A New Method for Obtaining Adsorption Isotherms on Colloidal

Colloidal Suspensions via Electrokinetic Sonic Amplitude. Measurement. Kleber L. Guimara˜es,† Ricardo H. R. Castro,*,‡ and Douglas Gouvêa†. De...
0 downloads 8 Views 139KB Size
Langmuir 2005, 21, 11645-11650

11645

A New Method for Obtaining Adsorption Isotherms on Colloidal Suspensions via Electrokinetic Sonic Amplitude Measurement Kleber L. Guimara˜es,† Ricardo H. R. Castro,*,‡ and Douglas Gouveˆa† Department of Metallurgical and Materials Engineering, Escola Polite´ cnica, Universidade de Sa˜ o Paulo, Av. Prof. Mello Moraes, 2463, Sa˜ o Paulo, SP, 05508-900, Brazil, and Centro Universita´ rio da FEI, Av. Humberto A.C. Branco, 3973, Sa˜ o Bernardo do Campo, SP, 09850-901, Brazil Received June 23, 2005. In Final Form: September 8, 2005 The standard methods for obtaining adsorption isotherms on colloidal suspensions are usually very time consuming and involve a large number of steps and assumptions that increase the experimental errors. In this work, an alternative method is proposed to evaluate the adsorption behavior of electrostericstabilized systems based on electrokinetic sonic amplitude signal measurements. The new method, entitled “zeta-sorption”, is noticeably less time-consuming when compared to conventional procedures but showed great precision and reliability confirmed by comparison with data obtained from conventional routes on alumina-polyacrylate and alumina-citric acid aqueous suspensions. The experimental conditions that restrict the applicability of the new method were identified and justified by discussing the possible ion exchanges.

Introduction Adsorption isotherms are of fundamental importance for colloidal processing since they are closely related to the optimal dispersant concentration, which can minimize costs and avoid undesirable rheological properties caused by dispersant excesses. However, the common procedures for obtaining these isotherms are very time consuming, and moreover, a number of assumptions and correction factors must be considered to obtain reasonable data, which may input considerable uncertainties on obtained results.1-4 Standard experimental procedures on solid/liquid interfaces usually involve a centrifugation or filtration step after stabilization of the solid-dispersant suspension. Analytical techniques, such as first derivative titration analysis and infrared and ultraviolet spectroscopies, are thereafter applied to determine the dispersant supernatant concentration. The adsorbed dispersant content is calculated from the difference between the total dispersant amount and the supernatant concentration.1-3,5-10 * To whom correspondence may be addressed. E-mail: rhrcastro@ fei.edu.br. † Universidade de Sa ˜ o Paulo. ‡ Centro Universita ´ rio FEI. (1) Hidber, P. C.; Graule, T. J.; Gauckler, L. J. J. Am. Ceram. Soc. 1995, 78, 1775. (2) Hoogendam, C. W.; Keizer, A.; Cohen Stuart, M. A.; Bijsterbosch, B. H. Langmuir 1998, 14, 3825. (3) Jucker, B. A.; Harms, H.; Hug, S. J.; Zehnder, A. J. B. Colloid Surf., B 1997, 9, 331. (4) Lee, D. H.; Condrate, R. A. Abstr. Pap. Am. Chem. S. 1995, 209, 314. (5) Hidber, P. C.; Graule, T. J.; Gauckler, L. J. J. Eur. Ceram. Soc. 1997, 17, 239. (6) Davies, J.; Binner, J. G. P. J. Eur. Ceram. Soc. 2000, 20, 1539. (7) Johnson, S. B.; Franks, G. V.; Scales, P. J.; Boger, D. V.; Healy, T. W. Int. J. Miner. Process. 2000, 58, 267. (8) Stuart, M. A. C.; Fleer, G. J.; Lyklema, J.; Norde, W.; Scheutjens, J. M. H. M. Adv. Colloid. Interface Sci. 1991, 34, 477. (9) Persson, M. In Surface and Colloid Chemistry in Advanced Ceramic Processing; Pugh, R. J., Bergstrom, L., Eds.; Marcel Dekker: New York, 1994; Vol. 51. (10) Hidber, P. C.; Graule, T. J.; Gauckler, L. J. J. Am. Ceram. Soc. 1996, 79, 1857.

The particle segregation procedures (centrifugation/ filtration) have a loss of reliability once practical situations involving fine particles and high molecular weight polymer chains are considered. The problem is related to the total elimination of particles in suspension after centrifugation. It has been reported that even after repeated centrifugation cycles a small quantity of alumina powder may remain in the supernatant of alumina-poly(acrylate) suspensions.6 This is due to the quite small particle sizes and high stabilization states. However, despite current procedures, repeated centrifugation cycles can cause the supernatant dispersant concentration to decrease as a function of the centrifuge cycle parameters.11 Specially when considering very high molecular weights (150 000 g mol-1), extending time and speed of centrifugation contributes to reduce the supernatant dispersant concentration, causing an overestimation of the adsorbed amount.11 Davis and Binner6 proposed a correction factor to compensate for the incapability of obtaining a particlefree supernatant, but a complete correction factor proposal is a difficult task since one must take into account several factors such as the following: (a) the fraction of remaining suspended particles is not a constant once flocculated and stabilized systems do not have equal settling conditions; (b) the surface coverage (adsorption efficiency) also depends on the total amount of dispersant added; (c) the active sites distributed over the powder surface also affect the titration readings once they also participate on the ion exchange reactions. Seeking to overcome these inconveniences, Castro and Gouveˆa11 proposed a method to directly evaluate the adsorbed amount based on the measurement of the dynamic electrophoretic mobility (µD) of colloidal particles present in dilute suspensions. The method was applied to quantify the adsorption efficiency of chitosan onto SnO2 particles at a specific pH condition. The advantage of this method is that adsorption was evaluated in a one-step experiment, without centrifugation or filtration, directly calculating the adsorbed amount by the surface charges (11) Castro, R. H. R.; Gouvea, D. J. Eur. Ceram. Soc. 2003, 23, 897.

10.1021/la051691l CCC: $30.25 © 2005 American Chemical Society Published on Web 10/22/2005

11646

Langmuir, Vol. 21, No. 25, 2005

Guimara˜ es et al.

generated by the dispersant. However, the method was not compared with data obtained by conventional routes, since the treated dispersant, a polysaccharide (chitosan), has a molecular weight that was too high and centrifugation effects were too pronounced. The intent of this paper is to present the method in more detail, testing its applicability by comparing the attained results with adsorption data obtained through a conventional route on a R-Al2O3-citric acid system. Since citric acid is a quite small molecule, the centrifugation effect can be neglected and comparison is reasonable. Also proposed is the improvement of standardizing the definitions and conditions while attempting to reduce involved errors. The new procedure, entitled the “zeta-sorption” method, was finally employed to study the adsorption efficiency of low molecular weight ammonium polyacrylate (PAA) on aqueous R-Al2O3 suspensions over a wide pH range.

the system is under continuum agitation during the experiment, particles are expected to be suspended and, therefore, φ was set as the nominal value (2 vol %) for calculation purposes. Under the assumption of spherical particles with thin double layers, the zeta potential (ζ) can be calculated from µD(ω) according to O’Brien’s formula13

ζ ) µD(ω) η G(R)-1 -1

(2)

where η and  are the solvent viscosity and permittivity, respectively. The G(R) term corrects for the inertia of the particles in the alternating field. The ESA probe was calibrated using 10 vol % Ludox in water. This silica suspension has a precise zeta potential at -38 mV (ESA signal ) -5.32 mPa‚m‚V-1) at 25 °C and a typical pH of 9.1. pH was monitored to ensure no pH variations during calibration. Conductivity of the suspension was also monitored during calibration to ensure the expected 2.1 µS/cm. The silica particles have density 2.2 g/cm3 and radius 0.014 µm.

Experimental Procedures

Zeta-Sorption Method

Materials and Chemicals. A commercial aluminum oxide (Alcoa A1000-SG), with mean particle size 0.4 µm and BET specific surface area of 8.4 m2/g, was used in the adsorption tests. Citric acid (Labsynth Inc. P.A.) and ammonium poly(acrylic) acid (R. T. Vanderbilt Inc. Darvan 821-A) with an average molecular weight of 3500 g/mol were used as the dispersants. Alkaline or acidic adjustments required were made using standardized analytical-grade potassium hydroxide (KOH) or nitric acid (HNO3) solutions (Labsynth Inc.). Conventional Route for Adsorption Experiments. Alumina suspensions (5 vol %) containing varying dispersant concentrations were allowed to achieve adsorption equilibrium over 24 h. The samples were then centrifuged at 2800 rpm for 1 h. A cloudy supernatant was removed from each sample, and the ionic strength was increased by the addition of a salt solution (0.2 mg of NaCl/mL of solution). Thereafter, the obtained solutions were recentrifuged under the same conditions in an attempt to remove the remaining suspended particles, and a visual transparent supernatant was obtained. A 30 mL aliquot of supernatant was extracted from each sample, and the pH was adjusted to 9.3 before analysis by first derivative titration. The same ionic strength and pH adjustments were performed when considering the reference samples used for the calibration plot. HCl (0.1 M) was used for all the titration experiments. Infrared spectroscopy experiments were carried out in a Nicolet Magna 560 spectrophotometer in ATR mode (attenuated reflectance). Electrokinetic Property Measurements. The measurements of the electrokinetic sonic amplitude (ESA) signal 12 as a function of dispersant concentration were performed in a ESA8000, Matec Applied Sciences, Hopkinton, USA, using 2 vol % Al2O3 suspensions. A dispersant-free suspension was prepared for each experiment, and dispersant was automatically injected in suspension with a standardized rate of 0.02 mL/30 s. Signals were measured before each injection. Dispersant solutions of 5 vol % were used in all experiments. Any pH modification was carried out using KOH or HNO3 solutions. The used injection rate was considered suitable for achieving adsorption equilibrium between each injection after experiments showed that longer times between injections caused similar results. Experiments were carried out at 25 ( 0.2 °C. Conductivity and pH were constantly monitored during the experiments. At the used solid loading, the magnitude of the ESA signal is related to the dynamic electrophoretic mobility (µD) by

The principle of the method is based on a linear relationship between the dispersant concentration and the surface charges generated by the adsorption (which maybe evaluated by the zeta potential ζ), observed in suspensions with very low dispersant concentrations. This linear behavior represents a condition where all dispersant molecules are adsorbed on the particles surface, and the bulk dispersant concentration may be disregarded. This happens since the dispersant concentration is so low that the chance of the molecules interacting with each other and hence staying in the bulk is much lower than that of interacting with the surface. We assume that the dispersant has enough affinity with the surface to adsorb and that the adsorbed molecules do not interact with each other, such that each molecule generates the same charge. The effects of the conductivity in charge developments were disregarded to consider the proposed linear behavior; moreover, this linear behavior would only be expected at low charge densities, and deviations from linear behavior should exist at high charge densities. Despite the fact that these assumptions may appear severe, the results showed the approximation to be a reasonable simplification. Therefore, a simple linear equation for the first points was considered. When extrapolated to higher concentrations, this equation will simulate a surface charge generation behavior if the dispersant adsorption had no saturation limit, i.e., considering total adsorption efficiency during the whole process. However, this corresponds to an unrealistic situation since a plateau is always attained. Therefore, the difference between the zeta potential expected in a total adsorption situation and the measured one will give the dispersant quantity that stays in the bulk (CS) and, hence, will indirectly quantify the adsorbed amount (CA) as a function of the total dispersant concentration (CL). The application of this method entitled “zeta-sorption” however requires the pH to be unaltered during the experiment. This is because any pH shifting would account for changes on the surface potential and therefore would represent an additional contribution to the zeta potential. The linear equation approach on the “first” points of the zeta potential curve as a function of dispersant concentration needs to be standardized to establish a general procedure for the adsorption isotherm construction. The following procedure is proposed: (a) smoothing of the potential curve is performed to eliminate noise effects

ESA(ω) ) µD(ω) c ∆F φ Gf

(1)

where c is the sound velocity in the suspension, ∆F the density difference between the particles and the solvent, φ the volume fraction of the particles, Gf a geometrical factor for the electrode geometry, and ω the angular frequency of the applied field. Since (12) Hunter, R. J. Colloids Surf. 1998, 141, 37.

(13) O’Brien, R. W. J. Fluid Mech. 1988, 190, 71.

Adsorption Isotherms on Colloidal Suspensions

Figure 1. Zeta potential and pH of the suspension as a function of citric acid concentration. xreal is the measured dispersant concentration needed to generate yzeta potential. xline represents the dispersant concentration needed to generate yzeta considering total adsorption described by the line shown.

using adjacent averaging with 10 points. (b) A curve fitting [ξ ) f(CL)] is performed using a polynomial. A sixth grade polynomial was efficient in the proposed study, but due to the presence of minima and maxima values, other polynomials may be adequate depending on data. The interception with the y-axis is taken to be equal to the intrinsic zeta potential (ζintr) measured for dispersantfree suspensions. (c) A tangent line

y ) a1 x + a0 )

∂p(x) | x + p(0) ∂x x ) 0

at the point x ) 0 is adjusted and represents the 100% adsorption condition. Differences between the experimental determined values and that expected for 100% adsorption lower than 0.1 mV are discredited Method Evaluation The zeta-sorption method was first evaluated by studying an alumina-citric acid suspension. Citric acid is commonly known as an efficient dispersant for alumina, and since the previously mentioned problems involving precipitation and filtration should not be so pronounced due to the small size of the dispersant in relation to alumina particles, this system is considered suitable for evaluating the new adsorption measuring procedure. Figure 1 shows the zeta potential variation of an alumina suspension as a function of citric acid concentration. Since citric acid is an efficient electrostatic stabilizer, with specific adsorption and high affinity to the alumina surface, the molecules promote surface charges, increasing repulsion forces between the suspended particles. Figure 1 shows the expected linear behavior representing total adsorption at low citric acid concentrations (