pubs.acs.org/Langmuir © 2010 American Chemical Society
Interaction between Biopolyelectrolytes and Sparingly Soluble Mineral Particles Peter Versluis, Alois K. Popp, and Krassimir P. Velikov* Unilever R&D Vlaardingen, Olivier van Noortlaan 120, 3133 AT Vlaardingen, The Netherlands Received August 15, 2010. Revised Manuscript Received November 2, 2010 We investigate the complex physicochemical behavior of dispersions containing calcium carbonate (CaCO3) particles, a sparingly soluble mineral salt; and carrageenans, negatively charged biopolyelectrolytes containing sulfate groups. We reveal that the carrageenans suspend and stabilize CaCO3 particles in neutral systems by absorbing on the particle surface which provides electrosteric stabilization. In addition, carrageenans provide a weak apparent yield stress which keeps the particles suspended for several months. The absorption measurements of carrageenan on the CaCO3 particle indicate that more carrageenan is removed from the solution than expected from the case of a simple monolayer adsorption. Confocal laser scanning microscopy observations confirm that polyelectrolyte-containing precipitate is formed in both CaCO3-carrageenan and CaCl2-carrageenan mixtures. On the basis of these results, we confirm that in the presence of carrageenan some CaCO3 dissolves and the Ca2þ ions interact with the sulfate groups leading to aggregation and formation of particle-like structures. These new insights are important for fundamental understanding of other mineral-polyelectrolyte systems and have important implications for various industrial applications where calcium carbonate is used.
Introduction Polymers and, in particular, polyelectrolytes play an important role in defining the stability of colloidal dispersions and other soft matter systems.1 The addition of (bio)polymers induces two different effects that lead to an improved stability of dispersions. First, polymers can increase the viscosity of the continuous phase by forming a weak network and slowing down the sedimentation/ creaming phenomena in disperse systems. Second, the polymers can adhere to the particle/droplet surface and alter their surface properties and interactions.2-7 Charged biopolymers, acting as polyelectrolytes, can increase the net charge on the particle surface, leading to increased electrostatic repulsion.8,9 Additionally, a sufficiently thick layer of absorbed biopolymer can serve as a steric barrier to reduce the effect of the van der Waals attraction since the overlap of polymer chains is energetically unfavorable.1 These properties have been actively used to facilitate the dispersion2-4 of nanoscale particles and for their stabilization during synthesis. Particularly in the area of nanoscale colloidal particles, polyelectrolytes can alter or enhance the stability of the suspension. When the polymer chains are repelled by the surface (i.e., nonabsorbing polymer), a depletion interaction is created which is attractive in general but can turn to a repulsion for neutral particles mixed with polyelectrolytes.10 At high concentration of *Corresponding authors. E-mail:
[email protected]. (1) Napper, D. H. Polymeric Stabilization of Colloidal Dispersions; Academic Press: New York, 1984. (2) Tripathy, S. S.; Raichur, A. M. J. Dispers. Sci. Technol. 2008, 29, 230–239. (3) Karimian, H.; Babaluo, A. A. Iran. Polym. J. 2006, 15, 879–889. (4) Ishikawa, Y.; Katoh, Y.; Ohshima, H. Colloids Surf., B: Biointerfaces 2005, 42, 53–58. (5) Shafir, A.; Andelman, D. Phys. Rev. E 2004, 70. (6) Dzubiella, J.; Moreira, A. G.; Pincus, P. A. Macromolecules 2003, 36, 1741– 1752. (7) Claesson, P. M.; Poptoshev, E.; Blomberg, E.; Dedinaite, A. Adv. Colloid Interface Sci. 2005, 114, 173–187. (8) Fritz, G.; Schadler, V.; Willenbacher, N.; Wagner, N. J. Langmuir 2002, 18, 6381–6390. (9) Einarson, M. B.; Berg, J. C. J. Colloid Interface Sci. 1993, 155, 165–172. (10) Spalla, O. Curr. Opin. Colloid Interface Sci. 2002, 7, 179–185.
Langmuir 2011, 27(1), 83–90
polyelectrolytes, the dispersion can also undergo phase separation due to depletion interacitons.11 The interactions between mineral particles and biopolyelectrolytes is also important for understanding of biological processes such as biomineralization.12,13 Dispersing of particles of sparingly soluble minerals such as calcium carbonate is important for several industrial products such as paints14 and coatings,15 inks, paper, composite materials,16,17 cosmetic products, and oral care18,19 and house cleaning products. In particular, sparingly soluble minerals of essential micronutrients such as calcium, iron, zinc, and magnesium find important application in enhancement of the nutritional profile of food products, which is highly important for the well-being of the human population.20,21 The fortification of drinks or other liquid products with minerals poses severe technical challenges.22,23 In order to be effectively absorbed, the mineral has to be taken up by the human body in a dissolved form. While the use of soluble mineral salts is usually linked to good bioaccessibility, several adverse effects on the product often prevent the use of this route. Mineral ions are rather reactive; chemical reactions with other ingredients take place that change taste and physicochemical characteristics of the product. As an example, the presence of free calcium ions (11) Tuinier, R.; Rieger, J.; de Kruif, C. G. Adv. Colloid Interface Sci. 2003, 103, 1–31. (12) Butler, M. F.; Glaser, N.; Weaver, A. C.; Kirkland, M.; Heppenstall-Butler, M. Cryst. Growth Des. 2006, 6, 781–794. (13) az-Dosque, M.; Aranda, P.; Darder, M.; Retuert, J.; Yazdani-Pedram, M.; Arias, J. L.; Ruiz-Hitzky, E. J. Cryst. Growth 2008, 310, 5331–5340. (14) Kalendova, A.; Vesely, D.; Kalenda, P. Pigm. Resin Technol. 2007, 36, 123– 133. (15) Alm, H. K.; Strom, G.; Karlstrom, K.; Schoelkopf, J.; Gane, P. A. C. Nordic Pulp Paper Res. J. 2010, 25, 82–92. (16) Kaully, T.; Siegmann, A.; Shacham, D. Polym. Adv. Technol. 2007, 18, 696– 704. (17) Kaully, T.; Siegmann, A.; Shacham, D. Polym. Compos. 2007, 28, 512–523. (18) Joiner, A. Int. Dent. J. 2006, 56, 175–180. (19) Lynch, R. J. M.; ten Cate, J. M. Int. Dent. J. 2005, 55, 175–178. (20) Vyas, H. K.; Tong, P. S. J. Dairy Sci. 2004, 87, 1177–1180. (21) Kruger, M. C.; Gallaher, B. W.; Schollum, L. M. Nutr. Res. (N.Y.) 2003, 23, 1229–1237. (22) Weaver, C. M. Int. Dairy J. 1998, 8, 443–449. (23) Velikov, K. P.; Pelan, E. Soft Matter 2008, 1964–1980.
Published on Web 12/02/2010
DOI: 10.1021/la103253a
83
Article
Versluis et al.
leads to bitterness24 and can cause precipitation of proteins25 or gelation of some polysaccharides (e.g., alginate,26 pectin27). The approach commonly used to introduce calcium in liquid systems is suspending insoluble minerals such as calcium carbonate, which is also used in the paper industry28 and paints and for encapsulation.29 The solubility of sparingly soluble calcium salts is highly dependent on the pH, enabling the formulation of stable neutral drinks containing insoluble calcium minerals that dissolve at low pH in the stomach. The size of the mineral particles, however, deserves consideration. Large calcium-containing particles can sediment quickly due to their high density (typically higher that 2 g/cm3) and become unavailable for consumption. Large hard particles can also give rise to a sandy or chalky mouthfeel. While small calcium carbonate particles would be advantageous, due to the low surface charge30 of pure calcium carbonate they can aggregate if appropriate stabilization (e.g., charge or steric) is not assured. In the case of mineral particles with high density, a perceivable sediment at the bottom of the container will be formed that can be difficult to resuspend. A common strategy to improve stability of particulate suspensions is the addition of biopolymers that increase viscosity and create apparent yield stress, which delays or fully suppresses sedimentation. Carrageenans are linear, negatively charged biopolyelectrolytes extracted from red seaweed. Carrageenans are high-molecularweight polysaccharides made up of repeating galactose units and 3,6-anhydrogalactose, both sulfated and nonsulfated. The units are joined by alternating R-1-3 and β-1-4 glycosidic linkages. The primary structure of κ-carrageenan consists of a regular alternation of R(1-3)-D-galactose-4-sulfate and β-(1-4)-3,6anhydro-D-galactose. Other monosaccharides such as galactose, galactose-2-sulfate, galactose-6-sulfate, and galactose-2,6-disulfate may be present in minor amounts. Carrageenans are commonly applied as viscosifiers, dispersants, stabilizers, and (at high concentrations) gelling agents.31 Due to the presence of sulfate groups, carrageenans interact with mono and divalent ions such as calcium and can undergo conformational changes32-34 and form gels.35-37 In this paper, we investigate the complex interactions between calcium carbonate particles and carrageenan in aqueous solutions. The results provide new insights into biopolyelectrolyte adsorption and interaction with sparingly soluble particles as well as into the role of free metal ions for the behavior of the polyelectrolyte.
Materials and Methods Materials. Synthetic calcium carbonate (CaCO3) as dry powder (Socal U1) was obtained from Solvay Chemicals. The material (24) Neyraud, E.; Dransfield, E. Physiol. Behav. 2004, 81, 505–510. (25) Canabady-Rochelle, L. S.; Sanchez, C.; MELLEMA, M.; Banon, S. J. Agric. Food Chem. 2009, 57, 5939–5947. (26) Sikorski, P.; Mo, F.; Skjak-Braek, G.; Stokke, B. T. Biomacromolecules 2007, 8, 2098–2103. (27) Fang, Y. P.; Al-Assaf, S.; Phillips, G. O.; Nishinari, K.; Funami, T.; Williams, P. A. Carbohydr. Polym. 2008, 72, 334–341. (28) Enomae, T.; Tsujino, K. Appita J. 2004, 57, 493. (29) Fujiwara, M.; Shiokawa, K.; Morigaki, K.; Zhu, Y. C.; Nakahara, Y. Chem. Eng. J. 2008, 137, 14–22. (30) Tobori, N.; Amari, T. Colloid Surf. A: Physicochem. Eng. Asp. 2003, 215, 163–171. (31) Rodd, A. B.; Davis, C. R.; Dunstan, D. E.; Forrest, B. A.; Boger, D. V. Food Hydrocolloids 2000, 14, 445–454. (32) Janaswamy, S.; Chandrasekaran, R. Carbohydr. Res. 2002, 337, 523–535. (33) Janaswamy, S.; Chandrasekaran, R. Carbohydr. Res. 2008, 343, 364–373. (34) Nickerson, M. T.; Paulson, A. T.; Hallett, F. R. Carbohydr. Polym. 2004, 58, 25–33. (35) MacArtain, P.; Jacquier, J. C.; Dawson, K. A. Carbohydr. Polym. 2003, 53, 395–400. (36) Morris, V. J.; Belton, P. S. Progr. Food Nutr. Sci. 1982, 6, 55–66. (37) Nickerson, M. T.; Paulson, A. T. Carbohydr. Polym. 2004, 58, 15–24.
84 DOI: 10.1021/la103253a
data sheet claimed a primary particle size of 70 nm and a surface area of 20 m2/g. Carrageenan (κ- and λ-) was obtained from CP Kelco. The κ-carrageenan used in this study was a sample of a special batch (X0909) that had been treated by the ion exchange process to remove all ions other than sodium. Both carrageenans were prepared by dissolution in water while heating above 70 °C, necessary to obtain a random coil confirmation. Solutions were cooled to room temperature before use. Acridine orange was obtained from Fluka A.G. Deionized water purified by a Millipore purification system (Millipore Synergy 185) was used in all reactions. All chemicals were used as received. Dispersion Preparation. Samples were prepared by mixing the appropriate concentrations of calcium carbonate and carrageenan from stock solutions upon stirring. The dry powder was predispersed at 5 wt % in demineralized water. The dispersion was either sonicated or subjected to high-shear mixing (Turrax) before use. The natural pH of this dispersion had a value close to 10, presumably as a result of the traces of calcium hydroxide. During the mixing process, the pH of the samples was carefully reduced to pH = 7.3 ( 0.2 by slow addition of diluted HCl while stirring. Since the solutions are not buffered around this pH, care was taken not to go below pH = 7.1. Few samples were prepared with 0.1 M Tris buffer. All samples contained 0.3 wt % of calcium carbonate. Characterization. Rheology. Rheological measurements were made on a TA AR1000 rheometer equipped with a 4 cm, 2° cone at 20 °C. Great care was taken to carry out the proper rheometer corrections. Before starting a measurement run, an air bearing mapping was performed, and the high-speed bearing friction was also calibrated. In order to determine the precision of the instrument and test the calibrations, demineralized water was measured over a suitable shear rate range (5-250 s-1). Typically, the viscosity of water was found to be in the range 0.98 10-3 to 1.02 10-3 Pa.s (nominal value 1.0 10-3 for 20 °C). Viscosimetric measurements on low-viscosity fluids containing calcium and carrageenan were performed at shear rates between 1 and 5000 s-1, while the viscosity of pure carrageenan solutions up to 1 wt % was measured at lower shear rates. On selected samples, the dynamic rheological properties (G0 , the elastic contribution, and G00 , the viscous contribution) were established in order to detect the presence of a gel network. These measurements were carried out at oscillation frequencies of 500.1 Hz at a strain well within the linear visco-elastic region (1 wt %). Particle Sizing. Static light scattering (SLS) measurements were used to determine the volume-weighted particle size distributions. Measurements were performed on Mastersizer 2000 (Malvern Instruments, Malvern, UK). A few droplets of the dispersion are placed into a water-filled tank of ∼100 mL volume, which is connected to a measurement window, equipped with lenses. A refractive index of n = 1.57 was assumed for CaCO3. In this case, for qualitative comparison only, particles were assumed to be spherical. The diluted solution is constantly cycled between tank and measurement window through a tube-equipped pumping system based on a mechanical stirring device integrated in the dispersing unit. In our investigation, a medium stirring speed of 1750 rpm has been chosen for all measurements. The instrument is able to perform repeated measurements on the same, recirculated dispersion over time to enable dissolution/aggregation kinetics to be measured. ζ-Potential. The ζ-potential was measured using a Zetasizer Nano ZS (Malvern Instruments Ltd.) which measures the distribution of the electrophoretic mobility and then converts them to ζ-potentials using Smoluchowski or H€ uckel theory depending on the conductivity of the sample measured. The samples were measured at different polyelectrolyte concentrations at a temperature of 25 °C. All the measurements were performed in disposable sizing cuvettes on diluted dust-free dispersions in water. Calcium Concentration Measurements. Total calcium concentration was measured using inductive coupled plasma-optical emission spectrometry on Perkin-Elmer Optima 3300 DV. Free Langmuir 2011, 27(1), 83–90
Versluis et al.
Article
calcium ion (Ca2þ) concentrations were measured with a calciumsensitive electrode WTW Ca 500 DIN combined with a reference electrode WTW R 503/D attached to a suitable voltmeter. The output was measured with a pH meter (mV mode). Measurement of Absorbed Polyelectrolyte. A series of samples were prepared containing κ-carrageenan at concentrations between 0 and 0.35 wt % and λ-carrageenan between 0 and 0.2 wt %, as well as 1:1 mixtures of κ- and λ-carrageenan. The samples were prepared without any buffer. After preparation and conditioning of the sample, the dispersion was left undisturbed for two weeks to allow complete particle sedimentation. Aliquots of the clear supernatant were taken, and the supernatant was analyzed with respect to viscosity, which is correlated to the concentration of polyelectrolyte still present in solution. It was assumed that (a) the adsorption process had been completed and (b) no more particles were present in the supernatant. From a calibration curve, consisting of the viscosities of known biopolymer concentrations at the same ionic environment and temperature (see Figure S1), the supernatant biopolymer concentration can be obtained. In order to apply this method to the problem studied here, viscosity calibration curves for κ-carrageenan, λ-carrageenan, and a mixture were measured in the presence of 2.2 mM Ca2þ (added as CaCl2), the same concentration of free calcium as present in the particle-carrageenan mixtures. Then, a simple subtraction of the supernatant concentration from the known initial biopolymer concentration allows a determination of the amount of the polyelectrolyte removed from the dispersions. A representative value of the viscosity, at a shear rate of 250 s-1, was taken for the calculations. At this shear rate, the viscosity is still at the first Newtonian plateau and will therefore be most sensitive to concentration changes. The required shear stress is also well above the minimum that the rheometer can achieve and can reliably be reproduced. At the applied shear rate, there is also no danger of turbulence effects that might lead to excessively high viscosities. Then, the viscosities of the supernatants of carrageenan/ particle mixtures after long storage were measured according to the same method, and the viscosity at a shear rate of 250 s-1 was used for further data analysis. The concentration of polyelectrolyte removed due to interaction with the particles was related to the total available particle surface of 0.3 wt % particle dispersion. The particle surface was taken as 20 m2/g, as specified by the supplier. The amount of adsorbed polyelectrolyte was calculated by dividing the concentration of precipitated polyelectrolyte (in weight %) by the total available surface of 0.3% particles (=6.66 m2). Confocal Scanning Laser Microscopy. Confocal scanning laser microscopy (CSLM) was used to inspect some of the calcium carbonate sediments that formed during storage. A small amount of the fluorescent stain acridine orange was added to a few milliliters of sample that was pipetted into sample chamber. This fluorescent probe binds to carrageenan and allows localization of the biopolyelectrolyte in the sediment.38 The chamber was placed onto the sample holder of the microscope equipped with a 40 oil objective and the bulk of the liquid was imaged immediately. The confocal microscope consists of a Biorad laser and scanning head system linked to an inverted Zeiss microscope. Fluorescence at 632 nm excited by the scanning krypton/argon laser (512 nm) was used for observation.
Results and Discussion Particle Characterization. ζ-Potential. In Figure 1, the ζ-potential of calcium carbonate particles in the presence of different concentrations of λ-carrageenan are displayed. Without biopolyelectrolyte, the calcium carbonate particles do not possess (38) Rees, D. A.; Williamson, F. B.; Frangou, S. A.; Morris, E. R. Eur. J. Biochem. 1982, 122, 71–79. (39) Vdovic, N.; Kralj, D. Colloids Surf., A: Physicochem. Eng. Asp. 2000, 161, 499–505.
Langmuir 2011, 27(1), 83–90
Figure 1. ζ-Potential of calcium carbonate particles as a function of λ-carrageenan concentration.
any significant surface charge density and tend to quickly aggregate when dispersed in water.39-41 Addition of λ-carrageenan at concentrations of at least 0.1 wt % leads to a ζ-potential of approximately -20 mV. The addition of carrageenan to the particles leads to significantly more stable systems, due to polymer adsorption onto the particles and the increase in surface charge density leading to electrostatic repulsion between the particles. Interestingly, the ζ-potential did not increase linearly with polymer concentration, but showed a saturation already at 0.1 wt % carrageenan. While the suspension stability is strongly improved by addition of a small amount of carrageenan, reflected by strong ζ-potential increase, further addition of polymer seems to have no further effect on the particle stability by electrostatic repulsion. However, steric repulsion and network formation by the polymer might be more important once polymer concentration is increased. The changes of volume-weighted average size of calcium carbonate particles without and with carrageenan was measured by SLS. The sample is placed into a dispersant tank of 100 mL volume filled with water as dispersing agent and then automatically pumped in a closed loop to the measurement cell and recirculated until the measurement is performed. While the use of a dispersion unit with pumping process involving a mechanical stirrer ensures good dispersion and dilution of concentrated samples, it also influences the sample properties over time. At the initial state of particle dispersal, changes in the particle size distribution were encountered in repeated measurements. Consequently, a series of measurements over an expanded time span was conducted, with the sample being recirculated between measurements with constant stirring speed. This procedure was performed on pure particle suspension without polymer and on mixtures of particles with λ-carrageenan. The mixtures were prepared to yield 0.3 wt % calcium carbonate in the presence of 0.2 wt % and 0.6 wt % λ-carrageenan before the measurement procedure. Small aliquots of these suspensions were pipetted into the dispersal unit, where they were diluted and pumped to the measurement window upon stirring. Measurements were taken every 2.5 min, over the time span of one hour. During that time, the sample was constantly recirculated by mechanical stirring. Samples consisting of calcium carbonate particles alone showed initially a size distribution of a main peak located at 3 μm, with a small satellite peak at 20 μm (Figure 2A). Over time, the peak size of the initially main peak slightly shifts toward a smaller size of ∼2 μm, with the peak at 15-20 μm steadily growing. After one hour, the size distribution became more complex, with a major peak at 15 μm whose shoulder extends to sizes above 100 μm (40) Chibowski, E.; Hotysz, L.; Szczes, A. Colloids Surf., A: Physicochem. Eng. Asp. 2003, 222, 41–54. (41) Moulin, P.; Roques, H. J. Colloid Interface Sci. 2003, 261, 115–126.
DOI: 10.1021/la103253a
85
Article
Versluis et al.
Figure 2. Temporal evolution of particle size of suspended calcium carbonate. The particles have been measured over a time span of 1 h, with sample recirculating between dispersion unit and the measurement window, driven by a stirring device. In (A), the whole size spectrum is displayed for pure calcium particles. Within the measurement duration, the calcium particles of initial size close to 4 μm aggregate (aided by the stirrer device), leading to a bimodal size distribution skewed toward a peak at 20 μm. In contrast, particles suspended in the presence of 0.2 wt % or 0.6 wt % carrageenan show an initial decrease in size, followed by no further changes in particle size after 5 min (B), with the final size being ∼2 μm.
(Figure 2A). We attribute this behavior to the growth of large aggregates of calcium carbonate particles. The situation changes when samples containing both particles and carrageenan are measured. Initially, one peak at ∼5-6 μm was obtained that after an initial decrease, probably caused by initial dissolution effects of the concentrated suspensions, stayed rather constant over the course of the complete measurement cycle (Figure 2B). The complete absence of aggregation effects can only be explained if particle stabilization by the polyelectrolyte is considered. After elimination of any effects from particle size changes, a small effect that might reflect dissolution of micrometer-sized particles under stirring can be detected. The D[3,2] values derived from the overall size distributions, shown as function of time after 86 DOI: 10.1021/la103253a
the first measurement, give an impression of the substantial changes in particle behavior once charged biopolymers are added. Suspending and Stabilization Properties of λ-Carrageenan Solutions. To investigate the stability, calcium carbonate suspensions (0.3 wt %) were mixed with different concentrations of λ-carrageenan, between 0 and 1% wt. In order to ensure longterm stability of the initially adjusted pH of 7, the samples were prepared in 0.1 M Tris buffer. Then, the stability of the suspensions in closed glass jars was monitored over two months. Photographs of the samples at 40 h and 2 months after preparation and storage at room temperature are shown in Figure 3. A white precipitate is already formed after 40 h in samples with less than 0.6 wt % carrageenan present. This result shows that the Langmuir 2011, 27(1), 83–90
Versluis et al.
Article
Figure 3. Photographs of calcium carbonate particles (0.3 wt %) suspended in λ-carrageenan solution after 40 h (top) and two months (bottom). Samples without addition of carrageenan sediment within 40 h, indicating the presence of large particles. Low polymer concentrations of 0.1-0.2 wt % do not prevent sedimentation, while the presence of 0.4-0.8 wt % carrageenan leads to more complex phase separation and sedimentation processes on a longer time scale. Only the sample with the highest concentration in carrageenan does not seem to show any sedimentation.
calcium carbonate particles, due to their high density compared to that of water, start to sediment, which can only be prevented by surprisingly high concentrations of polymer, far above concentrations necessary for network formation. The formation of sediment is due to the presence of particle aggregates which are much larger than the primary particle size of 70 nm (Figure S2). Upon drying, the particles were already aggregated. The particles are initially present with larger size, or the initially small particles, due to van der Waals attraction, have aggregated rather quickly, with the viscosity increase provided by the polymer only high enough at concentrations above 0.6 wt % to prevent quick aggregation. After two months storage, the particles in all samples apart from that with 1 wt % λ-carrageenan show different stages of sedimentation, with the samples containing 0.6% and 0.8% carrageenan showing separation into a turbid phase of high phase volume and a clear phase. The quick, seemingly complete particle sedimentation in the suspensions containing 0-0.2 wt % carrageenan can be interpreted as either the polymer network not being strong enough to keep the particles suspended or the polymer aggregate on the particles with too little polymer left for network formation. Then, a major part of the polymer would be present in the sediment also. The suspension contained 0.4% carrageenan sediments on a longer time scale, but after 2 months, the same white sediment as in the samples of lower polymer concentration is observed. After 40 h, however, in contrast to the samples with lower polymer content, a threephase system is obtained consisting of dense white sediment on the bottom and a turbid phase in the middle just separating from a clear phase close to the meniscus. At an even higher carrageenan concentration of 0.6 wt %, a single, highly turbid phase is seen after short storage of 40 h. After two months, the system phase-separates into a clear and highly turbid phase. The only system that seems to be stable is the one with 1 wt % carrageenan. In the following, this surprisingly complex behavior will be investigated in more detail, by analyzing supernatant and sediment with respect to their composition, in order to link the macroscopic phase behavior shown in Figure 3 to a certain type of particle-polymer interaction. Langmuir 2011, 27(1), 83–90
Polyelectrolyte Network Properties and Particle Sedimentation. Since the observation of short-term stable systems at carrageenan concentrations above 0.4 wt % could be explained by the formation of a weak polyelectrolyte gel that slowly collapses under the load of suspended particles, the samples were investigated by oscillatory rheology under low strain in a frequency sweep. The presence of a gel is generally correlated with rheological spectra showing a dominating elastic modulus G0 , and a much lower loss modulus G00 . Consequently, the samples were tested in the rheometer at 1 Hz and low strain for the detection of gel formation. For all concentrations studied here, G0 was much lower than G00 , indicating the absence of an elastic gel network (data for the highest concentration 1 wt % shown in Figure S3). The apparent resistance against sedimentation can thus not be the result of an elastic network but is an effect of high viscosities at low shear rates.31 Therefore, gel formation was ruled out, and the effect of carrageenan on particle sedimentation was attributed to its viscosifying function. Thus, carrageenan at the concentrations shown in Figure 4 should show a strong dependence of viscosity on shear rate, with very high viscosities at low shear rates. In Figure 4, the flow curves obtained for carrageenan concentrations between 0.2% and 1 wt % are shown, mainly for shear rates between 0.05 s-1 and 1 s-1. The shear thinning effect of carrageenan concentrations of 0.4% and higher can lead to viscosity differences of more than 2 orders of magnitude. In order to estimate the importance of the viscosifying effect on particle sedimentation velocity, U, was estimated using Stoke’s law, U = 2gr2ΔF/9η, where g is the gravity constant, ΔF is the density difference between particles and suspending fluid, r is the particle radius, and η is the viscosity of the continuous phase. From sedimentation velocity and particle radius, the shear rate around the falling sphere can be estimated to γ_ = 3U/2r. In turn, this calculation can also be used to estimate the shear rate at a given sedimentation velocity. For our estimates, values for the gravity constant g of 9.81 kg/m2, ΔF ∼2500 kg/m3, the particle radius r ∼4 μm, and the viscosity η were assumed. In order to check the relevance of the measured low-shear viscosities, we estimated the shear rate required to obtain a sedimentation rate of DOI: 10.1021/la103253a
87
Article
Figure 4. Viscosity as function of shear rate, for different λ-carrageenan levels. While the lowest carrageenan concentration of 0.2 wt % shows a flat line as a function of shear rate, higher carrageenan concentrations show shear rate dependent viscosities and thus the transition from dilute to semidilute. In comparison, the viscosities and shear rates experienced by sedimenting calcium carbonate particles of velocity 1 mm/month have been calculated using Stoke’s law. Particle sizes between 1 and 16 μm have been assumed (red squares). The red squares are the calculated viscosities and shear rates experienced by a sedimenting particle at 1 mm/ month. The particle size increases from 1, 2, 4, 8, and 16 μm from bottom to top. At very low shear rates, the data obtained for 1% carrageenan correspond very well to the viscosities and shear rates involved in slow sedimentation processes.
1 mm/month for particles of a different diameters. The red squares in Figure 4 represent results of these calculations, with the particle size varied between 1 and 16 μm. The measured viscosities for 1 wt % carrageenan at the lowest shear rates (∼0.001 s-1) seem to overlap for the estimate obtained for particle size of 4 μm. Although this type of calculation is only suited to provide a rough estimate of the suspension stability, the obtained prediction is remarkably accurate, especially in light of results obtained from particle sizing (Figure 2). Calcium-Polyelectrolyte Interactions. Carrageenans interact with calcium ions and can undergo conformational changes32-34 and form gels.35-37 In order to determine the effect of calcium ions on the mechanical properties of the carrageenan solution; the amount of free calcium ions in the supernatant of calcium carbonate-carrageenan suspension (Figure 3) was first determined using a calibrated calcium-selective electrode. From the calcium electrode measurements, constant total calcium levels in supernatants of 2.20 ( 0.15 mM were obtained. The effect of free calcium on network properties of carrageenan was investigated directly by rheology. The viscosity of various solutions of carrageenan of a concentration of 0.3 wt % containing different concentrations of calcium (added in the form of CaCl2) was measured at an extended range of shear rates in order to determine both the constant low shear viscosity and the influence of network effects leading to a significant reduction of viscosity at high shear rates. Figure 5 shows viscosity measurements of 0.3% wt carrageenan solutions, to which up to 20 mM calcium chloride was added. The strongest decrease in viscosity is already obtained by addition of only 1 mM Ca2þ, with the viscosity dropping by a third. Further increase in calcium concentration to 4 mM leads to a reduction in viscosity to about a third of the calcium-free sample. Concentrations higher than 4 mM do not lead to an additional viscosity reduction. At this state, 0.3 wt % carrageenans are already only six times more viscous than water at low shear. The shear-thinning effect at high shear rates is related to the network properties of 88 DOI: 10.1021/la103253a
Versluis et al.
Figure 5. Influence of free Ca2þ concentration on the viscosity of
0.3 wt % κ-carrageenan solution, measured at 25 °C. The viscosity of each solution was measured at shear rate between 1 s-1 and 5000 s-1, in order to obtain the Newtonian plateau of constant viscosity. Increasing calcium concentration weakens the polymer network. At concentrations above 4 mM calcium, the carrageenan viscosity becomes independent of calcium.
the biopolymer; carrageenan polymers at a concentration of 0.3 wt % are already strongly overlapping. With the reduction of low-shear viscosity upon calcium addition, the shear thinning effect at high shear rates is also reduced. A general interpretation of this behavior necessarily involves the exchange of the sodium ions of the sodium-κ-carrageenan by calcium ions, with the surplus positive charges of the bivalent ion leading to a repulsive force between the now-likewise-charged polymers, weakening the polymer entanglements. Alternatively, the presence of Ca2þ causes conformational changes of the polyelectrolyte coils to a more condensed (less space-filling) form. The opposite effect of calcium bridges between polymer chains seems less relevant in this case, since a strengthening effect of the polymer network and thus a viscosity increase would result. The slight increase in the viscosity at shear rates above 1000 s-1 as seen in the low-viscosity samples of Figure 5 is probably caused by the occurrence of turbulence. The measured values of 2 mM free calcium in supernatants of carrageenan-containing calcium carbonate dispersions in combination with rheological measurements show explicitly that calcium ion release from calcium carbonate particles does have a negative influence on the carrageenan network applied for stabilization of the particles. While free calcium has been shown to have a strong influence on the network properties of carrageenan solution (Figure 5), this effect might be due to a pure increase of electrostatic repulsion effects by the increase in ionic strength, or directly through an exchange of the sodium ions bound to the sulfate groups by calcium. In order to distinguish between these two effects and check whether Ca-carrageenan complexes according to the reaction Ca-ions þ CarrageenanTCa-carrageenan are preferably formed, a calcium electrode was used to the concentration of free calcium in solution. Thus, the calcium level in a CaCl2 solution upon addition of carrageenan can be monitored, and a drop in free calcium, detected by the electrode, would show the formation of Ca-carrageenan complexes. Unfortunately, for the Ca-electrode to perform properly the pH and ionic strength of the solution have to constant; therefore, it is required to add a buffer. Here, a 0.1 M Tris buffer of pH = 7.2 has been chosen. It cannot be ruled out that the presence of the buffer ions will influence the result. First, a calibration curve was Langmuir 2011, 27(1), 83–90
Versluis et al.
Figure 6. Concentration of free Ca2þ ions in 30 mM CaCl2 solu-
tion as a function of the added κ-carrageenan.
obtained for the calcium electrode. CaCl2 was titrated into the buffer solution in small steps and the known calcium concentration related to the potential of the electrode. Then, small quantities of κ-carrageenan solution were added. Binding of free Ca-ions to the biopolymer can be observed as a drop in the electrode potential. The precision of the method is high enough to detect a significant drop in free calcium concentration (Figure 6) for carrageenan concentrations above 0.05% wt. The exchange of monovalent ions by divalent ions bound to carrageenans can lead to strengthening of the polymer network, including possible gelation through bridging effects, as well as network weakening effects due to stronger electrostatic repulsion. The domination of one of these effects will then depend on ionic strength of further ions, carrageenan sulfate group density, and the concentrations of polymer and free calcium ions. In a suspension with solid calcium carbonate present, complex formation between Ca-ions and carrageenan would not lead to a significant drop in free Ca-ion level because of the equilibrium between solid calcium carbonate and free calcium ions whose disturbance by removing free Ca2þ will be restored by dissolution of some CaCO3. A significant increase in the concentration of free CO32- in solution caused by binding its counterion to the polymer could then be detected by an increase in pH. A rather simple experiment, pH-measurements upon titration of κ-carrageenan into a calcium carbonate particle suspension, proved to be sensitive enough to demonstrate the hypothesized changes in ionic environment. An increase in pH from 7.06 for pure particle dispersion to 7.4 for particles in the presence of 0.2 wt % κ-carrageenan was detected (Figure S4). The amount of CaCO3 that could dissolve to allow binding of Ca2þ to κ- or λ-carrageenan can be estimated by assuming that one Ca ion can bind to two sulfate groups of the biopolyelectrolyte. The equivalent molecular weight of one structure unit of κ-carrageenan is 384 (only one sulfate group) and that of λ-carrageenan is 559 (three sulfate groups). In the case of λ-carrageenan, the highest concentration we used was 0.2 wt %, and the loss after being in contact with calcium carbonate was 0.045% wt. This is equivalent to 0.012 wt % CaCO3 being dissolved. For κ-carrageenan, we used a maximum of 0.35 wt % polyelectrolyte and lost 0.11 wt %. This is 0.014 wt % of the total 0.3 wt % CaCO3 being dissolved. In both cases, this leads to a loss in CaCO3 of less than 5%. This amount is close to the 6.7 wt % (∼2 mM from 30 mM) of dissolved Ca measured in the supernatant, which consists very likely of a combination of free Ca ions and partially bound Ca ions to the polyelectrolyte. While the coating of calcium carbonate particles by charged carrageenans leads to an increased charge density on the surface, the charge density saturation at 0.1 wt % carrageenan gives rise to the question whether further increase of the polymer concentration Langmuir 2011, 27(1), 83–90
Article
Figure 7. Apparent adsorption isotherms for κ- and λ-carrageenan and of a 1:1 mixture of these biopolymers onto calcium carbonate particles (0.3 wt %). The curve is a guide to the eyes.
would still lead to polymer adsorption and thus in a further increase of steric repulsion effects. Similar behavior has been observed with other negatively charged polyelectrolytes.42 In Figure 7, the adsorbed amount of biopolymer is plotted against the polymer concentration in solution. The amount adsorbed is given as mg/m2 of particle surface. Since both calibration curves and supernatant viscosities showed no significant differences between κ- and λ-carrageenan, an influence of polymer type on absorbed polymer is absent. Apparently, for the interaction between biopolymer and particles at 2.2 mM free Ca2þ, the amount of charged sulfate groups on the biopolymer backbone is not important. Graphs like that shown in Figure 7 are known as adsorption isotherm. The classical Langmuir adsorption pattern, with a saturation effect leading to a plateau of adsorbed molecules at high concentrations of available molecules, is not observed here. Similar behavior is associated with multilayer adsorption or limited solubility of the adsorbing molecules (shown as the asymptote at high concentrations) and is known as BET-isotherm. Neither explanation offers a satisfying reason for the pattern observed here. Multilayer adsorption usually leads to distinct plateaus in the isotherm, which is are not observed in our measurements. Furthermore, the carrageenans that are used here are easily soluble in water to 3 wt % or more. It was hypothized that in some way the adsorption results are influenced by the presence of Ca2þ ions. It is well-known that at higher κ- and or λ-carrageenan the addition of Ca2þ ions can lead to the formation of a gel-like texture. The carrageenans used in this investigation give turbid gels at 1 wt % biopolymer and 20 mM CaCl2.35,37,43 While the supernatant has been investigated extensively, it is much more difficult to perform a similar investigation on the precipitate. While the hypothesis of an exchange of sodium by free Caþ has been proven by direct measurements of the ion concentration and the free carbonate, the consequences on polymer microstructure and adsorption of this finding are still not clear. Most importantly, we have not addressed whether the calciumcarrageenan complex is soluble or not, since it was found that the total calcium levels in the supernatants were (nearly) constant and did not depend on the κ-carrageenan level. This strongly suggests that the calcium-carrageenan precipitates, presumably as a coprecipitate with the calcium carbonate particles. In order to prove this, these samples were prepared with and without particles (0.3 wt %), with 0.3 wt % κ-carrageenan, and with 2 or 10 mM Ca-ions added. Acridine orange was added to fluorescently label the κ-carrageenan. The samples were put in CSLM sample (42) Tobori, N.; Amari, T. Colloids Surf., A: Physicochem. Eng. Asp. 2003, 215, 163–171. (43) Hermansson, A. M.; Eriksson, E.; Jordansson, E. Carbohydr. Polym. 1991, 16, 297–320.
DOI: 10.1021/la103253a
89
Article
Versluis et al.
Figure 8. CSLM observations of sediments in systems containing calcium and with κ-carrageenan. The width of the images is 65 μm. (A) 0.3% CaCO3; (B) 0.3% CaCO3 and 0.3% κ-carrageenan; (C) κ-carrageenan and 2 mM Ca2þ; and (D) κ-carrageenan and 10 mM Ca2þ. The pH of all dispersions was set to 7.3.
holders and allowed to stand for a few days before the samples were observed from the bottom. Figure 8A shows a fluorescent background without structural features with embedded black spots. The black spots are the CaCO3 particles that have sedimented to the bottom. Figure 8B shows fluorescent particles that are most likely Ca-carrageenan particles that have fallen out of solution and have coprecipitated with the CaCO3 particles. Figure 8C,D shows sediments from carrageenan solutions with added Ca2þ ions. It can be seen that a Ca2þ ion concentration of at least 10 mM is required to form a detectable quantity of Cacarrageenan particles. This supports the idea that some of the CaCO3 dissolves as carrageenan is added.
Conclusions We have investigated the complex interactions in suspensions composed of calcium carbonate particles in the presence of carrageenans, a negatively charged biopolyelectrolyte containing sulfate groups. The addition of solid particles of sparingly soluble minerals to aqueous solution containing biopolyelectrolytes is always linked to solubility equilibrium. In general, effects from slow dissolution and the appropriate ionic conditions to accelerate or suppress this effect have to be taken into account. Furthermore,
90 DOI: 10.1021/la103253a
effects of the dissolved ions on biopolymers and other constituents have to be taken into consideration. The increase in ionic strength or specific ion interactions (e.g., Ca2þ) that influence transitions, aggregation, or gelation can have consequences on microstructural level, leading to macroscopic effects like instability, failure, or long-term gelation effects. The current study indicates that carrageenan used to suspend and stabilize CaCO3 particles in neutral products are able to absorb on the particle surface and provide charge and steric stabilization that prevent reaggregation of the dispersed particles. In addition, carrageenans are able to provide a high enough viscosity to keep particles suspended for several months. The absorption measurements of carrageenan on the calcium carbonate particles indicated more carrageenan sediment than the amount expected from the case of a simple adsorption. Confocal measurement confirms that the biopolymer-containing sediment was formed in both cases: CaCO3-carrageenan and CaCl2-carrageenan mixtures confirming that CaCO3 dissolution and interaction with carrageenan take place. Additional experiments using a soluble calcium source revealed that Ca2þ ions can cause precipitation of carrageenans. On the basis of these results, we are able to confirm that some CaCO3 actually dissolves and the available Ca2þ ions interact with the sulfate groups of carrageenan leading to some aggregation and formation particle-like structures. These new insights are important for understanding the behavior of sparingly soluble minerals such as calcium carbonate that is used as for nutritional supplementation, as pigments or abrasives, or in other industrial applications in the presence of biopolyelectrolytes. Acknowledgment. We thank C. Remijn for performing the inductive coupled plasma-optical emission spectrometry analysis, H. Blonk and M. van Ruijven for performing the confocal measurements, and J. Hazekamp for performing the scanning electron microscopy. Part of this work is financially supported by the Food & Nutrition Delta 2 program (grant DFN0642300). Supporting Information Available: Viscosity (measured at 10 s-1) as a function of polyelectrolyte concentration calibration curves for -carrageenan, -carrageenan and a 1:1 mixture thereof, in the presence of 2.2 mM CaCl2. SEM micrograph of calcium carbonate particles. Storage and loss modulus of the supernatant containing 1 wt % -Carrageenan and 3 wt % calcium carbonate particles. pH increase as result of adding -carrageenan to a calcium carbonate dispersion. -carrageenan is titrated into a suspension of calcium carbonate particles, and the pH of the suspension was measured. This material is available free of charge via the Internet at http://pubs.acs.org.
Langmuir 2011, 27(1), 83–90
