assembly:
by addition of the NaOH solution. The oxidation was stopped by
adding 100 mL of ethanol, and oxidized fibers were filtered and
washed two times with a 1 mol L-1 NaOH solution to remove
traces of residual lignin. The recovered fibers were washed with
distilled water until the pH of the filtrate was close to 7. Cellulose The reference solution was a 10-3 mol L-1 solution of the
oxidation by TEMPO is depicted in Figure 2. surfactant studied and 5 × 10-3 mol L-1 KCl solution as an ionic
The content of carboxylic moieties in the oxidized fibers was strength buffer. Before use, the electrodes were conditioned
evaluated by using two techniques: (i) conductometric titration overnight in a 10-3 mol L-1 surfactant solution. The electromotive
of a 1% fiber suspension with 10-3 mol L-1 HCl39 (details in the force (EMF) of the electrochemical cell was stable to within 1
Supporting Information) and (ii) methylene blue (MB) sorption, mV. The electrodes gave a Nerstian response for surfactant
which is based on the replacement of carboxylic counterions by concentrations from the cmc down to 5 × 10-6 mol L-1:
the cationic dye.40
The X-ray diffraction patterns were measured with an X-ray E ) E0 + p log [surfactant]
diffractometer using Cu KR radiation at 40 kV and 30 mA. The
crystallization degree is obtained by comparing the area under where p is the experimental slope of the EMF (E) versus log
the crystalline peaks and the amorphous curve according to the [surfactant] plot at 25 °C and E0 is the EMF corresponding to
usual method.41 a surfactant concentration of 1 mol L-1. The slope observed was
Adsorption Isotherms. The adsorption isotherms of cationic in a range of 50-65 mV.
surfactants were carried out at 25 °C using the depletion method. The validity of the dosage method was confirmed by a
An appropriate volume of a high-concentration mother solution spectroscopic-UV dosage in the presence of hexadecylpyridinium
was added to a 1% cellulose fiber suspension to obtain surfactant chloride. The difference in the concentration determined via the
concentrations ranging from 10-5 to 8 × 10-3 mol L-1. The pH two analytical methods did not exceed 5%.
of the suspension was not affected by the addition of cationic ζ-Potential. A commercial ζ-potential analyzer (Malvern
surfactant, and the pH was close to 7.5-8 for all cellulose samples. 2000) was used to measure the electrophoretic mobility of cellulose
The suspension was then stirred for 24 h at room temperature fibers in the aqueous suspension. Measurements were conducted
to reach adsorption equilibrium. Dispersions were then centri- on the suspension fine fraction, which was obtained after filtration
fuged at 2500 rpm for 15 min, and the amount of free surfactant of the original suspension through a 45 µm screen. To avoid the
in the supernatant was determined by potentiometric titration fluctuation of ζ-potential due to ionic strength changes generated
using a surfactant-selective electrode for the dosage of C14 and by the increase in the surfactant concentration, the aliquot sample
C18 and by UV absorption at 260 nm in the presence of C16 and was diluted with a concentrated solution of KCl to buffer the
C12 surfactants bearing a pyridinium ion. ionic strength to 10-2 mol L-1. Four consecutive measurements
Chloride Ion Dosage. The concentration of free chloride ion were taken for each sample at room temperature and averaged.
in solution was determined by precipitation with Ag+ according The ζ-potential was calculated from electrophoretic mobility data
to a well-reported analytic technique.42 An excess amount of Ag+ by using the Smoluchowski equation. We verified in a preliminary
solution was added to the chloride ion solution, and the excess study that ζ-potential measurements repeated on the whole
of Ag+ was back-titrated with thiocyanate solution after removal suspension by using the streaming potential technique (Mütek
of the precipitated AgCl. SZP06 with a 40 µm screen electrode) matched with ζ-potential
Surfactant Dosage Using a Surfactant-Selective Elec- values obtained by electrophoresis, when measurements were
trode. The surfactant concentration was determined potentio- conducted on the suspension fine fraction.
metrically by using a cation-selective electrode. The electrode
was prepared as described by Ishibashi;43 its sensitive element Results and Discussion
was a 0.4 mm thick and 1 cm diameter membrane of poly(vinyl
Three samples of oxidized cellulose fibers were prepared
(39) Katz, S.; Beatson, R.P.; Scallan, A.M. Sven. Papperstidn. 1984, by soft oxidation with a TEMPO/NaBr/NaClO system.
87 (6R), 48. Their carboxyl content, expressed as moles per gram of
(40) Rohrsetzers, S.; Kovacs, P.; Kabai-Faix, M.; Papp, J.; Volgy, P. anhydroglycosyl units, is reported in Table 1.
Cell. Chem. Technol. 1995, 29, 65. 13
(41) Rabek, J. F. Experimental Methods in Polymer Chemistry: C solid NMR and FT-IR spectroscopic analysis did
Application of Wide-angle X-ray Diffraction (WAXS) to the Study of the not reveal a significant evolution in the chemical structure
Structure of Polymer; Wiley-Interscience: Chichester, U.K., 1980; p of the oxidized fibers, excluding the generation of car-
505. boxylic groups. FT-IR spectra of oxidized samples, pre-
(42) Harris. D. C. Exploring Chemical analysis; W. H. Freeman: New
York, 1997; p 112. sented in the Supporting Information, revealed a signifi-
(43) Ishibashi, N.; Masadome, T.; Imato, T. Anal. Sci. 1986, 2, 487. cant intensification of the CdO (∼1732 cm-1) and C-O
Surfactant Adsorption onto Cellulosic Fibers Langmuir, Vol. 21, No. 18, 2005 8109

adsorbed surfactant was low and did not exceed 30-40

µmol g-1. When the free surfactant concentration reached
a critical level, a significant increase in adsorption (onset
of region II) was noted. This critical level was attained at
a concentration close to 3.1 × 10-4 and 10-4 mol L-1 for
C12 and C16, respectively. The adsorption was slowed as
the adsorption attained a certain level (region III), and
then when the surfactant concentration was further
increased, it increased again (region IV) until it reached
a constant level (region V).
At first glance it seems astonishing that the adsorption
remains low and increases slowly in region I despite the
presence of a relatively high amount of available negative
sites on the surface. In this region surfactant monomers
are adsorbed via electrostatic interaction between surf-
actant headgroups and carboxylic sites on the cellulose
surface. Hydrocarbon tails are supposed to lie flat on the
substrate to minimize their contact area with water
molecules. The relatively high surface occupied by the
Figure 3. Adsorption isotherm of C12 onto oxidized cellulose adsorbed species reduces the possibility of adsorption of
fibers according to a linear-log scale at pH 7.5-8 and in the other surfactant molecules. The steep increase in the slope
absence of any added salts. of adsorption isotherms in region II is a result of a change
in the conformation of adsorbed molecules from a flat to
a perpendicular orientation, where alkyl chains form
highly packed aggregates, minimizing their contact area
with water molecules and therefore increasing the entropy
of the system. In this region the electrostatic interaction
between the charged headgroups and carboxylic sites on
the surface is still present since the adsorbed amount is
not sufficient to neutralize the surface charge. Surfactant
molecules are still adsorbed with headgroups facing
toward the surface, and self-association leads to the
formation of monolayered aggregates.
Region III appears when the adsorbed amount reaches
about 50-75% (depending on the surfactant and surface
charge density) of the total carboxylic sites. The decrease
in the slope of the isotherm could be attributed to the
progressive neutralization of the cellulose surface and to
the subsequent (i) reduction of the electrostatic surface/
headgroup attraction and (ii) increase of the repulsion
between surfactant headgroups. The restart of adsorption
Figure 4. Adsorption isotherm of C16 onto oxidized cellulose in region IV occurs when the substrate charge is close to
fibers according to a linear-log scale at pH 7.5-8 and in the neutral (the amount of adsorbed equivalents is close to
absence of any added salts. the corresponding density of the negative carboxylic sites).
The adsorption in this domain may occur with a bilayer
(∼1050 cm-1) bands with an increase in the carboxylic conformation, with surfactant headgroups facing into the
content. X-ray diffraction patterns of the original and solution. The solution activity of the surfactant is now
oxidized fibers shown in the Supporting Information sufficient to overcome the electrostatic repulsion among
revealed the presence of three peaks, 2θ ) 15°, 17°, and headgroups and to continue the growth of the aggregated
22.7°, which confirmed that only cellulose I was present domains. The stabilization of the adsorption in region V
and no conversion of the supermolecular structure or may be attributed to the saturation of the bilayer. Except
polymorphic forms occurred during the oxidation treat- for C16-cellulse-150, this region occurs at a surfactant
ment.44 These results are in agreement with the high concentration lower than or close to the cmc. When we
selectivity of the TEMPO oxidation procedure, which compare the different cellulose substrates and the two
allows the specific oxidation of primary alcohol groups of surfactants, we note that the slope of region II becomes
the glycosyl unit into carboxylic groups.45 However, a slight much steeper as the surface charge density and the
increase in the crystallization degree is noted after the surfactant chain length increase. Likewise, the extent of
oxidation (Table 1); the higher the carboxylic content, the region III follows the same trend. These results are
higher the crystallization degree. consistent with our explanation. Indeed, with an increase
The adsorption isotherms of C12 and C16 surfactants in the surface charge density, the distance between
on cellulose fibers carrying three different amounts of adsorbed surfactant molecules is reduced and the pos-
carboxylic groups, viz., 150, 300, and 600 µmol g-1, are sibility for lateral hydrophobic tail-tail interaction is more
shown in Figures 3 and 4. When cellulose 300 was used promoted. Moreover, with an increase in the chain length,
as the substrate, C12 and C16 adsorption isotherms clearly the hydrophobic attraction between surfactant tails is
displayed a five-region shape. In region I, the amount of more intense. These two effects can explain the transition
from a five- to a three-region adsorption isotherm when
(44) Ishikawa, A.; Okano, T.; Surgiyama, J. Polymer 1997, 38, 463.
the cellulose surface charge is increased from 300 to 600
(45) da Silva Perez, D.; Montanari, S.; Vignon, M. R. Biomacromol- µmol g-1. We can suppose that the hydrophobic contribu-
ecules 2003, 4, 1417. tion to both C12 and C16 adsorption was boosted by the
8110 Langmuir, Vol. 21, No. 18, 2005 Alila et al.

Figure 5. Adsorption isotherm of C16 surfactants onto

cellulose-300 and cellulose-150 and the corresponding evolution
of the released amount of chloride (Cl) counterion. Figure 7. ζ-potential and adsorption of cellulose-150 as a
function of C12 and C16 solution concentration at pH 7.5-8.

surfactant was adsorbed with its counterion and no

exchange with other counterions took place.
(iv) With C12, the surfactant adsorption proceeded
through an exchange mechanism even in region III. The
fraction of the surfactant adsorbed with its counterion
was relatively low compared to that of C16; it was about
10% and 20% for cellulose-300 and -600, respectively.
These results suggest that the C12 surfactant was
adsorbed as a monolayer (hemimicelle) whereas C16 was
adsorbed as a bilayer (admicelle) which was generated in
region IV of the isotherm. Presumably, the adsorption of
C12 on a charged cellulosic substrate occurs predominantly
head-on, with the compensation of the surfactant charge
by the surface charge. Adsorption is mainly driven by
electrostatic interaction. On the other hand, strong
hydrophobic interaction associated with the long alkyl
Figure 6. Adsorption isotherm of C12 and C16 surfactants chain of C16 leads to the formation of a second layer with
onto cellulose-600 and the corresponding evolution of the the surfactant headgroups facing toward the solution. The
released amount of chloride (Cl) counterion. comparison between chloride ion and C16 adsorption
isotherms confirmed that the switch from a five- to a three-
high surface charge and monolayered and bilayered region adsorption isotherm obtained when the cellulose
aggregates grew simultaneously, thus inducing the su- charge was increased from 300 to 600 µmol g-1 can be
perposition of regions II, III, and IV. associated with a nearly simultaneous self-assembly of
This hypothesis on the mechanisms of surfactant self- C16 into monolayered and bilayered aggregates.
assembly at the surface of cellulose fibers was based on To further support the difference in the configura-
the different phenomena reported to take place in the tion of C12 and C16 surfactants on a cellulose substrate,
interaction of ionic surfactants with oxide surfaces. Since the variation of the ζ-potential of cellulose fibers was
the validity of this adsorption model for surfactant/ measured as a function of the surfactant bulk concentra-
cellulose systems has not been proven yet, in this work, tion. The comparison of the cellulose ζ-potential with the
we will try to provide further arguments to support the corresponding adsorption isotherm of C12 and C16 on
extension of this model to surfactant/cellulose systems. cellulose-150 and -300, Figures 7 and 8, shows the
In Figures 5 and 6, the amount of chloride counterion following.
released by the adsorbed surfactant is superimposed to (i) The evolution of the ζ-potential with surfactant
the adsorption isotherms of C16 and C12 surfactants on concentration displayed the same transition points ob-
cellulose-150, -300, and -600, thus showing the following. served on the adsorption isotherms.
(i) For the two surfactants, the amount of released (ii) At a given adsorbed amount of C12 and C16
counterion followed the same trend of the adsorption surfactants, cellulose-300 displayed a ζ-potential lower
isotherm until adsorbed equivalents became close to the than that of cellulose-150. This difference is consistent
surface charge capacity. Up to this level, the chloride ion with the higher density of negatively charged sites on
was no longer released in the solution and the surfactant cellulose-300.
was adsorbed with its counterion. (iii) In the presence of C12, the ζ-potential at the
(ii) In the presence of C16, the fraction of the adsorbed saturation of the cellulose surface was close to zero.
surfactant with its counterion was close to 50%. (iv) With C16 the reverse in the sign of the ζ-potential
(iii) In the presence of cellulose-150 and -300, the plateau occurred in region III. As the onset of this region coincided
level of the released counterion coincided with the onset with the isoelectric point (iep) of the substrate, we could
of region III, which means that the C16 adsorption in this suppose that the electrostatic repulsion between the
region was no longer subjected to the electrostatic at- cationic surface and headgroups caused the surfactant
traction of the charged sites on the surface as the adsorption to decrease in region III. Nevertheless, ad-
Surfactant Adsorption onto Cellulosic Fibers Langmuir, Vol. 21, No. 18, 2005 8111

Figure 9. Adsorption isotherms of C16 SDBS onto cellulose-

150 at pH 7.5-8 and three different ionic strengths (addition
of KCl).
Figure 8. ζ-potential and adsorption of cellulose-300 as a
function of C12 and C16 solution concentration at pH 7.5-8.

sorption continues to operate as the result of the hydro-

phobic effect of the surfactant tail. The positive value of
the ζ-potential and its continuous increase in regions IV
and V are in agreement with the hypothesis of the
generation of a bilayer.
(v) The ζ-potential value in region V with C16 increased
as the substrate charge density increased. This result
provides another confirmation of the growth of a second
layer.
The analysis of ζ-potential measurements gave us
additional information to describe more precisely phe-
nomena associated with the adsorption of a cationic
surfactant on charged cellulose fibers. The adsorption of
the C12 surfactant was mainly driven by electrostatic
interactions, and the hydrophobic effect intervened es-
sentially to promote the self-assembly of adsorbed mono- Figure 10. Adsorption isotherm of C12, C14, C16, and C18
cationic surfactants onto cellulose-150 at pH 6.5-7 and in the
mers, namely, the transition from a flat to a perpendicular absence of any added salts.
conformation. The adsorption is limited to a monolayer.
The addition of four CH2 units in the alkyl chain length The influence of the alkyl chain length on the surfactant
increased the contribution of the hydrophobic interaction, adsorption is shown in Figure 10. The increase in the
overcoming repulsion forces among the cationic surface alkyl chain length led to a shift of the adsorption isotherms
and surfactant headgroups and promoting the growth of toward low concentrations and to a continuous increase
a second layer (region IV). in the maximum amount of adsorbed surfactant.
The evolution of the adsorption isotherm with ionic C12 gave a three-region adsorption isotherm which was
strength has been studied with C16 and cellulose-150. interpreted as reflecting the generation of mainly mono-
Figure 9 shows the adsorption isotherm at pH 7 for C16 loayered structures. C14 and C16 had five-region iso-
on cellulose-150 in the presence of different concentrations therms; we can suppose that the increase in the alkyl
of KCl. At a low amount of adsorbed surfactant (lower chain length favored the generation of bilayered ag-
than 20 µmol g-1) adsorption decreased with increasing gregates. The three-region isotherm obtained for C18
salt concentration, and it was enhanced above this level. shows that strong hydrophobic interactions among C18
Likewise, the amount of adsorbed surfactant at the onset alkyl chains enabled the electrostatic repulsion forces to
of region III and its slope increased with salt addition. At be overcome, thus inducing a continuous transition from
low surfactant concentration, region I, surfactant head- region I to region IV (a hemimicelle and an admicelle are
groups and cellulose surface charges are screened by the generated simultaneously). These trends are in agreement
increase in the ionic strength, thus provoking a decrease with the hypothesis of surfactant self-association into
in cellulose-surfactant interactions and adsorption. At monolayered and bilayered structures. Indeed, the in-
higher surfactant concentration, viz., in regions II, III, crease of the number of CH2 groups in the surfactant tail
and IV, surfactants form aggregates on the cellulose strengthened hydrophobic interactions among alkyl chains,
surface where cationic headgroups are closely packed. The and induced a decrease in the free energy of hemi-
charge screening due to the increase of the ionic strength micellization, as in the well-known correlation between
reduces the repulsion between the headgroups, thus the increase in surfactant chain length and the corre-
enhancing the formation of closely packed aggregates and sponding decrease of its cmc2.
surfactant adsorption. The contribution of ionic strength Despite the low surface area measured for the cellulose
to surfactant self-assembly was also shown by the fibers used in this study (3 m2 g-1), maximum adsorbed
progressive transition from a five- to a three-region amounts of surfactant were relatively high compared to
adsorption isotherm when the ionic strength was increased those of the charged oxide substrate, which displays such
from 0 to 10-2, Figure 9. adsorption levels only with specific surface areas higher
8112 Langmuir, Vol. 21, No. 18, 2005 Alila et al.

Figure 11. Possible adsorption isotherms and aggregation states of cationic surfactants at the surface of negatively charged
cellulose fibers. These isotherms and aggregation states are consistent with models proposed for surfactant/oxide systems.31-33 (a)
Three-region adsorption isotherm with monolayer aggregation, i.e., C12. Dispersive interactions among C12 molecules are too weak
to overcome electrostatic repulsion and to allow bilayer packing. (b) Five-region adsorption isotherm with monolayer and bilayer
aggregation, i.e., C14 and C16. Dispersive interactions among surfactant molecules are strong enough to allow the formation of
a bilayer. The monolayer to bilayer transition is retarded (presence of region III). (c) Three-region adsorption isotherm with
monolayer and bilayer aggregation, i.e., C18. Strong dispersive interactions boost surfactant self-assembly, monolayer and bilayer
aggregates form simultaneously, and region III (b) disappears.

than 200-300 m2 g-1.46 This result could be justified only was linked with the charge density of the substrate and
if we consider that surfactant molecules were adsorbed the length of the surfactant hydrocarbon chain.
not only at the external surface of the fibers but also inside (ii) The increase in the negative charge density of
their porous structures thanks to the fiber swelling cellulose improved electrostatic forces promoting surfac-
occurring during fiber oxidation. In other words, the tant adsorption and self-assembly, thus allowing the dense
cellulose surface available for surfactant adsorption in packing of adsorbed molecules into monolayered and
the water-swollen state was much larger than that bilayered aggregates.
measured for the dry substrate. (iii) The alkyl chain length affected surfactant self-
aggregation especially near the saturation of the substrate.
Conclusions The increase in the surfactant adsorption density and in
The adsorption and self-assembly of cationic surfactants the surface charge observed when the alkyl chain length
on oxidized cellulose displayed a complex behavior de- was increased from 12 to 18 C atoms was associated with
pending on the cellulose charge, the surfactant molecular the progressive transition from monolayer to bilayer
structure, and the ionic strength of the solution. Although aggregates. With C12 dispersive interactions are not
we did not rationalize experimental results with a rigorous strong enough to prevail over electrostatic repulsion and
mathematical treatment, the contribution of these vari- after the neutralization of the surface charge adsorption
ables to surfactant adsorption and aggregation was in decreases quickly; only monolayered aggregates are
agreement with tendencies which can be predicted by supposed to form. With C18 strong dispersive interactions
DLVO theory. promote surfactant adsorption and aggregation even after
(i) Adsorption of cationic surfactants proceeded via surface charge reversal; bilayered aggregates are supposed
electrostatic and dispersive interactions whose intensity to form.
(iv) The ionic strength influenced adsorption by screen-
(46) Thakulsukaant, C.; Lobban, L. L.; Osuwan, S.; Waritswat, A. ing the cellulose surface charge and the surfactant ionic
Langmuir 1997, 13, 4595. head. At low surfactant concentration, when adsorption
Surfactant Adsorption onto Cellulosic Fibers Langmuir, Vol. 21, No. 18, 2005 8113

is driven by electrostatic forces, adsorption was retarded cellulose-water interface and adsorption isotherms is
when the ionic strength was increased. At high surfactant given in Figure 11.
concentration, when electrostatic forces among surfactant
headgroups hinder a dense molecular packing on the Acknowledgment. This work was financially sup-
substrate, adsorption was boosted when the ionic strength ported by the Tunisian “Ministère de la Recherche
was increased. Scientifique, technologique et de developpement des
The negative charge of the cellulose substrate, the alkyl compétences” and by the International Foundation for
chain length, and ionic strength affected in similar ways Science through Grant W/3358-1.
surfactant adsorption and self-assembly. Indeed, an Supporting Information Available: Details pertain-
increase in one of these parameters was reflected by an ing to conductometric titration, FT-IR characterization, and X-ray
increase in surfactant adsorption and aggregation with a characterization. This material is available free of charge via
consequent variation in the shape of the adsorption the Internet at http://pubs.acs.org.
isotherm, which switched from a five- to a three-region
curve. A sum of possible surfactant configurations at the LA050367N