ARTICLE pubs.acs.org/Langmuir
Dispersions of Nanocrystalline Cellulose in Aqueous Polymer Solutions: Structure Formation of Colloidal Rods Yaman Boluk,*,†,‡ Liyan Zhao,§ and Vanessa Incani†,‡ †
Department of Civil and Environmental Engineering, University of Alberta, Edmonton, Alberta T6G 2G2, Canada National Institute for Nanotechnology, National Research Council of Canada, Edmonton, Alberta T6G 2M9, Canada § Alberta InnovatesTechnology Futures, Edmonton, Alberta T6N 1E4, Canada ‡
ABSTRACT: The steady-state shear and linear viscoelastic deformations of semidilute suspensions of rod-shaped nanocrystalline cellulose (NCC) particles in 1.0% hydroxyethyl cellulose and carboxymethyl cellulose solutions were investigated. Addition of NCC at the onset of semidilute suspension concentration significantly altered the rheological and linear viscoelastic properties of semidilute polymer solutions. The low-shear viscosity values of polymers solutions were increased 20490 times (depending on polymer molecular weight and functional groups) by the presence of NCC. NCC suspensions in polymer solutions exhibited yield stresses up to 7.12 Pa. Viscoelasticity measurements also showed that NCC suspended polymer solutions had higher linear elastic moduli than the loss moduli. All of those results revealed the gel formation of NCC particles and presence of internal structures. The formation of a weak gel structure was due to the nonadsorbing macromolecules which caused the depletion-induced interaction among NCC particles. A simple interaction energy model was used to show successfully the flocculation of NCC particles in the presence of nonadsorbing polymers. The model is based on the incorporation of the depletion interaction term between two parallel plates into the DLVO theory for cubic prismatic rod shaped NCC particles.
’ INTRODUCTION Colloidal suspensions are employed in a wide range of applications such as paints, coatings, cosmetics, ceramics, pharmaceutical formulations, food, and household products. There are two resourceful ways of formulating suspensions with effective structure formation and gelling characters. One way is the utilization of anisotropic particles such as rods and disks in suspensions. The other one is the incorporation of dissolved polymers in suspensions. Each of them is also the focus of extensive theoretical and experimental investigations by the discipline of colloid science. The aim of this paper is to investigate the structure formation and gelling properties of suspensions of rod shaped nanocrystalline cellulose (NCC) whiskers at low concentrations in polymers in solutions. Rod shaped particles may stay randomly oriented (isotropic) state in most of the circumstances. However in other situations, they may align depending on concentration, strength of interparticle interactions and flow field. Hence suspensions of nanometer size rod-shaped particles may exhibit interesting phenomena even at low fractions such as elasticity, thixotropy, birefringence, nematic phase, and phase separation. Langmuir, in one of his pivotal publications, reported the role of attractive and repulsive forces in the formation of tactoids and thixotropic gels of rod shaped fd virus and vanadium pent oxide particles.1 Onsager2 developed the theory of suspensions of rods with repulsive interactions and showed isotropicnematic phase transition. Experiments on rod-like boehmite,3,4 fd virus,5 NCC,6,7 and r 2012 American Chemical Society
chitin crystallite8 colloidal particles in aqueous systems are used to demonstrate the isotropicnematic phase transitions. Microstructural regimes of rod-like colloidal suspensions including nanosized ones are reviewed by Solomon and Spicer. 9 Nevertheless, experimental studies on the thixotropy and eleasticity of nanosized rod-shaped colloids at very low concentration range are very limited. 1012 Another common practice of alteration of the stability and rheological properties of colloidal suspensions is the addition of polymers in suspensions. Depending on the nature of particle polymer interactions, polymers either generate stability or flocculate colloidal systems. Colloidpolymer interactions have been studied both experimentally and theoretically during the last three decades.1316 Studies showed that polymers can cause two types of steric interactions among particles: either bridging by adsorption of polymers on particle surfaces or depletion by nonadsorbing polymers. Both of them can lead flocculation which form gels and carry stresses.17,18 Flocculation due to polymer adsorption occurs at low surface coverage, by bridging particles. Gels caused by bridging are transient in nature and revert back to the fluid state over a time scale of minutes to days. In the case of nonadsorbing polymers, the phase behavior and microstructure of depletion flocculates are influenced by the Received: September 9, 2011 Revised: November 25, 2011 Published: March 26, 2012 6114
dx.doi.org/10.1021/la2035449 | Langmuir 2012, 28, 6114–6123
Langmuir
Figure 1. Scanning electron micrograph of NCC.
range of interaction potentials. In literature, most of the viscoelastic and rheological properties of colloids in polymer solutions have been studied using spherical particles.1923 Meanwhile the interest on suspensions of rod shaped NCC particles has been growing.24,25 The hydrolysis of cellulose by concentrated sulfuric acid in a controlled mode removes the amorphous segments of fibrils and releases homogeneous and defect free NCC particles. They are typically 610 nm in width and 80200 nm in length. In addition, esterification reaction creates negatively charged sulfate ester groups on crystallite surfaces. Therefore, sulfuric acid processed NCC surfaces yield stable suspensions in water.24,26 Rod-shaped NCC particles generate birefringence and ordered liquid phases at sufficiently high concentrations in aqueous systems.6 After suspension reaches a certain concentration, the nanocrystals form a chiral nematic ordered phase displaying optical characteristics of a typical cholesteric liquid crystal. Particle geometry, addition of electrolytes, and nonadsorbing macromolecules also govern the phase separation.7 The transition from isotropic phase to ordered phases are investigated by using small-angle neutron scattering (SANS), small-angle X-ray scattering (SAXS).27 Electroviscous effects of NCC suspensions are investigated by studying the rheology of dilute NCC suspensions.28 As it is reviewed briefly here, the addition of nonadsorbing polymers is investigated mostly in suspensions of spherical particles. In the case of rod-like NCC particles, polymers are used only to induce the isotropic to nematic transitions and the phase separation of concentrated suspensions in the literature. The effects of polymers on the rheology and structure of NCC suspensions in dilute to semidilute concentration ranges have never been investigated. Our goal is to investigate and describe the structure of NCC suspensions in the presence of polymer solutions at low colloidal particle volume concentrations. We report the responses of NCC suspensions in 1.0% hydroxyethyl cellulose (HEC) solutions and carboxymethyl cellulose (CMC) solutions subject to steady state shear and linear dynamic viscoelastic measurements.
’ EXPERIMENTAL SECTION NCC. NCC sample was prepared by acid hydrolysis of commercial dissolved softwood pulp. Pulp was hydrolyzed with 64% sulfuric acid at 50 °C for 40 min and then diluted with deionized water to stop the reaction. The suspension was then centrifuged, neutralized with Na2CO3 and dialyzed to remove the salts. Finally, the suspension was further dispersed in an ultrasonic bath to achieve 12% concentration stable colloidal suspension. NCC was obtained in powder form by freeze-drying the suspensions. Detailed preparation method is discussed elsewhere.24,29 NCC particles in aqueous solutions carry negative
ARTICLE
electrical charges due to the sulfate surface groups. 0.66% (vol.) NCC suspensions were prepared for the experiments. Weight to volume conversion of NCC suspension concentration was carried out by considering the density of NCC as 1.5 g/cm3. Scanning Electron Microscopy (SEM). Scanning electron microscope was used to image the NCC particles. The instrument was a Hitachi model S-4800 equipped with a field emission source and operating at an accelerating voltage of 30 kV (Hitachi Tokyo, Japan). The individual NCC particles were prepared by drying a drop of a dilute suspension (0.1% by weight) on a carbon-coated microscope grid. A drop of uranyl acetate 0.2 wt % aqueous solution was then deposited on top of the previous grid, followed by a drying at room temperature. Figure 1 shows the micrograph of rod-like NCC particles. The average particle width (radius) determined by image analysis was 8 ( 1 nm. Cryo-Transmission Electron Microscopy (Cryo-TEM). Cryotransmission electron microscopy (Cryo-TEM) was used to image 0.66% NCC in 1.0% HEC720 solution. All images were obtained at 200 kV on a JEOL2200FS TEM equipped with a Schottky field emission gun and an in-column omega type energy filter. Samples for Cryo-TEM imaging were prepared by applying 0.66% NCC suspension in 1.0% HEC720 solution onto a TEM sample grid with holey carbon thin film supported. The excess solution in the droplet was blotted with filter paper. The grid with a thin film of solution was plunged into liquid nitrogen temperature. The grid was then transferred to cryo holder on cryo workstation and inserting the cryo holder to TEM column for imaging. Dynamic Light Scattering (DLS). The effective particle size of NCC particles based on translational diffusion constant was determined by dynamic light scattering (DLS). The instrument was a Zetasizer model Nano ZS from Malvern (Worcestershire United Kingdom). The wavelength of the 4 mW HeNe laser source was 633 nm. The particle size was measured at room temperature, and the NCC concentration was 0.1%. The effective particle size of NCC particles was measured as 90 ( 10 nm and taken as the length of particles. Zeta Potential Measurements. The zeta potential of NCC particles was determined by Zetasizer model Nano ZS from Malvern (Worcestershire United Kingdom). Zetasizer model Nano ZS is also a zeta potential analyzer based on electrophoretic light scatting. A 4 mW HeNe laser source with 633 nm wavelength was used as light source. All measurements were performed at 20 °C. The zeta potential of NCC particles determined by Zetasizer was 51.5 ( 0.8 mV. Water-Soluble Polymers. 2-Hydroxyethyl cellulose (HEC) and sodium salt of carboxymethyl cellulose (CMC) were selected as watersoluble polymers for our study and obtained from Sigma Aldrich. According to the supplier’s product specifications, HEC samples have weight average molecular weights (Mw) of 720 000, 250 000, and 90 000 Da. The molecular weights and radii of gyration of HEC polymers were determined by measuring intrinsic viscosities ([η]) of aqueous HEC solutions by using an Ubbelohde viscometer. The viscosity average molecular weights (Mv) and radii of gyration (RG) of nonionic HEC polymers were calculated by using [η] = 4.1 102 Mv0.73 and RG = 2.6 102 Mv0.59 relationships.30 The intrinsic viscosity and molecular weight of CMC were calculated from the Newtonian viscosity of 1.0% CMC700 solution by using the η0cMv and η0c[η] relationships.30 The Einstein equivalent hard sphere radius (RG) of CMC was determined from the intrinsic viscosity [η] and molecular weight, using the relationship [η] = 2.5πNA(2RG)3/600Mv where NA is the Avogadro constant. Values of the critical overlap concentrations (c*) were approximately determined according to Morris et al.31 by using c* = 4/[η]. Description, supplier listed molecular weight, degree of substitution (DS), molar substitution (MS), calculated intrinsic viscosity, viscosity average molecular weight, radius of gyration, and critical overlap concentration of HEC and CMC polymers are listed in Table1. 6115
dx.doi.org/10.1021/la2035449 |Langmuir 2012, 28, 6114–6123
Langmuir
ARTICLE
Table 1. Polymers Used for Solution Preparations in Experiments polymer
a
acronym
D.S.
M.S.
[η] (mL/g)
Mv (Da)
RG (nm)
c* (g/mL)
hydroxyethyl cellulose
HEC90
1.0
2.5
135
66 000
18
0.03
hydroxyethyl cellulose
HEC250
1.0
2.5
296
194 000
34
0.013
hydroxyethyl cellulose
HEC720
1.0
2.0
850
819 000
80
0.005
carboxymethyl cellulose
CMC700
0.7
2270a
650 000a
280
0.002
Calculated from the Newtonian viscosity of 480 Pa s. the frequency between 0.1 and 100 rad/s, at the 1.0% strain which was within the linear range. All of the measurements were done at 20 °C. Isothermal Titration Calorimetry (ITC). Isothermal titration calorimetry (ITC) was performed on a Nano-ITC from TA Instruments (New Castle, DE). NCC and HEC250 were dialyzed (MWCO 3500) against water or 10 mM of NaCl over a period of 2 d. The NCC suspension was later filtered (0.45 μm filter) in order to eliminate dust particles. NCC, HEC, 10 mM of NaCl, and water were properly degassed prior to use. For all experiments, the sample cell was filled with 950 μL of water, NaCl or NCC (1 wt %), and HEC (5 wt %) in a 250 μL syringe. Twenty five portions of 10 μL were injected at 1000 s intervals. After each addition, the instrument monitored the heat released or absorbed as a result of the different processes occurring in the solution. The heat of dilution of HEC was used as a blank and subtracted from the heat released or absorbed by the addition of HEC into NCC suspension. The molar heat of injection (ΔH in KJ/mol) was determined by integrating the injection peaks over each individual injection. All measurements were performed at 25 °C and the syringe was stirred at a constant rate of 250 rpm.
’ RESULTS
Figure 2. Steady shear viscosity of 0.66% NCC suspension and 1.0% solutions of polymers with and without 0.66% NCC: (a) as a function of shear-rate and (b) as a function of shear stress.
NCC Suspensions. A stock suspension with 5% by weight NCC was prepared from powder sample in deionized water. It was used to prepare 0.66% volume fraction of NCC suspension for further measurements. The pH of the suspension was 6.8. No further pH adjustment was done. NCC suspension was first sonicated for 3 min and then HEC and CMC polymers were incorporated. Both HEC and CMC polymers were in powder form. They were slowly added, mixed by stirrer for 3 h and completely dissolved in each of sample at 30 °C to prepare NCC suspensions in polymer solutions. Rheological Measurements. Steady state shear viscosity vs shear-rate measurements of NCC suspensions were carried out on a TA Instruments AR-G2 Rheometer (New Castle, DE) equipped with a 2° cone and plate geometry of 60 mm in diameter. The torque resolution is 0.1 μN. Dynamic Viscoelastic Measurements. TA Instruments AR-G2 rheometer equipped with a 2° cone and plate geometry of 60 mm in diameter was also used to measure linear viscoelastic properties of NCC suspensions in 1.0% HEC720 and CMC700 solutions. First, strain displacement sweep at 1 rad/s was carried out to determine the linear strain range. Then, all of the oscillation mode tests were done sweeping
Shear Viscosity. Steady state shear viscosities of 0.66 vol % NCC suspension and 1.0 wt % HEC and CMC polymer solutions with and without 0.66 vol % NCC presence are shown in Figure 2 a,b as a function of shear-rate and shear stress respectively. The transition from dilute to semi dilute concentration occurs at volume fraction of ϕ e (π/4)(d/L)2 where d is the diameter and L is the length of NCC rods. This equation suggests that NCC suspension at 0.66% volume concentration is just above the dilute concentration range where particles barely contact each other. The viscosity of 0.66% NCC suspension in water did not thicken and its viscosity can be predicted by Simha’s equation.28 As expected, viscosities of 1.0% HEC90 and HEC250 solutions were also Newtonian because of their low molecular weights and their 1.0% solution concentration was lower than their respective overlapping concentrations. 1.0% solutions of high molecular weight HEC720 and CMC700 polymers showed higher and shear-rate dependent thickening behavior. NonNewtonian character of 1.0% HEC 720 and CMC700 solutions were similar. HEC720 has higher molecular weight compared to CMC700. Nevertheless, much longer extension of anionic CMC chains, because of intramolecular repulsions and resulting RG compensated its somewhat lower molecular weight and produced very similar shear-rate dependent viscosity profile. On the other hand, viscosities of NCC suspensions in polymer solutions showed unusually big jumps in low-shear-rates and highly shearthinning behaviors. Viscosities of polymer solutions with 0.66% NCC at the low shear-rate (102 s1) surged between 20 and 490 times. Those viscosity raises depended on the molecular weight (66 000819 000 Da) and radius of gyration (18280 nm) of polymers tested in experiments. This unusually exceptional 6116
dx.doi.org/10.1021/la2035449 |Langmuir 2012, 28, 6114–6123
Langmuir
ARTICLE
Table 2. Yield Stress, Viscosity Index, and Shear-Rate Index of NCC Suspensions and Polymer Solutions viscosity yield shear
Figure 3. Steady state shear stress vs shear-rate plots of NCC suspensions in polymer solutions.
behavior is not observed in polymer solutions with conventional fillers. As shown in Figure 2b when viscosity is plotted against shear stress, the viscosity profiles of these NCC suspensions in polymer solutions displayed more than one inflection points: again this behavior is not observed in conventional polymer solutions. Viscosities of NCC suspension in HEC250, HEC720, and CMC700 increased with shear rate at very low shear rate region and then decreased with increasing shear rate above 0.1 s1. The first segment where the viscosity increases at very low shear rate range is expected to be due to shear induced NCC structure. The decrease in the viscosity with increasing shear-rate above 0.1 s1 is due to the disruption of the structure of NCC in polymer solutions. The rate of disruptions of connections exceeds the rate at which new connections can be reformed. Yield Stress. To make a comparative evaluation, the strength of NCC gels in polymer solutions were quantified by calculating yield stress values. Assessing the yield stress values has been challenging and there are more than few methods for measurements. In our case, the plots of shear stress vs shear-rate were fitted to HerschelBulkley model to calculate the yield stress. The HerschelBulkley is a generalized model of a shear thinning non-Newtonian fluid which also exhibits yield stress :n ð1Þ τ ¼ τY þ kγ where the viscosity index (k), the shear-rate index (n), and the yield shear stress (τY) characterize shear stress, τshear-rate γ_ relationship. Figure 3 shows the shear stress vs shear strain rate plots. As expected, 0.66% NCC suspension in water did not show any yield stress. Likewise, polymer solutions without NCC also did not show yield stress. Only NCC suspensions in HEC and CMC solutions had varying degree of yield stress values. The yield stress is a rheological property it is commonly observed for highly concentrated suspensions. Unlike them, rod shaped nanosized NCC particles in polymer solutions generated yield points just at the onset of semidilute concentration. The yield stress of NCC suspension in low molecular weight HEC250 solution was not detectable. On the other side, NCC suspensions in high molecular weight HEC720 and CMC700 solutions had significant yield values. In those suspensions, before the yield point, at very low strain values, the stress vs shear-rate increase was linear. HEC720 suspension had a higher initial slope than the CMC700 suspension. The initial slope corresponds to the elastic deformation of the NCC suspensions in polymer solutions. Then there was a
index, shear-rate standard
stress,τY (Pa) k (Pa s) index, n
sample
error
0.66% NCC
0.0
0.003
1.0
7.3
1.0% HEC90
0.0
0.004
1.0
n.a.
1.0% HEC90 + 0.66% NCC
0.0
0.022
1.0
1.9
1.0% HEC250
0.0
0.022
0.947
1.7
1.0% HEC250 + 0.66% NCC
0.35
0.31
0.651
6.7
1.0% HEC720
0.0
1.33
0.529
13.7
1.0% HEC720 + 0.66% NCC
4.47
4.74
0.383
25.48
1.0% CMC700 1.0% CMC700 + 0.66% NCC
0.002 7.12
0.99 2.42
0.639 0.53
12.09 13.86
Figure 4. G0 (ω) and G00 (ω) 1.0 wt % HEC250 solutions with and without NCC presence.
peak that defines the yield point. It was followed by a very brief decrease and then steady increase of the shear stress with shearrate. The yield stress corresponds to breakdown of the weak structure under the shear flow. The yield stress of CMC700 suspension was higher than HEC720 suspension. Herschel Bulkley model parameters of polymer solutions with and without NCC suspensions were calculated by using all of the measured shear stress-shear-rate data points (Table 2). As discussed in the preceding paragraph, NCC presence exhibited yield stress values of 0.35, 4.47, and 7.12 Pas in 1.0% HEC250, HEC720, and CMC700 solutions respectively. Yield stress value was affected by the chemical structure and molecular weight of polymers. The yield stress is caused by the network of NCC particles that can be broken by the shearing motion. Viscosity index in Table 2 can be considered as the viscosity value at 1 s1 shear-rate. NCC in HEC720 solution had lower yield stress but higher viscosity index than in CMC700 solution. However the stress generated at 1 s1 shear-rate was 9.21 Pa (4.47 + 4.74) for HEC720 suspension and 9.64 Pa (7.12 + 2.42) for CMC700 suspension, which are very similar. The shear-rate index shows the shear thinning character of the suspension in polymer solution. The presence of NCC particles in polymer solutions increased the shear thinning character thus decreased the shear-rate index. This drastic behavior is also clearly seen in Figure 2. 6117
dx.doi.org/10.1021/la2035449 |Langmuir 2012, 28, 6114–6123
Langmuir
Figure 5. G0 (ω) and G00 (ω) 1.0 wt % HEC720 solutions with and without NCC presence.
ARTICLE
Figure 7. Molar heat of injection for the titration of 5 wt % HEC250 into 0.66 vol % NCC (open squares) and heat of dilution of HEC250 5 wt % in 10 mM of NaCl (closed diamonds). The insert presents the characteristic ITC peaks of injecting HEC into NCC in 10 mM NaCl (raw ITC data correspond to the data display with open squares).
Figure 8. Molar heat of injection for the titration of HEC250 5 wt % into NCC 1 wt % (closed squares) and heat of dilution of HEC250 5 wt % in water (open diamonds). The insert presents the characteristic ITC peaks of injecting HEC into NCC in water (raw ITC data correspond to the data display with open diamonds). Figure 6. G0 (ω) and G00 (ω) 1.0 wt % CMC700 solutions with and without NCC presence.
Linear Viscoelasticity. To understand the influence of nonionic and anionic polymers on the viscoelastic properties of NCC dispersions within linear deformation range, oscillatory measurements were carried out in 1.0% HEC250, HEC720, and CMC700 solutions. Linear viscoelasticity is the simplest types of viscoelastic behavior with very small deformations (strains). Therefore it was used to observe the disturbance of NCC particles from their equilibrium configuration. The results are presented in Figures 46 respectively. As expected in each case, the polymer solutions without the presence of NCC gave the loss modulus (G00 ) much higher than the elastic modulus (G0 ). At all measured frequencies the viscous component was dominant, nevertheless as the frequency was increased, the elastic component increased more steeper than viscous component. 1.0% HEC and CMC polymer solutions had no G0 and G00 crossover points, terminal relaxation times and plateau regions within the measured frequency regions. This behavior is due to their fluid like nature with low viscoelasticity. Among them, HEC250 with the lowest molecular weight was the more fluid like and had almost none elastic behavior. HEC720 and CMC700 solutions were similar and more viscoelastic. There are some noticeable features of NCC filled polymer solutions as seen in Figures 46. Elastic modulus (G0 ) of each polymer solution increased drastically and reached to loss modulus, (G00 ) which is the sign of gelation. Nevertheless NCC presence had
the more drastic effect in HEC720 solution than others. G0 values of NCC suspension in HEC720 had a plateau region and it exceeded the G00 values at the measure frequency range. G0 and G00 values tend to converge at higher frequencies. The presence of the plateau region in the G0 curve indicated that some internal structures might exist, such as aggregated fractal-like structures, showing the gelation character of the NCC particles. The crossover point was observed for NCC gel in HEC250 solution. On the other hand, G0 and G00 were almost overlapped in the case of CMC700 solution. Since G0 ≈ ωa and G00 ≈ ωb and according to gel theory, the gel point is reached when a = b = 0.5 and the final gel state is reached when a = 0 and b = 1. In the case of NCC suspension in CMC700, a = 0.51 and b = 0.56. Similarly, NCC suspension in HEC720 (for the plateau range) has a = 0.21 and b = 0.76. As observed both in the yield stress and viscoelasticity measurements, NCC in HEC700 solution resulted more structured and high modulus of elasticity network. Nevertheless as shown in yield stress measurements, NCC in CMC700 withstands higher stresses before permanently destroying the elastic structure. PolymerHEC Interactions. The interaction potential of nonionic HEC with NCC was studied by isothermal titration calorimetry. In the first set of experiments the heat of dilution of HEC solutions were determined. The polymer was injected into water or NaCl to measure the exchange of heat associated with the interactions between HEC and the medium (Figures 7 and 8). We observed that at low concentration, hydration of the polymer leads to an exothermic signal that gradually decreases as the polymer concentration increases until it reaches a plateau. 6118
dx.doi.org/10.1021/la2035449 |Langmuir 2012, 28, 6114–6123
Langmuir
Figure 9. (a) SEM micrograph of 0.66% NCC in water; (b) Cryo-TEM micrograph of 0.66% NCC in 1.0% HEC solution.
This phenomena has been interpreted as two different processes (1) where there is a long-ranged restructuring of water molecules by the polymer32,33 (case for low polymer concentration) and (2) where the polymer and water interaction is short-ranged and the water shells of the polymers do not overlap (case for high polymer concentration). In a second set of experiments, HEC250 was added to NCC suspension. As shown in Figures 7 and 8 the value of the signals obtained is rather small in both cases. Titration in water yield a constant ΔH of 0.27 kJ/mol whereas in NaCl yield a constant ΔH of +0.39 kJ/mol. Such as small heat differences between the blank and HEC polymerNCC suspension suggest that the HECNCC interaction is negligible. Micrographs. Figure 9a shows a representative SEM micrograph of 0.66% NCC suspension in water without any polymer presence. NCC particles are well dispersed throughout water and no aggregates are visible. Representative Cryo-TEM micrograph of 0.66% NCC suspension in 1.0% HEC720 solution is shown in Figure 9b. This sample was chosen because it has the highest thickening characteristic. As shown in cryo-TEM micrograph, NCC particles are weakly flocculated and form a structure. 0.66% NCC suspension in 1.0% HEC solution is totally different than flocculate free NCC suspensions in water.
’ DISCUSSION Flocculation of NCC by Polymers. The rheological and viscoelastic characterization of suspensions of semidilute NCC in HEC and CMC polymer solutions demonstrated that addition of rod-shaped nanoparticles, even at the onset of semidilute concentration, significantly altered the linear viscoelastic and nonlinear rheology of semidilute polymer solutions. Two alternative mechanisms can be considered to explain the dramatic enhancements of the rheological behavior of NCC suspensions in polymer solutions: According to the first mechanism, polymer-bridging of particles can cause weak NCC networks. Suspensions of negatively charged NCC particles are electrostatically stable at low ionic strength. However in polymer solutions, if NCC surfaces prefer macromolecules to aqueous medium, polymers can adsorb on NCC surfaces. If polymer chains are long enough, some segments adsorb in the form of trains while tails protrude into the solution and are available to adsorb on other particles. By applying polymer self-consistent field theory and a “saturable” adsorption model to capture the effect of the particle size, Surve et al.34 demonstrated that polymers are adsorbed on nanoparticles significantly higher than larger particles. It was also suggested that long tails of macromolecules extends to a distance about radius of gyration RG of polymers. According to their prediction, interparticle interactions by polymer bridging tend to weaken with an increase in the concentration of polymers.
ARTICLE
However, isothermal titration calorimetry results excluded the possibility of adsorption of HEC on NCC. Hydroxyethyl cellulose (HEC) macromolecules have no preference to replace surface bound water molecules by adsorption on NCC surface. HEC is a nonionic polymer and like NCC surface has the ability both to accept and donate protons. Therefore its adsorption onto NCC, even it exists, is very weak by hydrogen bonding. Naturally, sodium carboxymethyl cellulose (CMC) is also expected not to adsorb on NCC. CMC is an anionic polymer and its adsorption on anionic sulfate groups contained NCC shall not be electrostatically favorable. Even in the case of adsorbing polymers, 1.0% polymer concentration employed in 0.66% NCC suspensions is extremely high to observe flocculation by bridging. Since polymer bridging occurs at low surface coverage (polymer additions of 0.0001%) bridging by adsorption mechanism is unlikely for NCC in HEC and CMC solutions. Therefore flocculation by bridging is not considered to explain rheological behavior of NCC suspensions in HEC and CMC solutions. The second potential mechanism is the flocculation of NCC particles by the depletion of nonadsorbing polymers. The rheological behavior of NCC suspensions in polymer solutions can be explained by depletion flocculation. If macromolecules do not adsorb at particle interfaces, the exclusion of the polymer segments from the particulate volumes leads to an effective “entropic” attraction between the particles known as the depletion interaction.17,35 Asakura and Oosawa were the first to formulate the qualitative analysis of depletion. They proposed that the exclusion of polymer molecules leads to an effective attractive interaction between the particles.36,37 This AO model and majority of experimental works deal with the regime where the size of colloidal particles (diameter, r) is much bigger than the size of polymers (radius of gyration, RG). This regime of r . RG is commonly called “colloid limit”. The opposite limit, where r e RG has been mostly investigated for the common practice of adding polymer to protein solutions in order to aid protein crystallization and therefore referred as “protein limit”.38 r e RG is the case not only for proteins but for all nanoparticles when particles are smaller than the radius of gyration of nonadsorbing polymers. We believe “nanoparticle limit” term, instead of the “protein limit” term can better describe such situation. NCC rods have a profile of infinite plane in one direction and a profile of sphere in the other direction. Therefore NCC rods in parallel orientation with L(length) > RG still satisfies the colloid limit while in perpendicular orientation with w(width) , RG falls within nanoparticles (protein) limit. As it was discussed in the case of adsorbing polymers, nonionic HEC has almost no interaction potential and anionic CMC has repulsive interactions with NCC. In addition, our polymer concentrations in NCC suspensions can only favor the depletion interactions. Therefore the gel-like structure of NCC dispersion in polymer solutions is due to the attractive interactions of particles in the presence of depleted nonadsorbing polymers. Similarly flocculation of micrometer sized spherical silica particles in the presence of HEC and CMC polymers at high concentrations due to the depletion interactions were also reported in the literature.22 Nevertheless in our case rod-like NCC particles with L = 90 nm and d = 8 nm and polymers with radius of gyrations between 18 and 280 nm posed special cases which we may call depletion interaction between rod shaped particles in nanoparticle limit. Experimental data will be further 6119
dx.doi.org/10.1021/la2035449 |Langmuir 2012, 28, 6114–6123
Langmuir
ARTICLE
compared with depletion flocculation model in the following sections. Interaction Energy Calculations. Rheological and viscoelastic measurements showed that polymers caused interactions among NCC particles. Before discussing the role of nonadsorbing polymers, first the total interaction energies between NCC pairs in water (with no polymer presence) were calculated in terms of the repulsive electrostatic (Vr), and attractive the van der Waals, (Va) energies, according to DLVO theory. As a second step, the depletion term (Vdep) was incorporated to include the effect of polymer molecules on the interactions between NCC particles. Sparnaay’s expressions derived for the interaction between rods were used, as given by Buining et al.3 The electrostatic repulsive interactions for parallel (VrR) and crossed (VrC) cylindrical rods are given by the following expressions: VrP ¼ 64ðπÞ0:5 nkTγ2 L VrC ¼ 128πnkTγ2 with
γ ¼ tanh
eψ0 4kT
ðkaÞ0:5 expð kHÞ k2
a expð kHÞ k2
ð2Þ ð3Þ
ð4Þ
where n is the number density of ions, k is the Boltzmann constant, T is the absolute temperature, k is Debye length for 1:1 electrode, ψ0 is the surface potential, e is the charge of the electron and a is the radius of rod, and L is the length of rod. The van der Waals attractive interactions between parallel (VPa ) and crossed (VCa ) flat square prism rods are given as 3 L VaP ¼ πA a5 U5 8 a " # A ð1 0:5m2 C EðkÞ KðkÞ Va ¼ 3 1 m2
ð5Þ
ð6Þ
where A is the Hamaker constant, U5 is function of cylinder radius, a and the distance between the cylinder axes (R = H + 2a). Method of calculation is given by Buining et al.3 In the expression for crossed cylindrical rods m is a/R and E and K are the first and second elliptical integrals respectively. Alternatively, the repulsive (Vr) and attractive (Va) interactions between parallel and crossed NCC particles were calculated by assuming cubic prismatic geometry and using simple parallel flat plates approximations: VrP, C ¼ 64nkTγ2 S VaP, C ¼
1 expð kHÞ k2
A S 12πH 2
ð7Þ
Figure 10. Interaction energy V = Va + Vr between cylindrical (—) and flat (---) rods. Calculations performed for particle diameter (width for flat rods) = 8 nm; length = 90 nm; Debye length = 3.0 nm; zeta potential = 51.5 mV; Hamaker constant = 1.20 1020 J.
1.1 1020 J. and 4.38 1020 J. respectively. Surface potential value of 51.5 mV was used based on zeta potential measurement. Debye length of 3.0 nm which corresponds to monovalent salt concentration of 20 mM was also used in all of the calculations. The total interaction potentials VP,C = (Va + Vr)P,C for two parallel and crossed oriented NCC particles, assuming cylindrical rod geometry with diameter of 8 nm and length of 90 nm were calculated by using eqs 26. Similarly, the total interaction potentials of NCC particles were calculated based on cubic prismatic rod geometry by using eqs 7 and 8. Figure 10 shows a large energy barrier (86.2 kT) for parallel and smaller maximum (12.6 kT) for crossed oriented cylindrical rods. Alternatively, cubic prismatic rods by using a simple parallel flat plate approximation resulted maximum barrier of 86.9 kT and 7.7 kT for parallel and crossed orientations, respectively. Since the barrier height is very high, repulsion between parallel rods are dominated. Repulsion between crossed oriented particles was much lower than parallel orientation. The repulsive maxima occur around 1 nm distance between two particles surfaces. According to isothermal titration calorimetry results, nonionic HEC did not get adsorbed on NCC surfaces. Naturally, the adsorption of anionic CMC on anionic NCC was also not expected. As we discussed in the preceding section, rheological and viscoelastic properties of NCC suspensions in those polymers suggested strong interactions among NCC particles. Therefore, when NCC is suspended in HEC and CMC solutions, the total free energy of interactions for parallel (P) and crossed (C) oriented NCC particles have to include the depletion interaction term among particles due to nonadsorbing polymers (Vdep)
ð8Þ
where S is the surface area and equals wL and w2 for parallel and cross oriented cubic prismatic rods respectively with width of w and length of L. Hamaker’s constant A = 1.2 1020 J was calculated by the √ √ equation A = ( A11 A22)2 where A11 is Hamaker’s constant for cellulose particles and A22 is Hamaker’s constant for water. Hamaker’s constants for cellulose and water are assumed equal to
V P, C ¼ ðVa þ Vr þ Vdep ÞP, C
ð9Þ
Since the analytical solution for the depletion interaction between two parallel plates in the presence of ideal polymers is straightforward,39 we incorporated it into the DLVO theory of cubic prismatic rods which also used Va and Vb terms for parallel flat plates. More accurate computer simulation technique is used by Dogic et al.5 for fd virus which is another rod shaped nanoparticle to plot isotropicnematic phase diagrams following 6120
dx.doi.org/10.1021/la2035449 |Langmuir 2012, 28, 6114–6123
Langmuir
ARTICLE
Figure 12. Semidilute NCC suspensions: (a) without the presence of nonadsorbing polymer; (b) percolated NCC network due to the presence of depleted nonadsorbing polymer.
Figure 11. The interaction energy V = Va + Vr + Vdep between flat rods. Calculations performed for particle diameter (width) = 8 nm; length = 90 nm; Debye length = 3.0 nm; zeta potential = 51.5 mV; Hamaker constant = 1.20 1020 J, HEC720 polymer with RG = 80 nm, Mv = 819 000, c* = 0.002 g/mL, Π = 317 N/m2.
Table 3. Secondary Minimum of Interaction Energies Due to Polymer Nonadsorption osmotic
secondary
secondary
pressure,
minimum,
minimum, distance at secondary
polymer
Π (N/m2)
V P (kT)
V C (kT)
minimum(nm)
HEC90 HEC250
707 442
1.70 2.23
0.15 0,20
20 21
HEC720
317
CMC700
1660
2.70 26.8
interparticle
0.24
24
2.38
20
the method described by Tuinier et al.39 4RG P, C Vdep ¼ ΠS pffiffiffi H HχðHÞ π
ð10Þ
where Π is the osmotic pressure of the dissolved polymer and RG is the radius of gyration of polymer chains. Π is defined as c ð11Þ Π ¼ np kT 1 þ 1:12 c The osmotic pressure is calculated from the number density of polymers in the solution, np and by taking into account the effect of mutual interaction between the polymer molecules. The term 1.12(c/c*) is the reduced concentration showing the degree of coil overlapping by taking into account mutual interaction between polymer molecules.40 In eq 10, χ(H) is defined as ! 8 1 p2 π 2 RG 2 χðHÞ ¼ 2 ΣP ¼ 1, 3, 5 2 exp ð12Þ π n h2 Equations 1012 were used to calculate the depletion interaction values for parallel and crossed oriented rods and then incorporated into eq 9 to obtain the total interaction energy. Figure 11 shows the calculated interaction energies for NCC particles, modeled as cubic prisms in 1.0% HEC720 solution. The depletion interaction term is calculated by taking RG = 80 nm,
Mv = 819 00, c* = 0.005 g/mL, and Π = 317 N/m2 for HEC720 solution. Due to the nonadsorbing polymer, a second minimum is formed at large (24 nm) interparticle distance. The position of the second minimum of the intermolecular potential does not change when the angle between two rods changes from parallel to crossed orientation; only the magnitude of the minimum changes. The secondary minimum (2.70 kT) of parallel oriented rods is sufficiently deep which makes weak flocculation of NCC particles possible. The secondary minimum in crossed orientation is shallower (0.2 kT) due to lower value of faced surface area (w2). Therefore, a parallel flocculation seems favorable to crossed orientation in the secondary minimum. Such weakly flocculated aggregates formed structured network and exhibit unique rheological and viscoelastic behavior. Table 3 lists osmotic pressures of HEC and CMC polymer solutions and secondary minimum energy values of NCC particles in those polymer solutions. Calculated osmotic pressure values of HEC solutions are in very good agreement with an experimental measurement reported in the literature.22 According to these measurements, osmotic pressure of 1.0% HEC (Mw = 270 000 Da) solution is 387.2 N/m2 and 0.4% CMC (Mw = 243 000 Da) solution is 5019 N/m2. Our calculated osmotic pressure 442 N/m2 of 1.0% HEC250 (Mw = 194 000 Da) solution seems reasonable. However our calculated osmotic pressure 1660 N/m2 of CMC700 is believed to predict less than the real value. Anionic CMC700 solution caused very deep secondary minimum at parallel (26.8 kT) and crossed (2.38 kT) orientations. They are the result of the greater radius of gyration due to the intramolecular charge repulsion effects. Such deep secondary energy mimimum explains the rheological and viscoelastic properties of NCC suspension in 1.0% anionic CMC700 solutions. NCC suspension in CMC700 solution shows strong interactions for both at parallel and crossed orientations with very deep secondary minimum that resulted in a high yield stress value. NCC suspension in HEC720 solution has lower secondary minimum compared to CMC700 and has lower yield stress value. Other HEC polymer solutions also followed the trend. Overall, the radius of gyration and osmotic pressure govern the depth of the secondary minimum. Since the secondary minimum at crossed orientation is negligible, the high linear elasticity of NCC in HEC720 solution can be explained by preferentially parallel orientation of interacted particles It should also be noted that our calculations of depletion interaction between cubic prisms which corresponds crossed (perpendicular) rods do not meet the required assumption of w (or 2r) > RG, colloid limit. The opposite, RG > w is the case for crossed rods and we define as nanoparticle limit (instead of protein limit). It causes depletion interaction lower than colloid limit. According to Dogic et al.5 eq 10 which is valid for the colloid model overestimates than computer simulation of 6121
dx.doi.org/10.1021/la2035449 |Langmuir 2012, 28, 6114–6123
Langmuir cylinders with a diameter of 6.6 nm while RG of the polymer is 11.1 nm. Despite of this deficiency, our simple model qualitatively showed the role of nonadsorbing polymers and also ranked HEC and CMC polymers. The presence of nonadsorbing HEC and CMC polymer molecules in NCC suspensions induce depletion attraction among rod-shaped NCC particles. As depicted in Figure 12 such weak interactions cause percolated network of NCC viscoelastic. The gel is described by Philipse and Wierenga41 as packing of “blobs” of the heterogeneous fiber fractals. The magnitude of yield stress increases with the depth of secondary minimum. The elasticity of weak gels before the yield point is the manifestation of NCC particles alignment. In addition, the weakly flocculated NCC structure results in a decrease in the accessible volume for polymer chains in the solution. Hence the apparent polymer concentration becomes higher in flocculated NCC free regions. Polymer coils which build high concentrated regions also contribute to the high yield stress of the NCC suspension.
’ CONCLUSION The addition of nonadsorbing HEC and CMC macromolecules caused depletion-induced interaction among NCC particles. A semiquantitative model for the flocculation of NCC particles due to nonadsorbing polymers was used. It is based on the incorporation of the depletion interaction between two parallel plates into the DLVO theory for cubic prismatic rod shaped NCC particles. The flocculation of NCC particles due to the presence of nonadsorbing polymers was caused by the secondary minimum. The rheological and viscoelastic characterization of suspensions of semidilute NCC in HEC and CMC polymer solutions demonstrated that addition of rod-shaped nanoparticles even in at the onset of semidilute concentration caused weak gels and significantly altered the linear and nonlinear rheology of semidilute polymer solutions. ’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected]. Tel: 780 492 6295.
’ ACKNOWLEDGMENT The work was financially supported by Alberta InnovatesBioSolutions (which includes former Alberta Forestry Research Institute). SEM and Cryo-TEM measurements were performed at National Institute for Nanotechnology of National Research Council of Canada. We thank Dr. Christophe Danumah for SEM micrograph, DLS and zeta potential measurements, and Ms. Hui Qian for Cryo-TEM sample preparation and measurements. We also thank Professor Hasan Uludag for providing Isothermal Titration Calorimetry (ITC) Equipment in his laboratory. ’ REFERENCES (1) Langmuir, I. The role of attractive and repulsive forces in the formation of tactoids, thixotropic gels, protein crystals and coacervates. J. Chem. Phys . 1938, 6, 873–896. (2) Onsager, L. The effects of shape on the interaction of colloidal particles. Ann. N.Y. Acad. Sci. 1949, 51 (4), 627–659. (3) Buining, P.; Philipse, A.; Lekkerkerker, H. Phase behavior of aqueous dispersions of colloidal boehmite rods. Langmuir 1994, 10 (7), 2106–2114.
ARTICLE
(4) Buitenhuis, J.; Donselaar, L. N.; BUiNING, P. A.; Stroobants, A.; Lekkerkerker, H. N. W. Phase separation of mixtures of colloidal boehmite rods and flexible polymer. J. Colloid Interface Sci. 1995, 175 (1), 46–56. (5) Dogic, Z.; Purdy, K. R.; Grelet, E.; Adams, M.; Fraden, S. Isotropic-nematic phase transition in suspensions of filamentous virus and the neutral polymer Dextran. Phys. Rev. E 2004, 69 (5), 51702. (6) Dong, X. M.; Kimura, T.; Revol, J. F.; Gray, D. G. Effects of ionic strength on the isotropic-chiral nematic phase transition of suspensions of cellulose crystallites. Langmuir 1996, 12 (8), 2076–2082. (7) Edgar, C. D.; Gray, D. G. Influence of dextran on the phase behavior of suspensions of cellulose nanocrystals. Macromolecules 2002, 35 (19), 7400–7406. (8) Li, J.; Revol, J. F.; Marchessault, R. H. Rheological properties of aqueous suspensions of chitin crystallites. J. Colloid Interface Sci. 1996, 183 (2), 365–373. (9) Solomon, M. J.; Spicer, P. T. Microstructural regimes of colloidal rod suspensions, gels, and glasses. Soft Matter 2010, 6 (7), 1391–1400. (10) Wierenga, A. M.; Philipse, A. P. Low-Shear Viscosities of (Semi-)Dilute, Aqueous Dispersions of Charged Boehmite Rods: Dynamic Scaling of Double Layer Effects. Langmuir 1997, 13 (17), 4574–4582. (11) Wierenga, A.; Philipse, A. P.; Lekkerkerker, H. N. W.; Boger, D. V. Aqueous Dispersions of Colloidal Boehmite: Structure, Dynamics, and Yield Stress of Rod Gels. Langmuir 1998, 14 (1), 55–65. (12) Huang, F.; Rotstein, R.; Fraden, S.; Kasza, K. E.; Flynn, N. T. Phase behavior and rheology of attractive rod-like particles. Soft Matter 2009, 5 (14), 2766–2771. (13) Tadros, T. F. Industrial Applications of Dispersions. Adv. Colloid Interface Sci. 1993, 46, 1–47. (14) Farinato, R. S.; Dubin, P. L. Colloid-Polymer Interactions From Fundamentals to Practice; John Wiley: New York, 1999. (15) Napper, D. H. Polymeric stabilization of colloidal dispersions; Academic Press: New York, 1983. (16) Fleer, G. J. Polymers at interfaces and in colloidal dispersions. Adv. Colloid Interface Sci. 2010, 159 (2), 99–116. (17) Jenkins, P.; Snowden, M. Depletion flocculation in colloidal dispersions. Adv. Colloid Interface Sci. 1996, 68, 57–96. (18) Fleer, G. J.; Scheutjens, J.; Stuart, M. A. C. Theroretical progress in polymer adsorption, steric stabilization and flocculation. Colloids Surf. 1988, 31, 1–29. (19) Schweizer, K. S.; Chen, Y. L.; Fuchs, M.; Shah, S. A.; Ramakrishnan, S.; Zukoski, C. F. Collective structure, gelation and elasticity in polymer-colloid suspensions. Abstr. Pap. Am. Chem. Soc. 2003, 226, U292–U292. (20) Shah, S.; Chen, Y. L.; Schweizer, K.; Zukoski, C. Viscoelasticity and rheology of depletion flocculated gels and fluids. J. Chem. Phys. 2003, 119, 8747. (21) Liang, W.; Tadros, T. F.; Luckham, P. Flocculation of sterically stabilized polystyrene latex particles by adsorbing and nonadsorbing poly (acrylic acid). Langmuir 1994, 10 (2), 441–446. (22) Snowden, M. J.; Clegg, S. M.; Williams, P. A.; Robb, I. D. Flocculation of silica particles by adsorbing and non-adsorbing polymers. J. Chem. Soc., Faraday Trans. 1991, 87 (14), 2201–2207. (23) Trappe, V.; Weitz, D. A. Scaling of the Viscoelasticity of Weakly Attractive Particles. Phys. Rev. Lett. 2000, 85 (2), 449–452. (24) Hamad, W. Y.; Hu, T. Q. Structureprocessyield interrelations in nanocrystalline cellulose extraction. Can. J. Chem. Eng. 2010, 88 (3), 392–402. (25) Habibi, Y.; Lucia, L. A.; Rojas, O. J. Cellulose nanocrystals: chemistry, self-assembly, and applications. Chem. Rev. 2010, 110 (6), 3479–3500. (26) Dong, X. M.; Revol, J.-F.; Gray, D. G. Effect of microcrystallite preparation conditions on the formation of colloid crystals of cellulose. Cellulose 1998, 5 (1), 19–32. (27) Lima, M. M. d. S.; Borsali, R. Rodlike Cellulose Microcrystals: Structure, Properties, and Applications. Macromol. Rapid Commun. 2004, 25 (7), 771–787. 6122
dx.doi.org/10.1021/la2035449 |Langmuir 2012, 28, 6114–6123
Langmuir
ARTICLE
(28) Boluk, Y.; Lahiji, R.; Zhao, L.; McDermott, M. T. Suspension viscosities and shape parameter of cellulose nanocrystals (NCC). Colloids Surf., A 2011, 377 (13), 297–303. (29) Bondeson, D.; Mathew, A.; Oksman, K. Optimization of the isolation of nanocrystals from microcrystalline cellulose by acid hydrolysis. Cellulose 2006, 13 (2), 171–180. (30) Clasen, C.; Kulicke, W.-M. Determination of viscoelastic and rheo-optical material functions of water-soluble cellulose derivatives. Prog. Polym. Sci. 2001, 26, 1839–1919. (31) Morris, E.; Cutler, A.; Ross-Murphy, S.; Rees, D.; Price, J. Concentration and shear rate dependence of viscosity in random coil polysaccharide solutions. Carbohydr. Polym. 1981, 1 (1), 5–21. (32) Dimova, R.; Lipowsky, R.; Mastai, Y.; Antonietti, M. Binding of polymers to calcite crystals in water: Characterization by isothermal titration calorimetry. Langmuir 2003, 19 (15), 6097–6103. (33) Mastai, Y.; Rudloff, J.; C€olfen, H.; Antonietti, M. Control over the structure of ice and water by block copolymer additives. ChemPhysChem 2002, 3 (1), 119–123. (34) Surve, M.; Pryamitsyn, V.; Ganesan, V. Nanoparticles in solutions of adsorbing polymers: pair interactions, percolation, and phase behavior. Langmuir 2006, 22 (3), 969–81. (35) Lekkerkerker, H. N. W.; Tuinier, R. Colloids and Depletion; Springer: New York, 2011; Vol. 100. (36) Asakura, S.; Oosawa, F. On Interaction between Two Bodies Immersed in a Solution of Macromolecules. J. Chem. Phys. 1954, 22 (7), 1255–1256. (37) Asakura, S.; Oosawa, F. Interaction between particles suspended in solutions of macromolecules. J. Polym. Sci. 1958, 33 (126), 183–192. (38) Mutch, K. J.; van Duijneveldt, J. S.; Eastoe, J. Colloid-polymer mixtures in the protein limit. Soft Matter 2007, 3 (2), 155–167. (39) Tuinier, R.; Vliegenthart, G. A.; Lekkerkerker, H. N. W. Depletion interaction between spheres immersed in a solution of ideal polymer chains. J. Chem. Phys. 2000, 113, 10768. (40) H€orner, K. D.; T€opper, M.; Ballauff, M. Assessment of the Depletion Forces in Mixtures of a Polystyrene Latex and Hydroxyethyl Cellulose by Turbidimetry. Langmuir 1997, 13 (3), 551–558. (41) Philipse, A. P.; Wierenga, A. M. On the Density and Structure Formation in Gels and Clusters of Colloidal Rods and Fibers. Langmuir 1998, 14 (1), 49–54.
6123
dx.doi.org/10.1021/la2035449 |Langmuir 2012, 28, 6114–6123
