Combet 2005

7456 Macromolecules 2005, 38, 7456-7469
Scattering Functions of Flexible Polyelectrolytes in the Presence of

Mixed Valence Counterions: Condensation and Scaling
Jérôme Combet,* François Isel, and Michel Rawiso

Institut Charles Sadron (CNRS-ULP), 6 rue Boussingault, 67083 Strasbourg Cedex, France
François Boué
Laboratoire Léon Brillouin (CEA-CNRS), CE Saclay, 91191 Gif-sur-Yvette Cedex, France
Received October 1, 2004; Revised Manuscript Received June 3, 2005
ABSTRACT: We have investigated by small-angle X-ray and neutron scattering the structure of salt-
free aqueous solutions of sulfonated polystyrene with variable fractions of monovalent (sodium, Na) and
divalent (calcium, Ca) counterions. We follow the variation of the position of the maximum in the scattered
intensity, q*, as a function of the concentration, c, and the fraction of Ca counterions. A relevant parameter
is the degree of condensation of Na and Ca counterions in the different mixtures. We survey the literature
and check that, in the case of monovalent counterions, the variation of q* with c can be described by the
scaling theory of Dobrynin, Colby, and Rubinstein, provided the counterion condensation is included
according to the Manning-Oosawa approach. Scattering data from mixed valence systems are still in
agreement with this scaling model at lower concentrations, except for pure CaPSS2 aqueous solutions.
For increasing concentration, a discrepancy is enhanced that extends toward higher Ca fractions. This
indicates that counterion condensation is not the only process. Another possible one is a retraction of the
macroions due to divalent counterions that induce bridging-type correlations between monomers, as
already observed on the form factors of CaPSS2 and NaPSS plus added CaCl2 salt. Such a retraction
modifies the scaling, in particular, via a change in the overlap concentration c*. Furthermore, it is
noticeable that q* measurements have to be handle with care. We show that misinterpretation of the
characteristic maximum in the small-angle scattering of polyelectrolyte solutions may occur, especially
at high concentration, if the contribution of the condensed counterions to the total scattering function is
predominant.
1. Introduction to distinct aggregated structures according to their

length and/or their stiffness. For rigid macroions, it is
Polyelectrolytes are ionic polymers which, when dis-
the formation of dense or ordered packages, as bundles
solved in a polar solvent such as water, dissociate into
for DNA condensates, that tend to form a network
macroions and small mobile counterions. Electrostatic
superstructure, which is observed; for flexible macro-
charges of one sign are localized on the macroions
ions, it is rather a retraction in the average conforma-
whereas an equivalent amount of oppositely charged
tion or the formation of compact precipitates or macro-
counterions is either condensed onto the macroions or
scattered in the solution. Because of their relevance in scopic physical gels, which is observed. Therefore, the
biology as well as in technological applications, increas- origin of multivalent counterion-mediated attractions
ing attention has been paid to these charged polymers.1-6 between like-charged chains, or monomers, is presum-
The understanding of their properties is, however, far ably different in both cases. It is the existence of diffuse
from complete. When dissolved in a good solvent (hy- electrostatic correlations among condensed counterions
drophilic polyelectrolytes), their properties are more that induce the formation of bundles, whereas the
complex than those of neutral polymers because elec- precipitation of linear flexible polyelectrolytes in mul-
trostatic interactions are long-ranged. In a poor solvent tivalent salts is attributed to the bridging of monomers
(hydrophobic polyelectrolytes), the complexity results via multivalent ions associated with more than one
from the competition between the long-range electro- monomer.
static repulsions and the short-range attractions associ- In this paper we wish to follow step by step the effect
ated with the poor solubility. of the counterion valence on the structure of a given
But other degrees of freedom are brought by coun- highly charged flexible polyelectrolyte. To consider only
terions that can also play a crucial role with respect to electrostatic effects, our system will not involve any
interactions. Numerous experiments6-19 as well as complexation (chemical bonding) between macroion and
simulations6,20-25 and theories6,26-39 have shown that counterions. This is the case for sulfonated polystyrene
like-charged macroions can attract each other through (PSS) linear macroions with sodium (Na) and/or calcium
effective forces of electrostatic origin. Contrarily to the (Ca) counterions.14 The static structure of their semi-
usual electrostatic repulsions, such attractive interac- dilute aqueous solutions is investigated here using
tions are short-ranged. They also require the presence small-angle X-ray and neutron scattering (SAXS and
of multivalent counterions, or small ions of the opposite SANS) techniques.
sign, in the solutions. For linear macroions, they lead SANS experiments have already been devoted to
solutions of that kind.16,17 They show first that the
* Corresponding author: Tel +33 3 88 41 40 96; Fax +33 03 88 characteristic broad maximum in the scattering function
41 40 99; e-mail combet@ics.u-strasbg.fr. is shifted toward smaller scattering vectors, when
10.1021/ma0479722 CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/21/2005
Macromolecules, Vol. 38, No. 17, 2005 Flexible Polyelectrolytes 7457
monovalent (Na) counterions are replaced by divalent Table 1. Characteristics of Polystyrene and Sulfonated
(Ca) ones. Correlatively, the average size of the macro- Polystyrene Samplesa
ions is found to be smaller in the presence of Ca Mn (g/mol) Nn I τs τw
counterions. Two distinct interpretations were proposed. NaPSS 77 600 745 1.04 0.90 0.07
On one hand, a decrease in the effective charge fraction CaPSS 77 600 745 1.04 0.90 0.07
associated with the counterion condensation process40,41 TMAPSS 82 200 733 1.16 1.00 0.10
was invoked;16 on the other hand, a conformational a M and N are the number-average molecular weights and
n n
change resulting from an electrostatic bridging between the number-average degrees of polymerization of the parent PS
monomers by Ca counterions was considered.17 The macromolecules (Nn ) Mn/m; m is the molar mass of PS monomers
second explanation was obviously inspired by previous (104.15 g/mol for PSh and 112.20 g/mol for PSd)); I is their
investigations on complexation between weak polyelec- polydispersity index. These characteristics have been obtained
trolytes and divalent counterions.9,14 It is also strength- from SEC measurements, by using THF as eluent and a low-angle
laser light scattering spectrometer (LALLS; down DSP) as molec-
ened by some numerical simulations on short macro- ular mass detector. τs and τw are the degree of sulfonation and
ions.25,42 However, both interpretations are reasonable the weight fraction of water content of the related dried Na, Ca,
since the experimental results cannot really distinguish and TMAPSS polyelectrolytes, respectively. Such characteristics
between them. are obtained from elemental analysis. Specifically, they allow to
To settle this problem, we intend to take SAXS and determine the concentrations (mol/L) of the various aqueous
SANS experiments up on more suitable systems, namely solutions.
salt-free aqueous solutions of PSS with mixtures of Na
and Ca counterions of variable ratios. Such mixed number of monomers, and the weight fraction of water content
valence systems have actually been studied in the τw were determined for each sample. The molecular weights
of the parent PS polymer samples (before sulfonation) were
past.43-47 We will see that they allow more specifically
characterized by size exclusion chromatography (SEC). Table
to distinguish condensed counterions from free coun- 1 lists the main characteristics of these samples.
terions and therefore to establish their respective real The salt-free aqueous solutions of Na-CaPSS were obtained
roles in attractive interactions between monomers when by dissolving NaPSS and CaPSS chains in ultrapure H2O
these counterions are divalent. For the sake of simplic- (Millipore grade, conductivity < 1 µS) for X-ray scattering
ity, scattering data will be analyzed in the framework experiments or in D2O (Euriso-top CEA group, 99.9% D) for
of the scaling theories48-50 combined with the counterion neutron scattering measurements. The salt-free aqueous solu-
condensation concept. More precisely, the validity of tions of TMAPSS were prepared in both H2O and D2O for
such an approach will be challenged for each of the neutron scattering measurements. All these solutions were
mixed valence systems and studied according to the heated until complete dissolution (at 50 °C for 1 h) and then
let stand for at least 2 days prior to the measurements at room
concentration in the semidilute regime.
temperature. Concentrations are determined from the masses
The rest of the paper is organized as follows: In of solute and solvent by using the tabulated partial molar
section 2, we will describe all the experimental details volumes56,57 and also by taking into account the water contents
and present the main characteristics of the investigated of the various PSS powders. Solutions are then characterized
PSS samples as well as the ones of other systems by their monomer concentration c (mol/L) and their fraction
extracted from literature for which particular attention of NaPSS macromolecules X, as defined by the molar ratio X
will be paid. In section 3, we will present SAXS and ) n(NaPSS)/[n(NaPSS) + n(CaPSS)] (because NaPSS and
SANS results obtained from salt-free aqueous solutions CaPSS samples have identical macroions, n can also be
of Na-CaPSS. Section 4 will be devoted to the Man- considered as a monomer concentration, in mol/L). Under this
ning-Oosawa approach of the counterion condensation procedure, X ) 0 and 1 correspond to pure CaPSS and NaPSS,
respectively.
process and the scaling theory of Dobrynin, Colby, and
2.2. Small-Angle X-ray and Neutron Scattering (SAXS
Rubinstein. We will demonstrate that a combination of and SANS) Measurements. SAXS experiments were carried
these models is able to describe the static structure of out in our laboratory with a diffractometer developed by
salt-free aqueous solutions of flexible polyelectrolytes Bruker-Anton Paar (Nanostar). This diffractometer operates
in the presence of monovalent counterions. At last, in with a pinhole collimator and a two-dimensional multiwires
section 5, we will focus on the mixed valence systems, proportional detector. A monochromatic (λ ) 1.54 Å with ∆λ/λ
and comparison with the previous theoretical models ) 4%) and almost parallel beam (divergence of the order of
will be carried out in order to elucidate the respective 0.03°) is obtained through cross-coupled Göbel mirrors. The
roles of monovalent (Na) and divalent (Ca) counterions, size of the incident beam on the sample is close to 300 µm.
as well as the ones of the free and condensed counter- The sample-detector distance was set at 1 m. This configu-
ration allowed us to perform measurements in the q range 0.02
ions. Å-1 < q < 0.2 Å-1, with a q resolution related to the beam size
on the sample and the beam divergence close to 0.003 Å-1. q
2. Experimental Methods is the magnitude of the scattering vector, defined by the
2.1. Materials. Hydrogenated and deuterated polystyrene wavelength of the incident beam λ and the angle between
(PS) chains with a narrow molecular weight distribution were incident and scattered beam θ such that q ) (4π/λ) sin(θ/2).
synthesized by anionic polymerization and then sulfonated The samples were held in calibrated mica cells of 1 mm
according to the Makowski et al. procedure.51-55 After neu- thickness, avoiding multiple scattering. Six concentrations
tralization with either sodium hydroxide (NaOH) or calcium were investigated at room temperature for each solute mixture
hydroxide (CaOH2), and tetramethylammonium hydroxide (X ) 0, 0.25, 0.5, 0.75, 1): c ) 0.0425, 0.085, 0.17, 0.34, 0.68,
(TMAOH) for deuterated chains, sodium (respectively calcium and 1.36 mol/L.
and tetramethylammonium)-sulfonated polystyrene (NaPSS, SANS experiments were performed at the Institut Laue
CaPSS2, and TMAPSSd) chains were purified by extended Langevin (Grenoble, France) on the D11 and D22 diffracto-
dialysis against pure water (conductivity of the order of 1 µS) meters. Measurements on D11 were done using two different
and obtained in powder after freeze-drying (note that CaPSS2 configurations (sample-to-detector distances: 4.10 and 13.50
and TMAPSSd macromolecules are only labeled CaPSS and m) covering a total q range 0.005 < q < 0.12 Å-1 at λ ) 4.51
TMAPSS hereafter). Their characterization was carried out Å (∆λ/λ ) 10%). To the subsequent 10% in ∆q/q have to be
by elemental analysis. In this way, the degree of sulfonation added the q resolutions related to the beam size on the sample
τs, defined as the ratio of sulfonated monomers to the total and the beam divergence, which are 0.003 and 0.001 Å-1, at
7458 Combet et al. Macromolecules, Vol. 38, No. 17, 2005
the respective sample-detector distances. A circular dia- In this relation m refers to macroions and c to counterions;
phragm of 10 mm in diameter imposed the size of the incident Km and Kc are the related contrast lengths. As Na-CaPSS
beam on the sample. The scattered intensity was recorded with aqueous solutions are studied by neutron scattering, counter-
a 2-dimensional 3He gas detector. Polyelectrolytes solutions ions can be neglected. The signal is then proportional to the
were held in a circular quartz container with a thickness of 4 macroion partial scattering function. For TMAPSS solutions
mm. The temperature was kept fixed during the whole in H2O, it is the same since the scattering length density of
experiment (T ) 22 °C). Only three concentrations were TMA (-0.64 × 1010 cm-2) is quite close to that of H2O (-0.56
investigated by using D11: c ) 0.0053 mol/L (X ) 0 and 1), c × 1010 cm-2). Contrarily, for TMAPSS solutions in D2O, the
) 0.0106 mol/L (X ) 0, 0.25, 0.5, 0.75, and 1), and c ) 0.0212 signal is proportional to the counterion partial scattering
mol/L (X ) 0, 0.25, 0.5, 0.75, and 1). Measurements on D22 function since the scattering length density of PSSd (6.39 ×
were done using two different configurations (sample-to- 1010 cm-2) matches the one of D2O (6.38 × 1010 cm-2). Now, as
detector distances: 1.35 and 14.40 m) covering a total q range Na-CaPSS aqueous solutions are studied by X-ray scat-
0.0025 < q < 0.5 Å-1 at λ ) 8 Å (∆λ/λ ) 10%). Sample thickness tering, the signal is no longer reduced to any partial scat-
was fixed to 1 mm. The other characteristics were close to the tering function, and we are bound to consider the three terms
previous ones (D11). Two kinds of solutions were then inves- of eq 4.
tigated: TMAPSS solutions in H2O and TMAPSS solutions in 2.3. Remarks on Other Scattering Measurements from
D2O. In this way, monomer (PSS) and counterion (TMA) the Literature. Besides our scattering results on mixed
partial scattering functions were measured. For each kind of valence systems, we will also recall in this paper various SAXS
TMAPSS solution, two concentrations were considered: c ) and SANS results reported in the literature in the case of
0.0425 and 0.34 mol/L. monovalent counterions only.60-64 They are related to salt-free
The scattering data were treated according to the usual solutions in good solvents of different dielectric constants and
procedures for isotropic small-angle scattering.58 After radial to linear flexible polyelectrolytes of distinct charge fractions,
averaging, the spectra were corrected for electronic noise of above and below the counterion condensation threshold. For
the detector, empty cell, absorption, and sample thickness. The all these polyelectrolyte solutions, we have considered the
normalization to the unit incident flux and the corrections of position in the reciprocal space, q*, of their scattering peak
geometrical factors and detector cells efficiency, for the SANS and studied its change as the concentration (in mol/L) is
data, were performed using the scattering of H2O. In the SAXS varied. The conversion from g/L (often used in the selected
experiments, the detector response was measured with the papers) to mol/L requires a precise knowledge of the mean
help of a 55Fe source. The data were set in absolute scale (with molar masses of the monomers, which depends on the charge
accuracy better than 10%) by using H2O as standard. In this fraction f and the water content τw of the initial dried samples.
way, the differential cross section per unit volume Σ(q) (cm-1) Unfortunately, these characteristics are not stated in all
was obtained for each solution. It is the sum of two terms: papers; therefore, a small uncertainty remains in the deter-
mination of the monomer concentration in mol/L. A brief
Σ(q) ) Σcoh(q) + ΣB(q) (1) description of the main features of the related systems is given
in Table 2. For some figures, we also consider additional light
Σcoh(q) is the coherent differential cross section containing all scattering results obtained from NaPSS salt-free aqueous
the information needed to describe the structure of the solutions65 by Drifford et al.
solution; ΣB(q) is the background which has to be removed from 2.4. Static Fluorescence. Pyrene with an excitation
∑(q) in order to obtain Σcoh(q). wavelength of 335 nm was used as fluorescent probe. Fluo-
For X-ray scattering, ΣB(q) is due to the scattering of the rescence spectra of pyrene molecules dissolved in NaPSS or
solvent. It is subtracted by using the differential cross section CaPSS aqueous solutions were recorded on a Hitachi F-4010
measured from the solvent alone ΣS(q) and by taking into spectrofluorometer between 350 and 500 nm. They allow to
account the volume fraction of the solvent in the solution ΦS; test the existence of possible hydrophobic microdomains.
i.e., we have Indeed, the vibronic fluorescence spectra of pyrene exhibit five
peaks (denoted 1 to 5), and the ratio I1/I3 of the intensities of
ΣB(q) ) ΦSΣS(q) (2) the first and third peaks depends on the polarity of the
environment of the pyrene molecules. It is close to 1.8-1.9 in
Such an approach is only approximate since the scattering water, at room temperature, and reaches 0.9 in nonpolar
related to density fluctuations of free counterions is neglected. solvents. This ratio was measured as a function of the
However, it is a rather good approximation since Σcoh(q) is large polyelectrolyte concentration.
compared to ΣB(q) in the q range explored here, which is mainly
concerned with the position of the maxima of the scattering 3. Results
functions. 3.1. Scattering Profiles at Low Concentrations
For neutron scattering measurements, ΣB(q) also involves
(c < 0.1 mol/L). Typical SANS and SAXS measure-
the incoherent scattering of the solute (macroion plus coun-
terions) that can be written as cNσinc/4π in which c is the ments at two distinct concentrations in the semidilute
monomer concentration (mol cm-3), σinc the monomer incoher- regime (c ) 0.0212 and 0.0425 mol/L) are presented in
ent cross section (cm2), and N Avogadro’s number (mol-1). Figure 1.
Neglecting inelastic effects, since the wavelengths of the used The scattering profiles exhibit a clear broad maximum
incident beams are rather low, σinc can be computed from the as encountered with salt-free aqueous solutions of
values listed in the literature.59 The background is then NaPSS or CaPSS in former works.66,60-69 We call q* the
described by eq 3. abscissa of this peak. One stringent feature is the q*
variation with X, i.e., the counterion valence. On one
ΣB(q) ) ΦSΣS(q) + cNσinc/4π (3) hand, q* is nearly constant for X ) 0.5, 0.75, and 1.
At this stage, to make more meaningful the comparisons
These three mixtures show almost identical scattering
between several solutions of the same polymer, we consider profiles and therefore should exhibit very similar struc-
the ratios of the coherent differential cross sections to the tures. On the other hand, for X ) 0.25 and 0, q* is
concentration Σcoh(q)/c (cm2/mol). shifted to lower q values while the intensity of the peak
Since we have a multicomponent solute made of large increases. Since the PSS macroions are identical for all
macroions and small counterions, Σcoh(q) actually involves the samples, the origin of these modifications has to be
three partial scattering functions: related to changes in the valence of the counterions.
As previously mentioned, the scattered intensity is
Σcoh(q) ) Km2Smm(q) + Kc2Scc(q) + 2KmKcSmc(q) (4) mainly related to macroions for neutron scattering while
Table 2. Characteristics of the Samples Used To Validate the Scaling Theory and the Manning-Oosawa Approach of the
Counterion Condensation for Polyelectrolyte Solutions in the Presence of Monovalent Counterionsa
sample solvent lB f ζ ref
NaPSS (X ) 1) H2O 78 7.14 0.90 2.51 this work
NaPSS* (X ) 1) (H2O)0.75 + THF0.25 60 9.22 0.90 3.21 this work
TMAPSS H2O or D2O 78 7.14 1 2.79 this work
NaPSS H2O 78 7.14 1 2.79 65
AMANPS (95%) H2O 78 7.14 0.95 2.65 63
AMANPS (60%) H2O 78 7.14 0.60 1.67 63
AMANPS (51%) H2O 78 7.14 0.51 1.42 63
MPCP DMF 36.7 15.3 1 5.98 64
MPCP PC 64.9 8.6 1 3.36 64
PMVP EGD 37.7 14.9 0.45 2.62 62
PMVP D2O 78 7.14 0.45 1.25 62
NaAD10 H2O 78 7.14 0.015 0.04 60
NaAD17 H2O 78 7.14 0.07 0.19 60
NaAD27 H2O 78 7.14 0.17 0.47 60
NaAD37 H2O 78 7.14 0.27 0.75 60
NaAD60 H2O 78 7.14 0.35 0.98 60
NaPVS (R ) 5.4%) H2O 78 7.14 0.054 0.15 61
NaPVS (R ) 10.7%) H2O 78 7.14 0.107 0.30 61
NaPVS (R ) 23.2%) H2O 78 7.14 0.232 0.65 61
NaPVS (R ) 31%) H2O 78 7.14 0.31 0.86 61
NaPVS (R ) 49.9%) H2O 78 7.14 0.499 1.39 61
a The monomer size b is 2.56 Å for all the samples; f is the fraction of charged monomers, or the charge fraction; the mean distance
between charges along the chemical sequence of the macroions is A ) b/f; is the dielectric constant; lB is the Bjerrum length (Å) for the
solvent considered, at T ) 25 °C; ζ ) lB/A is the Manning charge parameter. AMAMPS are samples of poly[acrylamide-co-sodium-2-
acrylamido-2-methylpropanesulfonate]; MPCP, a sample of atactic poly[2-(methacryloyloxy)ethyltrimethylammonium 1,1,2,3,3-pentacy-
anopropenide]; PMVP, a sample of poly[N-methyl-2-vinylpyridinium chloride]; NaAD series, sodium salts of acrylamide-acrylic acid
copolymers of distinct compositions; NaPVS, sodium salts of partially sulfuric acid esterificated poly(vinyl alcohol). The solvent DMF is
N,N-dimethylformamide; EG, ethylene glycol; PC, 1,2-propylene carbonate. NaPSS* has been dissolved in a solvent which is a mixture
of H2O and THF (tetrahydrofuran) with volume fractions of 75 and 25%, respectively. The corresponding dielectric constant has been
evaluated from those of its components, (H2O) and (THF), using the relationship ) 0.75(H2O) + 0.25(THF).
it is due to both counterions and macroions for X-ray (ii) For these lowest concentrations, the prefactor of
scattering. So, the peak intensity for X-ray scattering the c1/2 scaling law is almost identical for X ) 1, 0.75,
is very sensitive to the fraction as well as the nature of and 0.5 (a careful inspection for X ) 0.5 however denotes
condensed counterions. a small decrease) whereas a large discrepancy is ob-
3.2. Scattering Profiles at Higher Concentra- served for X ) 0.25, for which the prefactor is lowered,
tions (c > 0.1 mol/L). As the concentration is increased, and for X ) 0, for which it is even lower.
new features are observed (Figure 2). The scattering (iii) For higher concentrations (c > 0.1 mol/L), devia-
curves are now all different for all mixtures. Both the tions from the c1/2 scaling law are observed for all
position q* and the intensity I(q*) of the scattering peak mixtures containing Ca counterions. This trend is more
are affected: q* is shifted to lower q values, and I(q*) pronounced as the fraction of Ca counterions is in-
increases when X is varied from 1 to 0. creased; the scaling exponent being reduced from 0.5
The structure of the aqueous solutions is now strongly (the exact value being less than 0.5 as discussed later)
dependent on X. Upon further increase in concentration, to 0.2 as X varies from 1 (100% Na) to 0 (100% Ca).
the scattering peak gradually disappears, provided Ca (iv) For the highest concentrations and X ) 0, 0.25,
counterions are present, i.e., for X < 1. This effect is and 0.5, the peak disappears. The critical melting peak
more and more pronounced when X decreases. That concentration is highly sensitive to the presence of
point is illustrated in Figure 3 for X ) 0, i.e., for CaPSS divalent counterions and is displaced to lower c values
solutions. The scattering peak vanishes above c ) 0.68 when X is decreased.
mol/L. The scattering curves then look very similar to
To summarize sections 3.1-3.3, all observations
those of neutral polymers. For X ) 0.5, the peak
strongly suggest an important role of the counterion
disappears only around c ) 1.36 mol/L: the peak
valence on the structure of the polyelectrolyte solutions.
melting concentration is shifted to larger c values.
However, at low concentration, it is striking to see how
3.3. Variations of q* with Respect to c. The
valence effects are absent or masked (we will see why
variations of q* with respect to c for distinct X values
below) for the higher X values (1, 0.75, and 0.5).
are shown in Figure 4. The main features are the
following:
4. Theoretical Background: Condensation and
(i) For the lowest concentrations (inset in Figure 4),
Scaling
q* scales as c1/2 within experimental uncertainties for
all the solutions. This behavior is in accordance with Let us now recall the accepted picture for linear
scaling predictions for polyelectrolytes in good solvent polyelectrolytes in the presence of monovalent and or
conditions; it has already been observed for pure divalent counterions, at two levels: first, the counterion
NaPSS.65-69 It must be noted that other measurements condensation process; second, the scaling description of
(data not shown), performed on NaPSS with a still macroions and, more especially, that of the character-
higher degree of sulfonation (τs ) 1), give quite identical istic scattering peak of their semidilute solutions. In a
results (equal q* values). Thus, we will neglect the effect third part, we will show how the combination of coun-
of a nonperfect sulfonation in our discussion of the data terion condensation and scaling improve data agree-
and will consider our PSS samples as fully sulfonated ment with the scaling model, in the case of monovalent
ones. counterions.
Figure 3. SAXS profiles of salt-free aqueous solutions of pure

CaPSS polyelectrolytes (X ) 0), recorded at different monomer
concentrations.
Figure 1. Scattering from salt-free aqueous solutions of mixed

NaPSS and CaPSS polyelectrolytes. (a) SANS profiles recorded
at c ) 0.0212 mol/L for different fractions of NaPSS polyelec-
trolytes (X ) 0, 0.25, 0.5, 0.75, and 1). (b) SAXS profiles Figure 4. Scattering vector q* vs monomer concentration c
recorded at c ) 0.0425 mol/L for different fractions of NaPSS for different fractions of NaPSS polyelectrolytes (X ) 0, 0.25,
polyelectrolytes (X ) 0, 0.25, 0.5, 0.75, and 1). 0.5, 0.75, and 1). For X ) 0 and 0.25, the maximum of the
scattering function is no longer really observable as c > 0.34
mol/L. The insert is concerned with the results obtained for
the lowest concentrations (c < 0.05 mol/L). Dotted lines
correspond to a c1/2 evolution.
experimental and theoretical points of view.40,41,53,63 It

is therefore necessary to examine this process for any
X value.
For identical counterions of valence Z, according to
Manning, a condensation process takes place when the
charge parameter ζ ) lB/A is larger than 1/Z.40 As
defined in Table 2, lB is the Bjerrum length and A is
the mean distance between charges along the chemical
sequence of the macroions. The linear charge density
is then renormalized to a constant value such that A )
ZlB or ζ ) 1/Z. The condition ζ > 1/Z or lB/A > 1/Z, that
is also the strong-coupling limit, corresponds to A < lB
for monovalent counterions and A < 2lB for divalent
Figure 2. SAXS profiles of salt-free aqueous solutions of counterions. Such conditions are always fulfilled with
mixed NaPSS and CaPSS polyelectrolytes, recorded at c ) 0.34 our systems, and thus there is always a charge renor-
mol/L for different fractions of NaPSS polyelectrolytes (X ) 0, malization tending to reduce ζ to the net value 1/Z. The
0.25, 0.5, 0.75, and 1). charge fraction f (equals to 1 in the absence of any
counterion condensation, i.e., one electrostatic charge
4.1. Counterion Condensation: The One and every monomer of size b) is therefore decreased to the
Mixed Valence Cases. We have investigated different effective value feff ) b/ZlB. There should be one effective
mixtures of NaPSS and CaPSS. The relevant parameter charge every lB for pure NaPSS (X ) 1) and one effective
is the fraction of NaPSS chains, X, that varies from 0 charge every 2lB for pure CaPSS (X ) 0). Hence, we can
(pure CaPSS) to 1 (pure NaPSS). Since fully sulfonated view counterions as distributed in two populations:
polystyrene is a highly charged macroion, a counterion condensed counterions which are confined in small
condensation process is expected to take place from most volumes around the macroions by the macroion poten-
tial well (which let them some mobility along the

macroions) and free counterions which escape the mac-
roion potential well and scatter in the solution.41
For mixtures of monovalent (Na) and divalent (Ca)
counterions, the condensation process is more com-
plex.44,70 Counterions with the higher valence will
condense first and the Manning charge parameter will
be reduced to ζ′ according to the amount of condensed
counterions:
(i) If there are enough divalent counterions to reduce
ζ to ζ′ ) 1/2, then the condensation process stops and
the excess of Na and Ca counterions remains free in the
solution. Among the set of solutions investigated here,
this is only the case for the ones corresponding to X )
0 (pure CaPSS). The effective charge fraction is there-
fore feff ) b/2lB (0.18 in our particular case).
(ii) For solutions of larger X values (X ) 0.25, 0.5,
0.75), different regimes may appear. If there are not
enough divalent counterions to achieve ζ′ ) 1/2 but
enough to make 1/2 < ζ′ < 1, all Ca counterions will be
condensed whereas all Na counterions will remain free
in the solution. That is the case for X ) 0.25. In that
mixture, the condensation of all the Ca counterions
leads to one effective charge every four monomers
(approximately one effective charge every 10 Å along
the PSS chemical sequence). Here, we are within an
interesting regime where one can gradually change the
effective charge, with only Ca species condensed on the
macroion. Actually, the effective charge fraction becomes
equal to X.
(iii) Finally, if after condensation of the Ca counter-
ions ζ′ is still larger than 1, then Na counterions will Figure 5. Main characteristics of Na-CaPSS mixtures,
also condense to reduce ζ′ to ζ′′ ) 1 (one electrostatic according to the Manning-Oosawa approach. (a) Effective
charge fraction feff vs fraction of NaPSS polyelectrolytes X.
charge every lB ) 7.14 Å in water). Part of Na counter- Open circles correspond to experimental X values of the
ions will be condensed; the other part will be free. This present work. (b) Schematic representation of condensed and
is expected for X ) 0.5, 0.75, and 1. This is another free counterions for different fractions of NaPSS polyelectro-
surprising regime: although the ratio of monovalent to lytes (X ) 0, 0.25, 0.5, 0.75, and 1). Open and filled circles
divalent condensed counterions is changed according to correspond to Na and Ca counterions, respectively. Condensed
X, the effective charge fraction (feff ) b/lB, i.e., close to counterions are located in the shadow area around the
macroion. Note that the number of Na and Ca counterions is
0.36) and the amount of free counterions (all monova- only indicative. In this schematic representation, the macroion
lent) are constant. contains 24 charged monomers. X ) 1; the free and condensed
To sum up, the expected variation of the effective counterions are Na counterions; X ) 0.75, 0.5, and 0.25; the
charge fraction feff as a function of X is shown in Figure free counterions are Na counterions; X ) 0.75 and 0.5; the
5a. Moreover, a schematic picture is proposed in Figure condensed counterions are mixtures of Na and Ca counterions;
5b. X ) 0.25 and 0; the condensed counterions are Ca counterions;
X ) 0; the free and condensed counterions are Ca counterions.
We see now the interest of the mixed valence sys-
tems: they should allow us to gradually change the preferred average distance between the centers of mass
mean distance between effective charges along the of the macroions. In the semidilute regime (c > c*), the
chemical sequence of the macroions (within a certain X macroions interpenetrate each other. A single broad
range) and the proportion of Na and Ca counterions in scattering peak is still present but it scales now as c1/2.
the condensed and free counterion populations. In this The structure of the semidilute solutions has been
way, the role of the effective charge fraction as well as analyzed in terms of an isotropic net of rods by de
the nature of condensed and free counterions on the Gennes.48 The mesh size of this net, ξ, is the only
structure of the solutions can be probed. characteristic length. Specifically, it is the screening
This Manning-Oosawa description is strictly valid length of the electrostatic interactions. Thus, for dis-
only for an infinite dilution together with a rodlike tances smaller than ξ, the electrostatic interactions are
conformation of the macroions. However, it is in ac- dominant, and the average conformation of the macro-
cordance with experiments performed on flexible polyions remains highly stretched. For distances larger than
electrolytes in a finite concentration range.53,61,63,71,72 ξ, the electrostatic interactions are screened, and hence
Thus, we will consider it as a reasonable approximation. the average conformation of the macroions is a random
4.2. Scaling Description of Charged Linear Poly- walk of correlation blobs of size ξ. A scaling argument,
mers. In the dilute regime (c < c*), the average similar to that made for neutral polymers in a good
conformation of the macroions is highly stretched due solvent, allows to demonstrate that ξ scales as c-1/2.48,49
to electrostatic interactions. On the other hand, the From an experimental point of view, insight into the
scattering experiments reveal the existence of a peak structure of the semidilute solutions will be gained
at a position q* that scales as c1/3. This peak can be through the position of the single broad scattering
interpreted as the signature of a position order, i.e., a maximum that scales as ξ-1. Such a maximum is indeed
associated with an electrostatic correlation hole, each

part of a given macroion being surrounded by an
excluded volume of size ξ from which other macroions
are partly expelled.67,73
The scaling approach of de Gennes has been reviewed
by Dobrynin, Colby, and Rubinstein.50 The main new
feature is the introduction of a parameter that takes
the solvent quality into account. It is also emphasized
that the Debye screening length (due to free counterions
alone), κD-1, could not correctly describe the electrostatic
screening in semidilute solutions since κD-1 is shorter
than the average distance between macroions, ξ. In a
good solvent condition, the correlation or screening
length ξ can be written as
ξ ) f -2/7(lB/b)-1/7(bc)-1/2 (5)
Here, f is the charge fraction of the polyelectrolytes.

A network of interpenetrated rods can crudely sche-
matize this scaling model. The c dependence is then
related to the average distance between rods; mean-
while, the f dependence is introduced by considering
that the rods are alignments of electrostatic blobs, which
defines the mass per unit length of these rods. The
electrostatic blob size scales as f -6/7 and f -2/3 for good
and Θ solvent, respectively.
So, the scaling model mainly applies to weakly
charged polyelectrolytes for which the charge fraction
is the chemical charge fraction of the macroions. How-
ever, it is usually assumed also to apply to highly
charged polyelectrolytes. In the weak-coupling limit
(ζ < 1/Z), eq 5 remains valid. In the strong-coupling
limit (ζ > 1/Z), the macroions are then treated as weakly Figure 6. Scattering vector q* vs c and 2π/ξ for various
charged polyelectrolytes with an effective charge frac- flexible polyelectrolytes in the presence of monovalent coun-
tion feff taking into account the condensation of part of terions. (a) q* vs c; (b) q* vs 2π/ξ. The counterion condensation
the counterions. In this case, we propose to replace phenomenon has been taken into account according to the
Manning-Oosawa approach. Dotted line indicates the strict
relation 5 by equality q* ) 2π/ξ.
ξ ) feff-2/7(lB/b)-1/7(bc)-1/2 (5′) In this way, we take into account the counterion

condensation through the Manning-Oosawa theory.
Obviously, for the mixed valence systems (Na-CaPSS), If ξ correctly describes the systems, plotting q* as a
ξ therefore depends on X. Equations 5 and 5′ mostly function of 2π/ξ instead of simply the concentration
show that the screening length ξ and therefore the should lead to a master curve. The q* values for all the
position of the broad scattering peak q* depend on the polyelectrolytes listed in Table 2 are presented as a
charge fraction of the macroions (or their effective function of c in Figure 6a. The same data are reported
charge fraction when a condensation of part of the as a function of 2π/ξ in Figure 6b.
counterions takes place) and concentration. But they A master curve, linking q* and 2π/ξ, is clearly
also depend on the dimensionless parameter lB/b. obtained over more than 4 decades in concentration,
4.3. Scaling and Condensation: Comparison provided the counterion condensation process is taken
with Experiments Involving Monovalent Coun- into account. The scaling model is therefore able to
terions. In this section we focus on the case of monova- explain small modifications when changing from one
lent counterions in order to test the validity of the solvent to another as well as variations with the charge
scaling approach combined with condensation. density, when combined with condensation.
4.3.1. Master Curve. We will consider various flex- If the condensation of counterions is not accounted
ible macroions in good solvent and will compare the q* for, neither the nearly constant values of q* encountered
values measured by us or extracted from the literature in AMANPS (for different charge fractions) nor the
with the theoretical expression 5 for 2π/ξ, using the appearance of a plateau in NaPVS for the highest values
sample characteristics of Table 2. More precisely, ξ is of the dissociation degree (R) can be explained. More-
calculated from eq 5 in the weak-coupling limit (ζ < 1/Z), over, nor can the q* values for PMVP samples in
while it is rather obtained from eq 5′ in the strong- different solvents (EG, D2O) or MPCP in DMF and PC
coupling limit (ζ > 1/Z). Since only monovalent coun- be correctly rescaled.
terions are considered, feff ) b/lB, and eq 5′ can be Looking more in detail, we can point out a small
written as dispersion in the data forming the master curve in
Figure 6b. This is undoubtedly related to uncertainties
in scattering measurements as well as in the charac-
ξ ) (lB/b)1/7(bc)-1/2 (5′′) teristics of the polyelectrolyte solutions necessary to
evaluate ξ (Table 2). Furthermore, as will be seen in

the next subsection, scattering length density of both
counterion and macroion may also play a relevant role
in the data dispersion.
In this approach, we decided not to use the Debye
length as the typical length scale of the system. For
NaPSS solutions, the Debye length κD-1 (using κD2 )
4πlBI and the ionic strength I ) ∑iZi2ci) caused by the
free counterions is always smaller than the average
distance between chains (extracted from the relation d
) 2π/q*). This remains correct if the condensation is no
longer taken into account. For instance, κD-1 is close to
35 Å at c ) 0.0425 mol/L (and 9 Å at c ) 0.68 mol/L),
whereas the average distance between chains is close
to 175 Å (50 Å at 0.68 mol/L). Thus, the use of κD-1 as
the screening length is not appropriate without added
salt since electrostatic interactions occur at distances Figure 7. SANS profiles of salt free solutions of TMAPSS in
well above the Debye length. Furthermore, the Debye H2O and D2O. They are proportional to the macroion and
length is even smaller than the average distance counterion partial scattering functions (Smm(q) and Scc(q),
between counterions (with or without taking the con- respectively). Two monomer concentrations are considered: c
densation into account). However, the fact that κD and ) 0.0425 and 0.34 mol/L. For each monomer concentra-
tion, the scattering vector q* of Scc(q) is lower than the one of
ξ-1 scale as c1/2 demonstrate that both quantities are Smm(q).
proportional to each other. Thus, rescaling q* as a
function of κD instead of ξ-1 should lead to similar
results. In fact, this is not the case. For monovalent be compared to 2π/ξ is actually the one of the intermo-
counterions, κD is equal to (4πlBf)1/2c1/2 (without conden- lecular scattering function.75 By measuring the total
sation) or (4πb)1/2c1/2 (with condensation). Thus, modi- scattering function, involving the form factor as well as
fications of the solvent (lB) or the charge fraction f do the intermolecular correlations, discrepancies are ex-
not change κD and ξ-1 a similar way. It is known from pected. They are different for Scc(q) and Smm(q) because
an experimental point of view that q* scales as f 1/3 (or the form factor of the sheath of the condensed counter-
f 2/7). Such a result is not compatible with the Debye ions around one macroion differs from that of the
length description. macroion itself. A sharper decrease in the sheath form
factor shifts the maximum of the total scattering func-
4.3.2. Deviations at Higher Concentrations. A tion toward smaller q values. Since this decrease is
more systematic effect is however visible from closer marked at high q values, it leads to a larger shift as
inspection: the strict equality q* ) 2π/ξ is perfectly the concentration increases. Remarkably, discrepancies
obeyed only in a lower c range. Deviations toward lower in q* measurements are negligible as the macroion
q* values are observed beyond a certain c (or q*) value. (PSS) partial scattering functions are considered. On the
They correspond to deviations from the usual c1/2 scaling contrary, when the counterion partial scattering func-
law, which gradually increase with c. A first cause could tions are involved, they cannot be neglected. This is
be a progressive increase in the counterion condensa- especially the case for SAXS experiments. We finally
tion, leading to a continuous decrease in the effective performed SAXS measurements on TMAPSS (data not
charge fraction.20,74 In Appendix A, we rule this out shown). This counterion also presents remarkable prop-
because it leads to unrealistic changes in the effective erties toward X-ray scattering since the total scattering
charge fraction and/or the counterion condensation function is very close to Smm. In that case the c1/2 scaling
threshold. We can notice that the observed deviations law is perfectly obeyed. Furthermore, the position of the
depend on the explored q range. They are negligible at maximum q* is identical to the one observed with
low spatial resolution (low q), i.e., in light scattering neutrons in the Smm structure function.
experiments, and increase at high q, as the resolution We definitively see the problems in the q* values
is improved. They also depend on the contrast and/or determined from the total scattering function. According
the kind of radiation used for the experiments. For to the contrast of both counterions and polyions, q* may
example, with NaPSS salt-free aqueous solutions, such vary within a rather large interval (up to 10%). The
deviations are observed by SAXS whereas they are polyelectrolyte under investigation and the scattering
almost unobservable by SANS when D2O is used as technique also play a role in the data dispersion
solvent. This dependence on the contrast can be defi- observed in the previous section.
nitely demonstrated by comparing SANS results ob- 4.3.3. Size of the Electrostatic Blobs. The use of
tained with solutions of TMAPSS in H2O and D2O, at the scaling approach for highly charged polyelectrolytes
the same concentration. This comparison, which turns assumes that, after counterion condensation, macroions
out to compare the macroion and counterion partial simply behave as if they were partially charged and
scattering functions (Smm(q) and Scc(q), respectively), is even weakly charged though the actual charge fraction
made in Figure 7. can still be relatively large (0.36 for NaPSS). This can
Obviously, the q* of Scc(q) is lower than the one of be questionable. A key point is that the weakly charged
Smm(q). This is consistent with computer simulations polyelectrolyte model lies upon the concept of electro-
performed on salt-free solutions of rigid polyelectro- static blobs. In practice, even if the counterion conden-
lytes.82 This shift also depends on concentration: it is sation is taken into account, electrostatic blobs should
rather weak for low concentrations but increases with be rather small and thus contain only a few monomers
c. It is caused by a modification of the form factor rather (of the order of 3-4 monomers for NaPSS). The deriva-
than the intermolecular term. Indeed, the q* that should tion of the expression of ξ, in particular the different
exponents associated with f, lB, and b, lies upon the

assumptions about the statistic of the chain inside the
electrostatic blobs: e.g., the relation between the size
and the number of units is the same as in a self-avoiding
chain. This assumption may fail for chains of a few
monomer units. Despite this, the existence of a master
curve seems to validate the scaling description.
Hence, we reach a first intermediate conclusion:
except for the higher concentrations where other scat-
tering effects may come into play, the scaling approach
combined with condensation correctly describes the
polyelectrolyte peak position in the case of monovalent
counterions.
5. Na-CaPSS Mixtures: Interpretive Discussion

We have just seen that Figure 6b evidences the
existence of condensation threshold and charge renor- Figure 8. Scattering vector q* vs 2π/ξ for different fractions
malization effects on scattering curves in the case of of NaPSS polyelectrolytes (X ) 0, 0.25, 0.5, 0.75, and 1). The
line corresponds to the strict equality q* ) 2π/ξ. The dotted
monovalent counterions. Let us now return to the line corresponds to q* ) R/ξ with R < 2π.
measurements reported in this paper on mixtures of Na
and Ca counterions. Obviously, they open the way of only parameter that can be involved for CaPSS
modifying the effective charge fraction without changing (X ) 0) aqueous solutions, even in the low concentration
the chemical structure of the macroions. We will thus regime.
compare our scattering results on these mixtures in the At higher concentrations (c > 0.1 mol/L), there are
light of the scaling-condensation predictions. new features: q* is now different for each X value,
5.1. Qualitative Interpretation. Let us first com- which is not compatible with our description, and is
pare the scattering curves of Na-CaPSS mixtures followed by a modification in the scaling exponent that
according to their state of counterion condensation. is more pronounced when the proportion of Ca coun-
In the low concentration regime (c < 0.1 mol/L), terions is increased. We recall that all divalent coun-
our measurements show that q* is nearly identical for terions are supposed to be condensed for X ) 0.25, 0.5,
X ) 1, 0.75, and 0.5 and decreases for X ) 0.25 and 0. and 0.75 mixtures.
According to the scaling-condensation approach, this In summary, for the higher c regime (c > 0.1 mol/L)
suggests similar effective charge for the three higher X and contrarily to the lower c regime, the nature of the
values and a reduction for the two others. These condensed counterions plays a role that cannot be only
deductions are in accordance with our description of the explained by a charge renormalization as predicted by
counterion condensation process (section 4.1) which the Manning-Oosawa theory. Pure NaPSS is not really
predicts the same effective charge fraction for X ) 1, affected, at least within the investigated c range,
0.75, and 0.5 as well as smaller values for X ) 0.25 and although the counterion condensation is expected to be
0. Moreover, these deductions are reinforced by the slightly and continuously improved as the concentration
analysis of the height of the scattering peak I(q*). To is increased.20 The important influence of counterion
make it quantitative, let us recall that the contrast in valence on the structure of polyelectrolyte solutions is
X-rays is partly due to the condensed counterions or, therefore revealed. We will pursue this discussion in
more precisely, to their electronic density. Exchanging section 5.4.
Na for Ca counterions (as expected when changing the 5.2. Comparison with Scaling Predictions. In
X value) is ruled by a particular balance: the electronic section 4.3, we have shown that the correlation length
charge of Ca (18 e-) is about twice that of Na (10 e-), ξ developed in the scaling model is able to correctly
which is similar to the ratio between their ionic charge describe the evolution of q* for different flexible poly-
(2/1); changing one Ca (18 e-) for two Na counterions electrolyte solutions, at least in the case of monovalent
(2 × 10 e-) does not greatly affect the total number of counterions. We can now test this approach for poly-
electrons. For X ) 1, 0.75, and 0.5, since the effective electrolytes in the presence of mixed valence counterions
charge is constant, one condensed Ca counterion has to (specifically, mixtures of monovalent and divalent coun-
be replaced by two Na ones. Thus, the number of terions). The presence of divalent counterions will be
electrons in the condensed cloud of monovalent and taken into account through the effective charge fraction,
divalent counterions is roughly constant. Because the feff, which we introduce in the scaling model.
molar volumes of both counterions are almost identical, The evolutions of q* for distinct X values are pre-
their electronic density is constant (their contrast sented in Figure 8 as a function of 2π/ξ, calculated
lengths in water are similar). Hence, the fact that the according to eq 5′.
height of the scattering peak does not greatly change is Figure 8 shows that the scaling model accounts rather
a proof of an identical structure and may validate our well for SAXS measurements in the low concentration
condensation description at least for the highest X regime, except for pure CaPSS (X ) 0). In a first
values. approximation, these observations may validate the
The scattering features therefore provide a clear Manning-Oosawa description of the counterion con-
signature of a charge renormalization. A simple descrip- densation process for mixtures of monovalent and
tion of the condensation process associated with the divalent counterions, except for X ) 0.
scaling approach enables a qualitative interpretation of For pure CaPSS, i.e., X ) 0, it is obvious that another
the scattering curves. However, as will be shown in phenomenon is present which is not properly taken into
the next sections, the effective charge fraction is not the account by the scaling approach. We can therefore argue
that, except for pure CaPSS, the valence of the con- roions. There is more reliable information on the second,
densed counterions has practically no effect on the which we discuss now.
structure of the solutions other than the one related to 5.3.2. Conformational Changes. So far, we have
a charge renormalization, in agreement with the Man- considered that the only effect of the counterions was a
ning-Oosawa theory. modification of the charge fraction. However, a counter-
Hence, we reach a second intermediate conclusion: in ion-induced attraction between monomers may also
the low concentration regime, the scaling-condensation occur.9,10,14,31 Such a phenomenon that is more pro-
model correctly describes the polyelectrolyte peak posi- nounced with divalent counterions could be responsible
tion for a mixture of monovalent and divalent counter- for a retraction of the macroions. Then, the shift in q*
ions except in the range of high Ca fractions (X close to for pure CaPSS would be due to a collapse of the
zero). macroions, which in turn induces additional condensa-
5.3. Distinctive Features of the Large Ca Frac- tion and thus a decrease in the effective charge fraction.
tions (X e 0.25) in the Low Concentration Regime. Neutron scattering measurements of the radii of gyra-
The discrepancy for pure CaPSS (X ) 0) may have a tion of NaPSS and CaPSS polyelectrolytes have been
different origin. carried out by use of the zero average contrast (ZAC)
5.3.1. Main Scattering Results for X ) 0. Neutron method.16,17 They show that the macroions related to
scattering experiments on pure NaPSS and pure CaPSS CaPSS samples were shrunk with respect to those
aqueous solutions have been already performed.16,17 associated with NaPSS samples (Rg ratio of the order
They show that q* is shifted to lower q values for CaPSS of 1.5-2). The scaling approach cannot account for such
in accordance with our measurements. The average size a difference via a charge renormalization according to
of the macroions was also found to be smaller in the the Manning-Oosawa theory alone. This implies that
presence of divalent counterions compared to mono- the overlap concentration c* will not be correctly evalu-
valent counterions. These results, which have been ated for CaPSS samples. More generally, the slightly
confirmed by simulations,25 were interpreted by Zhang collapsed conformation through Ca bridges should: first,
et al.16 as a consequence of a charge renormalization, enhance the counterion condensation process; second,
namely a reduction in the effective charge fraction lead to a higher c* value that shifts the scattering peak
similar to the one encountered when the degree of toward lower q values and changes the exponent in the
sulfonation of the PSS macroions is decreased, i.e., when c dependence of the correlation length ξ.
the macroion becomes more hydrophobic.52,53,63,71,76 Now, one has to explain why a shrinking of macroions
Conversely, Dubois et al. argued that they were due to is only observed for X ) 0 (pure CaPSS) and not for X
a conformational change resulting from bridging-type ) 0.25 or for the other Na-CaPSS mixtures in which
intramolecular correlations between monomer units divalent counterions are also condensed (X < 1). At this
through divalent counterions, without implying any stage we wish to emphasize that a similar effect is
hydrophobic character. contained in the SANS study of Dubois et al.17 As just
In the frame of the scaling approach, the charge said, for pure CaPSS solutions without any added salt,
fraction or the effective charge fraction determines the the average conformation of macroions cannot be fitted
size of the electrostatic blobs. Any modification of that to the wormlike chain model, indicating a more shrunk
parameter may change the size of the extended chains conformation. But under addition of CaCl2 to NaPSS
in dilute solution and also the value of the overlap solutions, the chain conformation can still be viewed as
concentration c*. This is responsible for a shift of the wormlike, as long as the CaCl2 added salt concentration
position of the maximum q* values (as shown from is low enough (the persistence length of the macroions
relations 5 and 5′) as well as changes in the radius of decreases when the ionic strength is increased as usual).
gyration. If the counterion condensation process (and, The role of the salt is out of the scope of that paper, but
therefore, the effective charge of the macroion) certainly we can consider in a first approximation that part of
plays an important role as mentioned by Zhang et al., the added Ca cations replace the initially monovalent
it is thus not necessary to invoke any hydrophobic condensed counterions while another part is scattered
property to describe the q* behavior. Nevertheless, this in the aqueous solution with the Na counterions.
effect is not large enough in the scaling model to Therefore, solutions with low added salt concentration
correctly explain the experimental observations. We will correspond to our X range, 0.25 e X < 1 (no free divalent
return to this discussion below; before we would like to counterions), whereas large concentrations correspond
stress the main difference between X ) 0 (pure CaPSS), to X < 0.25 (free divalent counterions). We see that here
where the discrepancy is pronounced, and X ) 0.25, also aqueous solutions with free Ca counterions only
where it is marginal. For X ) 0.25, which is very close exhibit the conformational shrinking, in agreement with
to 0.18, the nature of the free counterions is different. our measurements.
Pure CaPSS is the only one among our systems for It is remarkable that this chain retraction is just
which some divalent counterions are free in the aqueous observed as free Ca counterions are free in the solution.
solutions. It is also paradoxical because one electrostatic link
Here, we reach a third intermediate conclusion: it is should be more easily formed from a divalent counterion
mostly when some divalent counterions are free in the closely localized to at least one part of the chain, leading
solution that we observe disagreement with the scal- to a charge reversal. However, the coincidence of the
ing-condensation model. chain shrinking with the presence of free Ca counterions
The scaling approach does not explicitly take into in the solution may be not fortuitous. Actually, electro-
account the nature of the free counterions. Therefore, static bridges or loops, if they exist, could only be formed
it may not be valid in this particular case if counterions through some exchanges between condensed and free
have some additional influence. This influence could act counterions. Since in addition divalent counterions have
on macroion interactions, through the electrostatic to be involved in such processes, the condition X < 0.25
screening, or on the average conformation of the mac- (or 0.18) is then required.
To proceed further in this debate would be facilitated as the effective charge fraction) and cause deviation
by numerical simulations. Few have been performed on from the c1/2 scaling law.
flexible polylectrolytes in the presence of multivalent As mentioned above, in this concentration range, the
counterions. They indicate shrinkage in the macroion structure of the solutions is highly sensitive to the
conformation.25,42 Specifically, they enlighten the exist- nature of condensed counterions. In the previous part
ence of intramolecular bridges formed by divalent of this work, we emphasized the importance of the free
counterions. They consider the macroion partial scat- divalent counterions in the bridging process. Bridging
tering function that exhibits a broad maximum whose phenomena observed for pure CaPSS could be enhanced
the position q* depends on the charge fraction.25 In the at high concentration due to the reduction of the
dilute regime, q* is found to be identical for monovalent persistence length. This kind of interaction could also
and divalent counterions (for all charge fractions) and be conceivable for other mixtures even in the absence
scales as c1/3. In the semidilute regime, q* is found to of free divalent counterions, provided that the concen-
scale as c1/2 for both counterion valences Z ) 1 and 2, tration is high enough.
but the prefactor is found to vary with Z and the charge
fraction. For divalent counterions, the prefactor is lower 6. Summary and Conclusion
than for monovalent counterions. Numerical simula- This study brings some new insights for understand-
tions are therefore in agreement with our measure- ing the so-called “polyelectrolyte peak” encountered in
ments in the low concentration regime. Moreover, they the small-angle scattering from salt-free polyelectrolyte
show that a master curve can be obtained by rescaling solutions, using the trick of mixing monovalent and
q* as a function of c/c*. Thus, everything is correlated divalent counterions. These insights mainly concern the
to a change in c*, or in the average size of the macroions, position in the reciprocal space, q*, of the broad maxi-
when monovalent counterions are replaced by divalent mum observed in the scattering profiles, first associated
ones. The origin of the shift in q* for X ) 0 is therefore with an electrostatic correlation hole within the isotropic
related to a shrinkage in the macroion conformation due model.48 We used both small-angle X-ray and neutron
to an electrostatic bridging phenomena via Ca counter- scattering techniques to investigate the structure of the
ions. It must be noticed that even if these simulations salt-free aqueous solutions of sulfonated polystyrene
capture the main characteristics of the effect of the (PSS) macroions in the presence of various mixtures of
counterion valence on the structure of the macroions, sodium (Na) and calcium (Ca) counterions. q* has been
they are restricted to very short macroions. studied as a function of X, which is related to the ratio
5.4. High Concentration Regime. At higher con- of monovalent (sodium, Na) to divalent (calcium, Ca)
centrations, neither the scaling approach nor the nu- counterions as well as of the concentration of macroions,
merical simulations are able to explain the q* deviations c, and compared with the inverse of the correlation
with respect to the c1/2 scaling law, observed for CaPSS length ξ predicted by scaling theories.48-50
and Na-CaPSS mixtures. In this concentration range, One of the most relevant parameters shows to be the
the nature of the condensed counterions is believed to degree of condensation of counterions, successfully
play a relevant role in the structure of the solutions. evaluated according to the Manning-Oosawa ap-
As discussed previously, these deviations are probably proach.40,41
related to the misunderstanding with the peak position First, we have reviewed the case of monovalent
that results from the influence of the counterion form counterions by a survey of the literature which com-
factor upon the total scattering function measured by pletes our measurements for the case X ) 1 (i.e., pure
SAXS, as demonstrated for monovalent counterions in NaPSS). We show that a universal curve exists, linking
section 4.3. However, since the relative importance of q* and 2π/ξ, not only for weakly charged polyelectrolytes
this scattering effect on the q* deviation is difficult to but also for highly charged ones, provided one intro-
determine, other origins may be introduced. duces an effective charge fraction estimated from the
The origin could also be a progressive increase in the Manning-Oosawa approach of counterion condensation.
counterion condensation related to the presence of In practice, the charge fraction f should only be replaced
divalent counterion and leading to a continuous de- by the effective charge fraction feff ) b/lB, in the scaling
crease in the effective charge fraction.20,74 For the law for the correlation length ξ as a function of f and
highest concentrations, the scattering peak also disap- concentration c:
pears (X ) 0, 0.25, and 0.5) and the scattering curves
become similar to those of neutral polymers. Both the ξ ) f -2/7(lB/b)-1/7(bc)-1/2
decrease in the scaling exponent for c > 0.1 mol/L and
the disappearance of the scattering peak at higher
(in which lB and b are the Bjerrum and monomer
concentrations are presumably correlated or could have
lengths, respectively; we are here in the strong-coupling
the same origin.
limit, since some of the counterions are bound or the
Because such deviations resemble the ones observed counterion concentration in the vicinity of the macroions
for hydrophobic polyelectrolytes,52-54,63,64,71,76 we checked is finite). Conversely, information about the condensa-
the possible existence of hydrophobic microdomains in tion of the monovalent counterions could be gained from
CaPSS and Na-CaPSS aqueous solutions using fluo- q* measurements for salt-free aqueous solutions.
rescence spectroscopy (see Appendix B). The answer is Such relation as q* ) 2π/ξ is surprisingly perfectly
negative: the observed deviations do not seem to be obeyed in a lower concentration range.
correlated to any hydrophobic effect. However, slight deviations from this equality (hence
In the case of pure CaPSS, divalent Ca counterions from the usual c1/2 scaling law), which gradually in-
induce the formation of intramolecular bridges and thus crease with c, can be observed beyond a certain c or q*
a slight collapse that could mimic the one of a polyelec- value. The contrast with the good agreement at low c
trolyte in poor solvent conditions. Also, the number of prompted us to a better understanding. It turns out that
Ca bridges could vary with the concentration (as well the relevant maximum is the one of the macroion partial
intermolecular scattering function, while the total scat- another along the chemical sequence, i.e., the formation
tering function can be affected by a slight change in the of temporary loops in the conformation. Such a conjec-
form factor.75 The total scattering function involves ture seems to be quite reasonable, since it is in ac-
intra- and intermolecular correlations as well as mac- cordance with previous form factor measurements and
roion and counterion partial scattering functions. Dis- simulations.16,17,25 One has, however, still to derive a
crepancies are expected, all the more important as the deeper theoretical description, especially allowing to
counterion part of the scattering function is large (e.g., precise the respective roles of free and condensed
metal counterions for X-rays) and as the concentration counterions. Finally, let us note that we have chosen
increases (because q* enters a higher q range with not to discuss the effect of Ca free counterions on the
higher contribution of the counterion part). This could Debye screening length because we do not expect it to
explain why the exponent of the scaling law linking q* be relevant in the investigated c range.
to c is found to vary from 0.46 to 0.5 in the literature At this stage, we can conclude that divalent counter-
(in good solvent), according to the scattering technique ions are efficient in linking parts of chains only when a
(X-ray or neutron), the nature of the counterions, the fraction of them is free (condensation has reached its
macroions, and the solvent. maximum) which may seem surprising.
Second, we used our results to show that this scaling- Fourth, when the concentration is increased, failures
condensation model applies when monovalent (Na) are present even for lower fractions of divalent (Ca)
counterions are replaced by divalent (Ca) ones, up to a counterions. We have found that the scaling theory of
certain fraction of Ca counterions. The same PSS Dobrynin, Colby, and Rubinstein, associated with the
macroion was considered; this is a model case because Manning-Oosawa approach of the counterion conden-
of the absence of complexation between macroion and sation, fails to explain our scattering results. Deviations
counterions. from the usual c1/2 scaling law are observed that
Following the Manning-Oosawa approach for the enhance as the fraction of Ca counterions increases.
counterion condensation process, when X is varied from Thus, the exponent of the scaling law linking q* to c is
1 (pure NaPSS) to 0 (pure CaPSS), three distinct X reduced down to 0.2 as X is varied to 0. Also, the peak
regimes are expected (Figure 5a). For low X values (X in the scattering function vanishes for X ) 0, 0.25, and
< 0.18), only Ca counterions are condensed and free 0.5 above a certain concentration that increases with
counterions are composed of Na and Ca counterions. The X. The role of the condensed divalent (Ca) counterions
effective charge fraction feff is then constant. In a is clearly revealed. At this stage, such modifications are
particular X range (0.18 < X < 0.36), all Ca counterions mainly due to the influence of the intramolecular
are condensed and all Na counterions are free. In this correlations as in the case of monovalent counterions.
regime, feff just scales as X. Finally, for high X values However, a slight change in the condensation threshold
(X > 0.36), condensed counterions consists of both Na is also conceivable. Bridging phenomena due to divalent
and Ca counterions, and there is a constant number of counterions may also play a role even in absence of free
Na free counterions. Again, feff is constant. divalent counterions, provided that the concentration
In the low concentration range, the measured q* is large enough. This high concentration regime also
values were found to be in qualitative agreement with needs other investigations to be fully interpreted.
such a description. In the range X g 0.25 (the theoretical In summary, at low concentration, the scaling model
limit is actually 0.18), everything seems to be governed for flexible polyelectrolytes in good solvent, combined
by the valence of the free counterions (monovalent). with the Manning-Osawa estimation of the effective
That of the condensed counterions does not appear to charge fraction, agrees with data for monovalent coun-
be essential: changes in the static structure are de- terions alone or partly replaced by divalent ones, as long
scribed by changes in the effective charge fraction alone. as those are all condensed on the macroions. When some
This is supported by the existence of a master curve divalent counterions are free, they introduce new in-
connecting q* and 2π/ξ. teractions. At higher concentrations, these different
Third, we have characterized deviations from the effects may be even more intricate.
scaling model for larger fractions of divalent (Ca)
counterions. For X < 0.25 (the theoretical limit is still Acknowledgment. We thank Bruno Demé, from
0.18), i.e., when Na free counterions are also replaced Institut Laue Langevin, and Fabien Schnell, from
by Ca counterions or as part of Ca counterions are free, Institut Charles Sadron, for their help during neutron
a more significant change in the structure is observed and X-ray small-angle scattering measurements. We are
that cannot be accounted for by a simple charge renor- indebted to the Institut Laue Langevin for use of
malization. Specifically, for X ) 0 the variations of q* neutron beam time. Thanks are also due to Albert
with c and f can no longer be described by the previous Johner for valuable discussions.
scaling law, even if one includes the counterion conden-
sation according to the Manning-Oosawa approach. Appendix A. Condensation Process for Highly
Obviously, the counterion condensation is not the only Charged Macroions in the Presence of
process. Monovalent Counterions
We then have turned our interest toward the influ- The universal curve, linking q* and 2π/ξ, in Figure
ence of divalent counterions on the macroion conforma- 6b does not correspond to a strict equality q* ) 2π/ξ,
tion, as already investigated by use of the zero average and deviations toward lower q* values are observed
contrast (ZAC) method in SANS from aqueous solutions beyond a certain c or q* value. Actually, they are related
of CaPSS, as well as NaPSS plus added Ca salt.16,17 The to deviations from the usual c1/2 scaling law, which
main effect is a retraction of the macroions that, in gradually increase with c. As remarked in section 4.3,
particular, modifies the scaling account for c*. This such deviations could be due to a progressive increase
retraction can be due to the existence of pure electro- in the counterion condensation process, i.e., a continuous
static Ca bridges between monomers apart from one decrease in the counterion condensation threshold. The
Figure 9. Effective charge fraction feff vs monomer concentra- Figure 10. Ratio I1/I3 of the intensities of the first and third
tion c for NaPSS. Here feff is an adjustable parameter derived peaks of the fluorescence spectra of pyrene for pure NaPSS
from the scaling law (5′), by assuming that the q* deviations (X ) 1) and CaPSS (X ) 0) aqueous solutions at room
from the strict equality q* ) 2π/ξ in Figure 6b are exclusively temperature.
due to an increase in the counterion condensation with c.
trolytes are in qualitative agreement with such a
effective charge fraction would then be expected to vary behavior.52-54,64,71,76 The c dependence of q* observed
with c. In this appendix, the effective charge fraction is with fully sulfonated PSS macroions in the presence of
considered as an adjustable parameter, and its c de- mixed valence or divalent counterions has also a certain
pendence is derived by using the scaling law (5′) as similarity and therefore could be related to a poor
reference. We can emphasized that a comparison be- solvent effect or an hydrophobic character of the mac-
tween q* and 2π/ξ is not straightforward. A prefactor roions. Partially sulfonated polystyrenes are hydropho-
is indeed missing. Nevertheless, as suggested by the bic polyelectrolytes of which the structure of their
light scattering part of the master curve reported in salt-free aqueous solutions has been extensively stud-
Figure 6b, this prefactor should be close to unity. The ied.52,53,71,76 Such a hydrophobic character can be evi-
results obtained from measurements on pure NaPSS denced by studying: either the change in the scattered
samples are presented in Figure 9. intensity by the aqueous solutions as the temperature
Whereas the effective charge fraction is close to 0.36 is varied53,54,80 or the fluorescence of pyrene molecules
in the low concentration regime, i.e., the expected value dissolved in the aqueous solutions.53,71 This last method
according to the Manning-Oosawa theory, it is continu- specifically allows to probe the existence of hydrophobic
ously reduced when the concentration is increased. This microdomains.
decrease is related to the fact that q* does not rigorously We also performed SAXS measurements on pure
follow the c1/2 power law. NaPSS and CaPSS at different temperatures. No major
The effective charge fraction of fully sulfonated poly- modification of the scattered intensity is observed as the
styrene has been measured by Essafi53 and Baigl71,81 temperature is varied from 20 to 80 °C. Such a behavior
using osmotic pressure and potentiometric techniques. is quite different from the one of the aqueous solutions
The experimental value was found to be very close to of partially sulfonated polystyrenes. We also made
the expected theoretical one (0.36) in the concentrations measurements on pure NaPSS in mixtures of water and
range 0.01-0.1 mol/L in contradiction with our extrapo- THF (THF is a good solvent for polystyrene). A very
lations from the q* measurements. If we cannot rule slight displacement of q* is observed that is only
out a slight decrease of the effective charge fraction with explained by a change in the dielectric constant. The
the concentration, the departure to the c1/2 power law Na-CaPSS mixtures can also be considered as hydro-
in the q* evolution or, in other words, the deviations philic, as dissolved in water.
from the strict equality q* ) 2π/ξ has to be connected This is still proved by measurements of the fluores-
to another phenomenon. cence of pyrene molecules. Indeed, if the pyrene mol-
ecules were trapped in hydrophobic microdomains, the
Appendix B. Absence of Hydrophobic ratio I1/I3 should be reduced. As shown in Figure 10,
Microdomains in CaPSS Solutions this is not the case since I1/I3 remains similar as Na
An increasing attention has been devoted to the counterions are replaced by Ca counterions and close
properties of polyelectrolytes in poor solvent.74,77-78 The to the value associated with pyrene molecules in pure
macroions are expected to adopt a pearl-necklace con- water.
formation with dense charged aggregates (pearls) con-
nected by stretched chain parts (narrow elongated References and Notes
strings).77-79 In the semidilute regime, according to the
concentration, two distinct scaling laws are then pre- (1) Hara, M., Ed.; Polyelectrolytes: Science and Technology;
Marcel Dekker: New York, 1993.
dicted for the c dependence of q*. In the string controlled (2) Schmitz, K. S. Macroions in Solution and Colloidal Suspen-
regime (c such as the distance between two successive sions; VCH: New York, 1993.
beads along the chemical sequence is larger than ξ), q* (3) Förster, S.; Schmidt, M. Adv. Polym. Sci. 1995, 120, 51.
scales as c1/2; in the bead-controlled regime (c such as (4) Barrat, J.-L.; Joanny, J.-F. Adv. Chem. Phys. 1996, 94, 1.
this distance is less than ξ), it is a c1/3 scaling law that (5) Radeva, T., Ed.; Physical Chemistry of Polyelectrolytes; Marcel
prevails.77 Some experiments on hydrophobic polyelec- Dekker: New York, 2001.
(6) Holm, C., Kekicheff, P., Podgornik, R., Eds.; Electrostatic (44) Dolar, D. In Polyelectrolytes; Sélégny, E., Ed.; Reidel Publish-
Effects in Soft Matter and Biophysics; NATO Sci. Ser. II; ing: Dordrecht, 1974; pp 97-113.
Kluwer Academic Press: Dordrecht, 2001; Vol. 46. (45) Rinaudo, M.; Loiseleur, B. Bull. Soc. Chim. 1973, 124, 1.
(7) Widom, J.; Baldwin, R. L. J. Mol. Biol. 1980, 144, 431; (46) Rinaudo, M. In Polyelectrolytes; Sélégny, E., Ed.; Reidel
Biopolymers 1983, 22, 1595. Publishing: Dordrecht, 1974; pp 157-193.
(8) Bloomfield, V. A. Biopolymers 1991, 31, 1471; Curr. Opin. (47) Tomsic, M.; Bester Rogac, M.; Jamnik, A. Acta Chim. Slov.
Struct. Biol. 1996, 6, 334. 2001, 40, 333.
(9) Huber, K. J. Phys. Chem. 1993, 97, 9825. Ikeda, Y.; Beer, (48) de Gennes, P.-G.; Pincus, P.; Velasco, R. M.; Brochard, F. J.
M.; Schmidt, M.; Huber, K. Macromolecules 1998, 31, 728. Phys. (Paris) 1976, 37, 1461.
(10) Axelos, M. A.; Mestdagh, M. M.; François, J. Macromolecules (49) Pfeuty, P. J. Phys., Colloq. 1978, C2, 149.
1994, 27, 6594. Heitz, C. Thesis, Université Louis Pasteur (50) Dobrynin, A. V.; Colby, R. H.; Rubinstein, M. Macromolecules
de Strasbourg, 1996. François, J.; Heitz, C.; Mestdagh, M. 1995, 28, 1859.
M. Polymer 1997, 38, 5321. Heitz, C.; François, J. Polymer (51) Makowski, H. S.; Lundberg, R. D.; Singhal, G. S US Patent
1999, 40, 3331. 3870841, 1975, to Exxon Research and Engineering Co.
(11) Olvera de la Cruz, M.; Belloni, L.; Delsanti, M.; Dalbiez, J.- (52) Essafi, W.; Lafuma, F.; Williams, C. E. In Macroion Char-
P.; Spalla, O.; Drifford, M. J. Chem. Phys. 1995, 103, 5781. acterization from Dilute Solutions to Complex Fluids; ACS
(12) Yoshikawa, K.; Takahashi, M.; Vasilevskaya, V. V.; Khokhlov, Symposium Series 548; Schmitz, K. S., Ed.; American Chemi-
A. R. Phys. Rev. Lett. 1996, 76, 3029. cal Society: Washington, DC, 1994; Chapter 21, pp 278-286.
(13) Tang, J. X.; Janmey, P. A. J. Biol. Chem. 1996, 271, 8556. (53) Essafi, W. Thesis, Université Paris VI, 1996.
Tang, J. X.; Wong, S.; Tran, P.; Janmey, P. A. Ber. Bunsen- (54) Heinrich, M. Thesis, Université Louis Pasteur de Strasbourg,
Ges. Phys. Chem. 1996, 100, 1. Tang, J. X.; Ito, T.; Tao, T.; 1998.
Traub, P.; Janmey, P. A. Biochemistry 1997, 36, 12600. (55) Baigl, D.; Seery, T. A. P.; Williams, C. E. Macromolecules
(14) Sabbagh, I. Thesis, Université Paris VII, 1997. Sabbagh, I.; 2003, 36, 6878.
Delsanti, M.; Lesieur, P. Eur. Phys. J. B 1999, 12, 253. (56) Tondre, C.; Zana, R. J. Phys. Chem. 1972, 76, 3451.
Sabbagh, I.; Delsanti, M Eur. Phys. J. E 2000, 1, 75. Drifford, (57) Millero, F. J. In Water and Aqueous Solutions; Horne, R. A.,
M.; Delsanti, M. In Physical Chemistry of Polyelectrolytes; Ed.; Wiley-Interscience: New York, 1972; Chapter 13.
Radeva, T., Ed.; Marcel Dekker: New York, 2001; Chapter (58) Glatter, O., Kratky, O., Eds.; Small-Angle X-ray Scattering;
4, pp 135-161. Academic Press: London, 1982. Cotton, J.-P. In Neutron,
(15) Raspaud, E.; Olvera de la Cruz, M.; Sikorav, J.-L.; Livolant, X-ray and Light Scattering: Introduction to an Investigate
F. Biophys. J. 1998, 74, 381. Tool for Colloidal and Polymeric Systems; Lindner, P., Zemb,
(16) Zhang, Y.; Douglas, J. F.; Ermi, B. D.; Amis, E. J. J. Chem. Th., Eds.; North-Holland Delta Series: Amsterdam, 1991;
Phys. 2001, 114, 3299. Chapter 2, pp 19-31. Lindner, P. In Neutrons, X-rays and
(17) Dubois, E.; Boué, F. Macromolecules 2001, 34, 3684. Light: Scattering Methods Applied to Soft Condensed Matter;
(18) Wong, G. C. L.; Lin, A.; Tang, J. X.; Li, Y.; Janmey, P. A.; North-Holland Delta Series: Amsterdam, 2002; Chapter 2,
Safinya, C. R. Phys. Rev. Lett. 2003, 91, 18103. Butler, J. C.; pp 23-48.
Angelini, T.; Tang, J. X.; Wong, G. C. L. Phys. Rev. Lett. 2003, (59) Sears, G. D. Neutron News 1992, 3, 26.
91, 28301. Angelini, T. E.; Liang, H.; Wriggers, W.; Wong, (60) El Brahmi, K. Thesis, Université Louis Pasteur de Stras-
G. C. L. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 8634. bourg, 1991.
(19) Sorci, G. A.; Reed, W. F. Macromolecules 2004, 37, 554. (61) Nishida, K.; Kaji, K.; Kanaya, T. Macromolecules 1995, 28,
(20) Stevens, M. J.; Kremer, K. Phys. Rev. Lett. 1993, 71, 2228; 2472.
J. Chem. Phys. 1995, 103, 1669. (62) Ermi, B. D.; Amis, E. J. Macromolecules 1997, 30, 6937.
(21) Stevens, M. J. Phys. Rev. Lett. 1999, 82, 101. (63) Essafi, W.; Lafuma, F.; Williams, C. E. Eur. Phys. J. B 1999,
(22) Gronbech-Jensen, N.; Mashl, R. J.; Bruinsma, R. F.; Gelbart, 9, 261.
W. M. Phys. Rev. Lett. 1997, 78, 2477. (64) Waigh, T. A.; Ober, R.; Williams, C.; Galin, J.-C. Macromol-
(23) Chu, J. C.; Mak, C. H. J. Chem. Phys. 1999, 110, 2669. ecules 2001, 34, 1973.
(24) Liu, S.; Muthukumar, M. J. Chem. Phys. 2002, 116, 9975. (65) Drifford, M.; Dalbiez, J.-P. J. Phys. Chem. 1984, 88, 5368.
(25) Chang, R.; Yethira, A. J. Chem. Phys. 2003, 118, 11315. (66) Cotton, J.-P.; Moan, M. J. Phys., Lett. 1976, 37, L-75.
(26) Wittmer, J.; Johner, A.; Joanny, J.-F. J. Phys. II 1995, 5, 635. (67) Nierlich, M.; Williams, C. E.; Boué, F.; Cotton, J.-P.; Daoud,
(27) Gonzales-Mozuelos, P.; Olvera de la Cruz, M. J. Chem. Phys. M.; Farnoux, B.; Jannink, G.; Picot, C.; Moan, M.; Wolf, C.;
1995, 103, 3145. Rinaudo, M.; de Gennes, P.-G. J. Phys. (Paris) 1979, 40, 701.
(28) Rouzina, I.; Bloomfield, V. A. J. Phys. Chem. 1996, 100, 9977. (68) Ise, N.; Obuko, T.; Kunugi, S.; Matsuoka, H.; Yamamoto, K.
(29) Ha, B.-Y.; Liu, A. J. Phys. Rev. Lett. 1997, 79, 1289; Phys. I.; Ishii, Y. J. Chem. Phys. 1984, 81, 3294.
Rev. Lett. 1998, 81, 1011; Phys. Rev. E 1998, 58, 6281; Phys. (69) Kaji, K.; Urakawa, H.; Kanaya, T.; Kitamaru, R. J. Phys.
Rev. E 1999, 60, 803. (Paris) 1988, 49, 993.
(30) Brilliantov, N. V.; Kuznetsov, D. V.; Klein, R. Phys. Rev. Lett. (70) Manning, G. S. In Polyelectrolytes; Sélégny, E., Ed.; D.
1998, 81, 1433. Reidel: Dordrecht, 1974; pp 9-37.
(31) Schiessel, H.; Pincus, P. Macromolecules 1998, 31, 7953. (71) Baigl, D. Thesis, Université Paris VI, 2003.
(32) Golestanian, R.; Kardar, M.; Liverpool, T. B. Phys. Rev. Lett. (72) Reddy, M.; Marinsky, J. A.; Sarkar, A. J. Phys. Chem. 1970,
1999, 82, 4456. 74, 3891.
(33) Kuhn, P. S.; Levin, Y.; Barbosa, M. C. Physica A 1999, 266, (73) Hayter, J.-B.; Jannink, G.; Brochard-Wyart, F.; de Gennes,
413. P.-G. J. Phys., Lett. 1980, 41, L-451.
(34) Solis, F. J.; Olvera de la Cruz, M. Eur. Phys. J. E 2001, 4, (74) Micka, U.; Holm, C.; Kremer, K. Langmuir 1999, 15, 4033.
143. (75) Nishida, K.; Kaji, K.; Kanaya, T.; Shibano, T. Macromolecules
(35) Solis, F. J. J. Chem. Phys. 2002, 117, 9009. 2002, 35, 4084.
(36) Borukhov, I.; Lee, K.-C.; Gelbart, W. M.; Liu, A. J.; Stevens, (76) Baigl, D.; Ober, R.; Qu, D.; Fery, A.; Williams, C. E. Europhys.
M. J. J. Chem. Phys. 2002, 117, 462. Lett. 2003, 62, 588.
(37) Ermoshkin, A. V.; Olvera de la Cruz, M. Phys. Rev. Lett. 2003, (77) Dobrynin, A. V.; Rubinstein, M.; Obukhov, S. P. Macromol-
90, 125504. ecules 1996, 29, 2974. Dobrynin, A. V.; Rubinstein, M.
(38) Ermoshkin, A. V.; Olvera de la Cruz, M. J. Polym. Sci., Part Macromolecules 1999, 32, 915.
B: Polym. Phys. 2004, 42, 766. (78) Limbach, H. J.; Holm, C.; Kremer, K. Europhys. Lett. 2002,
(39) Naji, A.; Netz, R. R. Eur. Phys. J. E 2004, 13, 43. 60, 566.
(40) Manning, G. S. J. Chem. Phys. 1969, 51, 924, 934; Annu. Rev. (79) Kantor, Y.; Kardar, M. Europhys. Lett. 1994, 27, 643. Kantor,
Phys. Chem. 1972, 23, 117; Ber. Bunsen-Ges. Phys. Chem. Y.; Kardar, M. Phys. Rev. E 1995, 51, 1299.
1996, 100, 909. (80) Boué, F.; Cotton, J.-P.; Lapp, A.; Jannink, G. J. Chem. Phys.
(41) Oosawa, F. Biopolymers 1968, 6, 134; Polyelectrolytes; Marcel 1994, 101, 2562.
Dekker: New York, 1971. (81) Qu, D.; Baigl, D.; Williams, C. E.; Möhwald, E.; Fery, A.
(42) Holm, C.; Kremer, K. Proceedings of Yamada Conference L Macromolecules 2003, 36, 6878.
Polyelectrolytes; Yamada Science Foundation: Osaka, 1999; (82) Shew, C. Y.; Yethiraj, A. J. Chem. Phys. 1999, 110, 11599.
p 27.
(43) Dolar, D.; Peterlin, A. J. Chem. Phys. 1969, 50, 3011. MA0479722

Combet 2005

Uploaded by

Copyright:

Available Formats

Combet 2005

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Combet 2005

Uploaded by

Copyright:

Available Formats

7456 Macromolecules 2005, 38, 7456-7469

Scattering Functions of Flexible Polyelectrolytes in the Presence of

Jérôme Combet,* François Isel, and Michel Rawiso

1. Introduction to distinct aggregated structures according to their

Figure 3. SAXS profiles of salt-free aqueous solutions of pure

Figure 1. Scattering from salt-free aqueous solutions of mixed

experimental and theoretical points of view.40,41,53,63 It

tial well (which let them some mobility along the

associated with an electrostatic correlation hole, each

Here, f is the charge fraction of the polyelectrolytes.

ξ ) feff-2/7(lB/b)-1/7(bc)-1/2 (5′) In this way, we take into account the counterion

evaluate ξ (Table 2). Furthermore, as will be seen in

exponents associated with f, lB, and b, lies upon the

5. Na-CaPSS Mixtures: Interpretive Discussion

You might also like