Article pubs.acs.org/Langmuir
Distribution of Counterions around Lignosulfonate Macromolecules in Different Polar Solvent Mixtures Ulla Vainio,*,† Rolf A. Lauten,‡ Sylvio Haas,§ Kirsi Svedström,⊥ Larissa S. I. Veiga,∥ Armin Hoell,# and Ritva Serimaa⊥ †
HASYLAB at DESY, Notkestr. 85, D-22607 Hamburg, Germany Borregaard Lignotech, P. O. Box 162, N-1701 Sarpsborg, Norway § Humboldt-Universität zu Berlin, Institut für Chemie, Brook-Taylor-Strasse 2, D-12489 Berlin, Germany ⊥ Department of Physics, P. O. Box 64, FI-00014 University of Helsinki, Finland ∥ Instituto de Fisica ″Gleb Wataghin″, UNICAMP, CP 6165, 13083-970 Campinas, Brazil # Helmholtz-Zentrum Berlin für Materialien und Energie, Institute of Applied Materials, Hahn-Meitner-Platz 1, D-14109 Berlin, Germany ‡
S Supporting Information *
ABSTRACT: Lignosulfonate is a colloidal polyelectrolyte that is obtained as a side product in sulfite pulping. In this work we wanted to study the noncovalent association of the colloids in different solvents, as well as to find out how the charged sulfonate groups are organized on the colloid surface. We studied sodium and rubidium lignosulfonate in water−methanol mixtures and in dimethyl formamide. The number average molecular weights of the Na- and Rb-lignosulfonate fractions were 7600 g/mol and 9100 g/mol, respectively, and the polydispersity index for both was 2. Anomalous small-angle X-ray scattering (ASAXS) was used for determining the distribution of counterions around the Rb-lignosulfonate macromolecules. The scattering curves were fitted with a model constructed from ellipsoids of revolution of different sizes. Counterions were taken into account by deriving an approximative formula for the scattering intensity of the Poisson−Boltzmann diffuse double layer model. The interaction term between the spheroidal particles was estimated using the local monodisperse approximation and the improved Hayter−Penfold structure factor given by the rescaled mean spherical approximation. Effective charge of the polyelectrolyte and the local dielectric constant of the solvent close to the globular polyelectrolyte were followed as a function of the methanol content in the solvent and lignosulfonate concentration. The lignosulfonate macromolecules were found to aggregate noncovalently in water−methanol mixtures with increasing methanol or lignosulfonate content in a specific directional manner. The flat macromolecule aggregates had a nearly constant thickness of 1−1.4 nm, while their diameter grew when counterion association onto the polyelectrolyte increased. These results indicate that the charged groups in lignosulfonate are mostly at the flat surfaces of the colloid, allowing the associated lignosulfonate complexes to grow further at the edges of the complex.
■
INTRODUCTION Lignosulfonate (LS) is a colloidal polyelectrolyte, which is produced in the sulfite pulping process of wood. During pulping, lignin is separated from cellulose and made soluble in water by sulfonate groups which get grafted onto the amorphous aromatic polymer. Lignosulfonate is used widely as a dispersant in many industrial processes,1 and recently, sodium lignosulfonate has been found to be a superb dispersant for carbon nanotubes.2 Unlike many dispersants which form well-defined association colloids, lignosulfonate is a branched polyelectrolyte, a colloid by itself. Lignosulfonate is known to be surface active, particularly at high concentrations.3 Long chain alcohols have been shown to improve slightly the surface activity of calcium lignosulfonate,3 and it has been suggested that lignosulfonate forms mixed surfactant systems with anionic surfactants.4 Both the ability to form mixed surfactant systems and association with long chain alcohols, in addition to the © 2011 American Chemical Society
possible self-association, are related to the shape and distribution of charged as well as uncharged groups on the macromolecular surface. The shape of lignosulfonate macromolecules has been studied with various methods. Thicknesses from about 1.3 to 3.5 nm were observed for different lignins on liquid surfaces in the beginning of the 1970s.5 Remarkably, the thickness did not depend on the molecular mass of lignin, indicating that the molecules adopt a flat conformation on the interface. Later, the result was corroborated by electron microscopy studies, which showed a 2 nm thickness for lignosulfonate macromolecules on surfaces.6 In solution, the shape may be different, but in the past years, solute low molar mass macromolecules of lignin Received: November 14, 2011 Revised: December 20, 2011 Published: December 22, 2011 2465
dx.doi.org/10.1021/la204479d | Langmuir 2012, 28, 2465−2475
Langmuir
Article
derivatives have mostly been described as being flat.7,8 However, the shape of the lignin particle has been in most cases determined using solutions where electrostatic interactions between macromolecules have not been present. By using pure water and methanol−water mixtures as solvents for lignosulfonate and by using several different concentrations of lignosulfonate, we attempt here to find structural information on both the shape and charge of the macroion under different conditions. Instead of screening the electrostatic interactions, these interactions are included in the structural model that is fitted to the experimental data. Anomalous small-angle X-ray scattering (ASAXS) has been recently applied to study the distribution of counterions around spherical and rod-like colloids,9−11 including DNA,12 while a related method called resonant X-ray reflectivity has been used to study counterion distribution close to flat lipid surfaces.13 Applying the ASAXS method to the study of lignosulfonate is not straightforward due to the polydisperse and random nature of this natural product. High statistical accuracy is needed for the ASAXS data, and therefore we decided to apply it to concentrated solutions in order to get enough scattering signal in a reasonable amount of time. The theoretical handling of concentrated solutions of nonspherical charged or even noncharged particles relies very much on approximations, and a robust analytical way to approach the problem is still missing. In this study we develop and test a model which can be applied to study the counterion condensation in ellipsoidal macroion systems.
■
knowledge of the molecular weight of the molecules as well as the atomic weight of Rb, we can calculate that Rb and Na lignosulfonate molecules with the number average molecular weight given in Table 1 contain about 19 charged groups, which corresponds to an effective surface charge density of about 1.16 e−/nm2 (16 μC/cm2), when assuming a spherical shape with volume equivalent radius 1.1 nm for the particles and no counterion association. By measuring the density of solutions of sodium and rubidium lignosulfonate dissolved in water and using the knowledge of the concentrations of the solution, the densities of sodium and rubidium lignosulfonate were calculated to be 2.02 g/cm3 and 2.44 g/cm3 (error 2%), respectively, the ratio of which is the same as that of the molecular weights. It is noteworthy that the measured density for sodium lignosulfonate fraction is larger than the density 1.4 g/cm3 that we used for the same fraction in our previous publication,8 where we simply assumed the density to be the same as the literature value15 for lignin. The main results of that publication, such as the particle shape determination, are not affected, but unfortunately the volume fraction might be systematically wrong by a scaling factor. Small-angle X-ray scattering (SAXS) was measured from all concentration series of sodium and rubidium lignosulfonate. Selected rubidium lignosulfonate samples were also measured at several photon energies using anomalous small-angle X-ray scattering (ASAXS). In this method the energy of the X-rays is tuned below the absorption edge of an element in the sample in order to achieve resonant scattering effects in the SAXS pattern. The SAXS and ASAXS experiments were carried out at two ASAXS dedicated experimental stations 7T-MPW SAXS at BESSY synchrotron in Berlin and B1 at DORIS III in Hamburg.16,17 The sample solutions were prepared a few weeks before the measurement and let to incubate in room temperature. At 7T-MPW SAXS, the size of the beam on the sample was 0.5 mm × 0.5 mm. The sample solutions were injected into glass mark tubes (Hilgenberg) that had about 50μm-thick walls and inner diameter of 3.9 mm. The thickness of each tube was measured separately, because the standard deviation of the thickness of the tubes was 0.6 mm. The mark tubes were sealed with Teflon, or for long measurements also glued tight, and placed vertically to the sample holder. A two-dimensional multiwire proportional counter was placed at two different distances (3749 mm and 1404 mm) behind the sample to detect the SAXS patterns and to cover a suitable scattering vector range. One SAXS pattern was typically measured in 8 min. ASAXS measurements from samples were made by measuring SAXS at 14803 (E1), 15085 (E2), and at 15166 eV (E3) near the Rb K-edge, which is at 15200 eV. A typical ASAXS measurement of one sample took a couple of hours in total. All measurements were made at (22 ± 1) °C. The obtained SAXS data was corrected for dead time, normalized by detector flatfield, incoming photon flux, as well as transmission and thickness of the sample, integrated over the azimuth angle, calibrated to absolute intensity units (1/cm) using a glassy carbon reference sample, and corrected for solvent background while taking into account the volume fraction of the polyelectrolyte.8 Half a year later, ASAXS measurements of one Rb lignosulfonate sample (ϕ = 0.17) in water were repeated at beamline B1 using similar experimental parameters as at 7T-MPW SAXS. Furthermore, SAXS measurements of samples with DMF as the solvent were measured at an energy of 11021 eV at B1, and these samples were placed in quartz capillaries of 2 mm in diameter.
EXPERIMENTAL SECTION
Two lignosulfonate salts prepared from one sample of carefully fractionated lignosulfonate were used in this work. The preparation of fractionated sodium lignosulfonate is described elsewhere.8 To obtain rubidium lignosulfonate, an ion-exchange resin (Amberlite IR120) was used to convert the sodium salt to lignosulfonate acid. The lignosulfonate acid was neutralized with an aqueous solution of RbOH (Aldrich). The metal content, i.e., rubidium and sodium, was determined using inductively coupled plasma spectrometry (ICP) to verify that treatment with the ion-exchange resin removed the metal counterions and that rubidium replaced sodium as counterion during neutralization. Molecular weight was determined with a gel permeation chromatography (GPC) technique described elsewhere.14 The lignosulfonate salts with, respectively, sodium or rubidium as counterions differ in molecular weight and the difference between the two fractions corresponds to the difference in the counterion atomic weight. Table 1 summarizes the characteristics of the two lignosulfonate salts. The polydispersity of the fraction was 2.07.
Table 1. Lignosulfonate Salt Molecular Weight, Density ρm, and Puritya
a
type
Mn (g/mol)
Mw (g/mol)
ρm (g/cm3)
Na
Ca
Rb
NaLS RbLS
7600 9100
18 000 21 000
2.02 2.44
4.7% 0.09%
a(E) to have the same shape as the potential (while neglecting a constant background). The surface charge density of our polyelectrolyte may be high enough to violate the 2467
dx.doi.org/10.1021/la204479d | Langmuir 2012, 28, 2465−2475
Langmuir
Article
assumption |eψ| < kT, depending on degree of counterion association. However, we are interested only in an approximate shape of the ion cloud around the polyelectrolyte and actually also assume that the associated counterions are found throughout the polymer instead of forming a compact counterion layer on the surface. If the Debye−Hückel approximation is valid, an approximate thickness for the electrical double layer is then given by the Debye screening length 1/κ, where we now define the Debye−Hückel parameter κ as23
κ=
ze 2 ϕ Vpεε0kBT
The well-known form factor amplitude of a sphere is
Fsphere(q , E) = ρr < a(E)V × ⎡ sin(qr ) − qr cos(qr ) ⎤ ⎢3 ⎥ ⎢⎣ ⎥⎦ (qr )3
where r is the radius of the sphere (here r = a), and V is the volume of the sphere (4πa3/3). Integration of eq 4 by employing the second row in eq 5 gives us the form factor amplitude for the diffuse double layer
Fions(q , E) = ρr > a(E)V × ⎡ κr sin(qr ) + qr cos(qr ) ⎤ ⎢3 ⎥ ⎢⎣ ⎥⎦ q κ 2r 3 + (qr )3
(6)
Here z is the electric charge number and Vp is the volume of the polyelectrolyte particle, ϕ is the volume fraction of polyelectrolyte in solution, ε is the dielectric constant of the solvent, and ε0 is the vacuum permittivity. Any contribution from background electrolyte is not included in this formula. Regarding the scattering contrast that depends on energy ρr>a(E) in eq 5, for the special case of Rb lignosulfonate this scattering length density of ions compared to solvent, changes as a function of energy E near the absorption edge of Rb due to the change in the anomalous scattering factors of the ions. If some ions are located inside the particles the scattering length density ρr
