You are on page 1of 27

Advances in Colloid and Interface Science 106 (2003) 5581

Review of the measurement of zeta potentials in concentrated aqueous suspensions using electroacoustics
R. Greenwood*
School of Chemical Engineering, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK

Abstract This paper reviews the technique of electroacoustics as it has been applied to aqueous suspensions of inorganic particles. It starts by charting the development of the technique from its earliest beginnings in the 1930s to the present day. The technique has become well established in the last decade with the advent of the Acoustosizer and the Acoustosizer II. Illustrations of how the technique can be used are given based on the authors own experience, especially the measurement of iso electric points, the adsorption of polyelectrolytes, the effect of ionic strength, the effect of powder surface area and the dissolution of material from powders. Some new data on (a) the adsorption of different molecular weights of polyacrylic acid onto alumina and (b) titanium dioxide suspensions are also presented. 2003 Elsevier B.V. All rights reserved.
Keywords: Electroacoustics; Zeta potential; Suspensions; Polyelectrolytes

1. Introduction Over the last twenty years electroacoustics has proved to be a powerful tool for the measurement of zeta potentials in concentrated aqueous suspensions. This paper first reviews the history behind the technique, and in particular, the application of the technique to aqueous suspensions of inorganic particles especially ceramic ones. The paper then reviews what type of experiments have been carried out using electroacoustics, using some previously published examples from the authors own work. Finally, some new data are presented on the effect of molecular weight of
*Tel.: q44-121-414-7234; fax: q44-121-414-3626. E-mail address: r.w.greenwood@bham.ac.uk (R. Greenwood). 0001-8686/03/$ - see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016/S0001-8686(03)00105-2

56

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

polyacrylic acid adsorbing onto alumina under different salt concentrations. For an excellent review of the theory behind the technique and the study of non-inorganic particles see Hunter w1x. Although the first reported reference to the phenomena of electroacoustics was approximately 70 years ago, it was only since OBrien w24x formulated his theories in the early 1990s that led to the development of commercial machines that the technique has taken off. There is an increasing trend within the ceramics industry towards finer and finer particle sizes, as this leads to better strengths in the final product, shorter sintering times and lower sintering temperatures. However, this reduction in particle size does have some associated problems. As the particle size range enters that of the colloidal size range, i.e. sub micron, the particles will aggregate together due to the attractive van der Waals force. If these aggregates persist during processing they can cause flaws in the final sintered product. Not only are these flaws potentially athsetically unappealing, they also weaken the strength of the product. From the Griffiths w5x equation the strength of a ceramic body is inversely proportional to the flaw size. If the size of these flaws can be minimised to that of an individual particle then the mechanical properties of the final product can be vastly increased. This technique is known as colloidal processing w6x and involves optimising the processing conditions for a powder from a fundamental understanding of the interparticle forces. As well as reducing the mechanical properties these flaws can also reduce the thermal, optical and electrical properties of the final product. The Acoustosizer and Acoustosizer II (Colloidal Dynamics USA) are especially suited for studying the colloidal processing of advanced ceramic materials. Electroacoustics can refer to two types of processes. Firstly, if a sound wave passes through a colloidal suspension it creates a macroscopic potential difference called the ultrasonic vibration potential or UVP effect. Secondly, the opposite effect occurs when an alternating electric field is applied to a colloidal suspension, this generates a sound wave, which is called the electrokinetic sonic amplitude or ESA. In both cases the applied field and the response occur at the same frequency. The first reported reference to these phenomena was by Debye w7x in 1933, who gave a theoretical treatment of the UVP effect when applied to an electrolyte solution in water, which he termed the Ionic Vibration Potential, IVP. As the sound wave passed through the electrolyte solution it produced a separation of charge due to differences in the effective masses and frictional coefficients of the solvated anions and cations. The resulting sum of these tiny dipoles leads to a macroscopic potential difference, which depends on the sound wave frequency. However, it was not until over a decade later, in which the experimental verification of the IVP was established by Yaeger et al. w8x. Meanwhile, it had been established that much larger effects could be described theoretically in colloidal suspensions, the so-called colloidal vibration potential, CVP w911x. This effect was initially over estimated by Enderby w12x, but improved calculations were later made by Booth and Enderby w13x. However, this was limited to the lower frequency range (<1 MHz) and such approaches have little modern value as todays machines work at much higher frequencies ranges of 0.110 MHz.

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

57

The early work on the IVP has been reviewed by Zana and Yeager w14,15x who also described the CVP and its relationship to proteins and polyelectrolytes. More recently, Marlow et al. w16x have reviewed the CVP and its extension to more concentrated suspensions. Since then there has been very little theoretical work on the CVP, but there has been extensive development of the ESA effect since its discovery in the early 1980s w17x. The ESA technique has transformed the characterisation of colloidal suspensions especially concentrated ones, because it allows simultaneous measurement of both the zeta potential and the particle size. Until the advent of this technique, the only methods for measuring zeta potentials were for dilute systems only, e.g. electrophoresis. Early work on the ESA only appeared in conference proceedings or in the manufacturers technical notes, so the early reviews were limited w1820x. The time taken from the first theoretical treatment of the UVP effect and its scientific exploitation reflects the technical difficulties involved. These were due to the fact that the devices used to generate the sound waves require the application of electric fields that interfere with the measurement of the resulting potentials. Additionally, measurement of the small electroacoustic signals required the development of low noise, high frequency amplifiers. Nevertheless, substantial advances were made and by the 1980s commercial CVP devices were in development. However, one group of engineers conceived the idea of reversing the procedure and listening for the sound wave instead. This they were able to develop quickly, building on the previous experience gained from the CVP devices, but also due to the inherent advantages of the ESA techniques over the alternative CVP. The resulting machine the ESA 8000 is still in commercial use, although it only operates at one frequency and requires the input of a particle size to calculate the zeta potential. The CVP is developed as the sound wave travels through the colloidal suspension, it moves the particles and their associated double layers in slightly different ways provided there is a density difference between the particles and the medium. The particle charge becomes slightly separated from its counter charge in solution, creating an array of dipoles, which increase and decrease in magnitude on the same length scale as the sound wave amplitude. It is the sum of these separate dipole fields that gives rise to the macroscopic potential, as long as the electrodes are not exactly one or more wavelengths apart. For a given sound wave intensity the maximum effect occurs when the electrodes are separated by an odd number of wavelengths. In the ESA effect the applied electric field causes the charged particles to oscillate backwards and forwards at the same frequency and this generates tiny acoustic dipoles associated with each particle (provided there is a density difference between the particle and the medium). These dipoles cancel one another throughout the body of the suspension except near the electrodes where the acoustic dipoles generate a sound wave, which can emerge from the suspension and move down the delay rod where it is detected by the transducer. The transducer then detects the amplitude and phase angle of this sound wave as a function of the applied frequency. The phase angle measures the time lag between the applied field and the subsequent particle motion. This is zero at low frequencies and increases with increasing frequency and increasing particle size.

58

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

The first theoretical treatment of the ESA effect was provided by OBrien w21x who showed that the CVP and the ESA were reciprocally linked. An analogy can be drawn with the link between electrophoresis and electro-osmosis established by Mazur and Overbeek w22x. The initial analysis was limited to dilute solutions, but was subsequently extended to systems of arbitrary concentration by OBrien w3x, provided that the particles are small compared to the sound wave length. A condition that is always satisfied for colloidal particles for frequencies less than 20 MHz. An experimental verification of this reciprocal effect for polyelectrolytes and electrolytes was provided by OBrien et al. in 1994 w23x. It is important to note the difference between the CVP and the ESA. The quality required is the dynamic mobility. The ESA technique gives this directly, whereas the alternative route requires not only the CVP, but also the complex conductivity. CVP devices only operate at one frequency and the conductivity measurement is performed only in the low frequency limit, which leads to an inherent error in estimating the dynamic mobility especially for low conductivity systems due to neglect of the imaginary part of the conductivity. The problem of relating particle mobility to the surface properties of a particle has been examined since the start of the twentieth century when von Smoluchowski solved the problem for d.c. electrophoresis in the special case where the double layer (ky1) is thin compared to that of the particle radius (a), see Hunter w24x. Smoluchowskis solution breaks down for ka-50 and large zeta potentials. The Acoustosizer operates in the frequency range 0.311 MHz allowing particle sizing from 0.1 to 10 mm. For smaller particles the zeta potential can be calculated, but the inertia forces are too small to allow sizing. The dynamic mobility spectrum is measured over a frequency range to give the mobility spectrum, from which the particle size and zeta potential can be obtained. If the real part of the dynamic mobility is plotted against the imaginary part then an Argand diagram can be obtained where each point represents a value of the dynamic mobility. The magnitude of the mobility is thus given by the distance from the origin and the phase lag is measured by theta. For the small particles the magnitude does not change greatly with frequency and the phase lag increases slightly. The larger particles, however, show a smaller magnitude at the lowest frequency and both the magnitude and phase angle change significantly with increasing frequency. To determine the size and charge, it is then a question of obtaining the best fit to the spectrum using the OBrien theory. The limiting phase lag is 458 and when the sign on the particle changes the mobility points move from one quadrant of the graph to the opposite one, i.e. 1351808. This 1808 phase change allows an unequivocal and accurate determination of the iso electric point. The initial device that was developed the Matec ESA 8000 could operate in either ESA or CVP mode, but only worked at one frequency (1 MHz). At the time theory did not exist for relating the ESA to the particle properties through the dynamic mobility. One could only measure the relative ESA signal and standardise its significance by calibrating with a colloidal sol whose properties were assumed to be constant, e.g. colloidal silica. However, it did work in aqueous and non-aqueous media and was accurate to approximately 2% and up to volume fractions of 0.05,

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

59

but as it only operates at one frequency it could not measure particle size. These limitations led to the development of the Acoustosizer. The Acoustosizer measures the magnitude and phase of the dynamic mobility spectrum of any particle concentration in the size range 0.110 mm. For the smallest measurable size the inertia effect is only significant at the highest frequencies, whilst for the largest particle size the amplitude at the lowest frequency can still be detected. Generally, the amplitude of the particle motion is extremely small. At 10 MHz a 1 mm particle in a field of 10 V y cm moves a few picometres. The Acoustosizer cell is manufactured from highly chemical resistant epoxy resin with a capacity of 400 ml. The contents of the cell can be agitated with a propeller mixer. Three probes are inserted into the cell to measure the temperature, conductivity and pH and four thermally conducting ceramic rods are immersed into the suspension to provide temperature adjustment, by both positive (Joule heating) and negative (Peltier effect). All of which are monitored and recorded during a measurement. Calibration is with an electrolyte. All electrolytes give an ESA signal particularly if the cation and anion differ significantly in mass and y or frictional drag i.e. radius. OBrien et al. w25x have chosen the potassium salt of alpha dodecatungstosilicic acid as a standard calibration. The octadecahydrate (K2wSiW12O40x18H2O) is an effective standard as long as it is pure. The Acoustosizer II has been developed for both laboratory and pilot plant environments with the capacity for in stream monitoring. Particle sizing is carried out by a combination of electroacoustics and ultrasonic attenuation covering the size range from 0.02 to 10 mm. Again the pH, conductivity and temperature are recorded although the sample volume can be as small as 20 ml and up to 150 ml. Instead of stirring the colloidal suspension it is transported by a peristaltic pump, which means the maximum working viscosity is 0.25 Pas. Another improvement is the ability to work at a much higher conductivity, i.e. 5 S y m. Recently, another commercial device for measuring zeta potentials in concentrated suspension has been launched onto the market by Colloidal Dispersions, Mount Kisco, New York, USA. This calculates the zeta potential from the colloid vibration current (CVI), see Dukhin and co workers w19,2628x. Much the same set of experiments has been carried out with this device as the ESA8000 y Acoustosizer, e.g. optimal dispersant dosage to prevent aggregation w29x and measurement of iso electric points w30x. However, here only work carried out by the Acoustosizer and its predecessor will be discussed. 2. Literature review 2.1. ESA 8000 Although the ESA8000 was extremely limited in its ability to measure the absolute mobilities and zeta potential it proved to be of great use in moderately concentrated (5% v y v) suspensions of minerals in aqueous media. Early results were to correlate ESA measurements with electrophoresis w3,16,18,31,32x. Many early papers on ESA measurements, however, failed to take into account the particle

60

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

size distribution of the material resulting in some strange results with zeta potentials as high as several 100 mV! Scales and Jones w33x demonstrated the importance of particle size and polydispersity for adequate correlation of electroacoustic data to that obtained by electrophoresis. The earliest publications investigated; cetyl pyridinium chloride adsorption on kaolinite w34x and its charge reversal by hydrolysable metal ions w35x and the charge reversal of silica w36x. In 1992, the National Institute of Standards and Technology (N.I.S.T.) published the proceedings of a special symposium on the ESA technique w37x. This covered a wide range of topics including; colloidal diamond particles, corrections for background electrolyte effects in mineral oxide systems, effect of pH on silicon nitride and yttrium oxide, adsorption of polyacrylate on calcium carbonate, effect of pH on gamma alumina and coal, effect of pH on non-ionic polyacrylamide treated hematite and correlation of the zeta potential with viscosity measurements, cement in the presence of plasticisers, rotogravure inks and applications to phosphors and toners. Other particles studied include ZnO and ZnS w38x, mullite particles mixed with a latex binder w39x, silicon carbide platelets mixed with alumina w40x and pigment particles w41x. The technique has also been applied to the paper making process w4244x. Paik et al. w45x reported that the particle surface charge of barium titanate was influenced by the solids concentration of the suspension causing a shift in the i.e.p. These papers reflect the broad spectrum of applications the device can be used for rather than any development of theory. The ESA technique is also able to cope with unusual particle geometries, e.g. Texter w46x studied the adsorption of sodium oleoylmethyltaurine onto a ceramic powder of parallelepiped particle morphology and obtained fair agreement with adsorption isotherms for monolayer coverage. Numerous researchers have also concentrated on the processing of that important ceramic material; silicon nitride w4752x. Paik et al. w53x studied the adsorption of a PMMA dispersant and a polyvinylalcohol binder onto silicon nitride. The PMMA caused the effective i.e.p. of the system to be moved to lower pH, whereas the PVA had no shifting effect but did reduce the magnitude of the zeta potential. The ESA technique was also used by Walldal w54x to study the flocculation between two cationic polyelectrolytes and negatively charged silica particles. Polyacrylamide was more efficient than polyamide in neutralising the charge on the particles. A similar study was carried out by Baltar and Oliveira w55x using polyacrylamide as a flocculant for silica. The ESA technique was used by Smith and Haber w56x to study structure formation of alumina suspensions in terms of its filtration behaviour. A well-dispersed suspension cast slower than a flocculated one at the same volume fraction. One type of colloidal particle missing from the above list appears to be polymer latex particles, which do not appear to have been investigated as widely as inorganic particles. Shubin et al. w57x studied carboxylate latex using d.c. electrophoresis, ESA measurements and dielectric relaxation behaviour. They were able to rationalise their data by introducing the concept of anomalous surface conduction by which ions in the Stern layer are able to contribute to the electrical conductivity. Klingbiel et al. w58x investigated polymethylmethacrylate (PMMA) lattices, whilst James et

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

61

al. w31x investigated polystyrene and PMMA lattices amongst other particles. OBrien w2x verified his electroacoustic theory using a monodisperse latex and a cobalt phosphate sol. However, beyond these few isolated examples, latex particles have been massively under utilised in terms of electroacoustic experiments. 2.2. Acoustosizer The earliest studies with an Acoustosizer were also published by the N.I.S.T. w59x and initial ceramic y minerals work was described by OBrien w60x. The latter paper describes how the sample is titrated over a pH range to determine the iso electric point, an important characteristic of many colloidal systems (see experimental section). Stable oxides show a reversible titration curve with respect to the zeta potential. The size of the particles is a minimum at pH values away from the i.e.p. as they are well dispersed. Near the i.e.p. the size appears larger, but this is to some extent an artefact. As a system passes through its i.e.p. the system flocculates and the particle size is undefined. The size of the aggregates depends on the time scale and stirring speed. In 1995 OBrien w61x developed a formula to calculate the dynamic mobility of a porous particle, which is relevant to studying flocculated systems. As with the ESA technique there are numerous studies comparing electroacoustic measurements with other techniques, e.g. Pettersson et al. w62x investigated the adsorption of polyacrylic acid and polymethacrylic acid onto alumina, zirconia and yttria doped zirconia using a combination of adsorption isotherms, particle size measurements and adsorbed layer thickness measurements. Beattie and Djerdjev w63x compared electroacoustic results from alumina suspensions in the presence of polymethyacrylate dispersants with viscosity measurements and sedimentation volumes. Electroacoustic measurements have also been compared with electrophoresis and streaming potential experiments by Knosche et al. w64x in the case of silica, titania and alumina powders and by Rasmusson and Wall w65x for alumina modified silica nanoparticles. Another powder that has been investigated by electroacoustics is titanium nitride. Shih et al. w66x studied the adsorption of polymethacrylic acid at different pH values with rheology and zeta potential measurements. The adsorbed amount increased as the pH decreased, an effect reported by numerous researchers w6769x. OBrien and Rasmusson w70x measured the zeta potential and size of bentonite. It is normally very difficult to obtain reliable information on bentonite due to its unusual shape and viscosity characteristics, but it behaved well in electroacoustic analysis. Another mineral system of great interest is the metal sulphides. Prestidge and Rowlands w71x compared electrophoretic and electroacoustic measurements of the zeta potential for zinc sulphide and lead sulphide. These systems are complicated due to the possibility of surface oxidation, which produces time effects, but can also depend significantly on the surface area to volume ratio. It is also possible to study semi conducting particles, e.g. silicon particles w72x by making a slight modification to the theory. Another area of potential danger is in studying materials that display anomalous surface conduction, in particular kaolinite. Surface conduc-

62

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

tion can lead to erroneous sizing, which can be detected in the phase lag data. Instead of increasing from low values to a limiting value of y458, it may go through a maximum at low frequency and even turn into a phase lead over part of the range, see Subhin et al. w57x and OBrien and Rowlands w73,74x. The Acoustosizer is in principle capable of working at elevated salt concentrations, but the factory calibration only covers the range from 0.01 to 1 S y m i.e. in the order of 0.5 mM to 0.1 M salt. To extend the range requires standards of known ESA and higher conductivity. Caesium chloride solutions w75x are suitable for this purpose and extend the range to approximately 20 S y m, which is the conductivity of 3 M NaCl. Early work at high salt concentrations was carried out by Kosmulski and Rosenholm w76x, but unfortunately they did not know the limitations of the machine, so their results under estimate the zeta potential w77,78x. Rowlands et al. w75x studied the effect of high salt concentrations on the behaviour of a suspension of aluminium hydroxide. The i.e.p. of the material shifted from 9.1 for dilute salt concentrations, but moved to over 11 for 0.5 M NaCl and showed no i.e.p. at all for 3 M NaCl. Thus, what is normally considered an indifferent electrolyte shows a distinct tendency to specific adsorption at high enough concentrations, something that has always been considered possible in the current descriptions of the oxide solution interface. They also came up with the idea of determining the i.e.p. from the Argand diagram of real and imaginary components to the dynamic mobility. Johnson et al. w79,80x published two papers on the binding of monovalent electrolyte ions on alpha alumina at high electrolyte strengths using the techniques of yield stress rheology measurements and electroacoustics. At high (approx. 1 mol dmy3) electrolyte concentration, the monovalent cations bind to the negative alumina y in the order Liq)Naq)KqfCsq. By contrast Bry, Cly, Iy and NO3 anions adsorb to an almost identical extent over the range of concentrations and pH conditions investigated. The cation binding sequence was consistent with the structure making y structure breaking model proposed in the 1960s w81,82x. Similar work on alumina and zirconia has been carried out by Franks w83,84x and coy y y y workers investigating IOy 3 , BrO3 . Cl , NO3 , ClO4 ions. The adsorption of other electrolytes namely cobalt chloride, molybdenum chloride and ammonium heptamolybdate onto silica and alumina has been investigated by Deboer et al. w85x. The importance of correcting for the background electrolyte signal has been demonstrated by Desai et al. w86x. As mentioned previously for a solid in suspension the measured ESA signal is a combination of that due to the particles and that of the background electrolyte. Under certain circumstances this background must be subtracted. They noted that the i.e.p. of two silica suspensions at an ionic strength of 0.1 M was 4.7 and 5.6, but with the background correction the particles did not change sign until at least a pH of 2. One of the most common substances for dispersing colloidal samples is the sodium salt of polyacrylic acid. This adsorbs very strongly on positively charged surfaces via electrostatic forces and will adsorb on poorly charged negative surfaces through hydrogen bonding and van der Waals forces. Fortunately, polyacrylic acid adsorbs so strongly on most surfaces that it lies almost flat, with negligible loops

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

63

and tails. As a result, the zeta potential is strongly affected, which can be easily detected by electroacoustic measurements. Djerdjev w87x studied the adsorption of polyacrylic acid onto alumina over a range of pH values and obtained a threedimension surface of how the zeta potential varies as a function of the coverage and the pH. The experiment was carried out by addition of successive amounts of polyacrylic acid at the end of each titration run. The whole process taking only a few hours. Kaolinite, silicon nitride and calcium carbonate all respond well to polyacrylic acid as a dispersing agent w88x. Kaolin has also been investigated, alongside alumina, by Johnson and Scales w80,89x using both electroacoustics and rheology. Polysulphonic acid seems to behave in a similar manner and is used to disperse colloidal materials. Ukigai et al. w90x used electroacoustics to study the adsorption of polysulphonates on coal water slurries at volume fractions of 65%. They were able to identify the molecular weight of polysulphonate that generated a maximum in the zeta potential and a minimum in the particle size. A low molecular weight polycarboxylate dispersant has been investigated by Costa et al. w91x. Pradip et al. w92x investigated the dispersion of three different ceramic powders, i.e. alumina, zirconia and silicon nitride using two different polyelectrolytes; one cationic and one anionic. All the powders could be dispersed when the working pH was at least two pH units away from the effective i.e.p. of the system in the presence of the dispersant. Premachandran and Malghan w93x, however, recommend at least 2.5 to 3 pH units difference in their study of silicon nitride, silicon carbide, alumina, aluminium nitride and yttria powders. They also recommend cationic dispersants over anionic ones as they produce suspensions with higher stability, with a narrower particle size distribution over a wider range of pH values. As previously mentioned, one of the most commonly investigated powders is silicon nitride. Hackley w94x used the ESA 8000 and the Acoustosizer to study the adsorption of polyacrylic acid onto silicon nitride. In an earlier paper, Hackley and co-workers w95x used soxhlet extraction to clean five silicon nitride powders and obtained a pristine i.e.p. of 9.7 for unoxidised silicon nitride. In another paper by the same group of workers Hackley and Malghan w47x noted that fluoride was specifically adsorbed onto the surface of silicon nitride powders. Galassi et al. w96x noted that although the milling of a silicon nitride powder did not alter the i.e.p. it did alter the magnitude of the zeta potential. A follow-up paper w97x on the addition of sintering aids noted that they dramatically altered the surface chemistry and the i.e.p. Silicon nitride has also been studied by Laarz and Bergstrom w98,99x using electroacoustics in conjunction with rheological measurements. They compared copolymers of PEO and methacrylic acid with polyacrylic acid and demonstrated that the grafted PEO chain had a minor influence on the colloidal stability, whereas the polyacrylic acid extended perpendicular to the surface. They also attributed the effect of excess polyelectrolyte to the increased suspension viscosity due to increased ionic strength caused by the release of associated counter ions of the polymer functional groups. Titania powders have been investigated by Greenwood and Kendall w100x they noted that due to an organic coating on one of the titania powders the adsorbed

64

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

amounts were very low (in the order of micrograms per m2). This organic coating could be removed by calcining the powder, which then lead to increased adsorbed amounts. Roncari et al. w101x investigated a magnesia doped mullite suspension and noted that the magnesia undergoes pronounced solubilization that strongly affects the suspension properties. The dissolution of material from powders has also been noted for doped zirconia powders i.e. cerium w102x and yttrium w100,103x. This is further discussed in Section 3.7. Maier et al. w104x made an early attempt to measure the adsorption of polyethylene oxide onto silica dispersions. They were able to relate the drop in zeta potential to a shift in the position of the plane of shear, as had been suggested by Koopal and Lyklema w105x. Carasso and OBrien w106,107x have also investigated this effect and developed a model which can distinguish effects on the mobility due to the fluid motion within the polymer layer itself and also those due to a shift in the plane of shear. The model allows an estimation of how much the shear plane has been shifted by the adsorbed layer i.e. the adsorbed layer thickness. A PVA sample of molecular weight 50 000 Daltons reduced the zeta potential by approximately 30%. The same results were obtained for a series of nonylphenylethoxylates of varying molecular weights. The adsorbed layer becoming thicker with increasing concentration and molecular weight. Similar results were obtained by Miller and Berg w108x studying the adsorption of PVA onto titania. Thus, it has been shown that the Acoustosizer has been used alongside numerous other techniques to study concentrated suspensions of ceramic particles. The most commonly studied subjects are the iso electric point of the powder and the adsorption of polyelectrolyte dispersants onto these powders. The technique is robust enough that even the dispersion of cement can be studied, e.g. Hodne and Saasen w109x related the consistency of cement slurries to the zeta potential, despite the nature of the material and its high pH)12. In the next section, a review of some of the experiments that have been carried out by the author using electroacoustics to study the manipulation of the interparticle forces is presented. The exact experimental details if not reported are given in the references. 3. Examples of electroacoustic experiments 3.1. Measurement of iso electric points (i.e.p.) For purely electrostatically stabilised systems it is important to know the exact location of the i.e.p., i.e. the pH value at which the particles have zero zeta potential. At this pH value there are no repulsive forces whatsoever and the particles will be heavily flocculated due to the dominance of the attractive van der Waals forces. Normally, this is an undesirable state of affairs and systems are designed such that the suspension pH is well away from the i.e.p. The further away the suspension pH from the i.e.p. the greater the surface charge and hence the greater the zeta potential. The magnitude of the repulsive force between particles is directly related to the magnitude of the zeta potential. As a rough rule of thumb working approximately 23 pH units away from the i.e.p. in either direction is sufficient to generate a large

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

65

Fig. 1. Effect of NaCl concentration on the iso electric point of SDK160 alumina. DiamondssDouble distilled water, Triangless1 mm NaCl, Circless50 mm NaCl, Open symbolssincreasing pH, Closed symbolssdecreasing pH.

enough zeta potential for stabilisation. A zeta potential of greater than "30 mV is considered sufficient for stabilisation and in fact this usually occurs approximately 23 pH units away from the i.e.p. so the rough rule has some basis. However, it is also well known that electrostatic stabilisation is insufficient to stabilise particles at high volume fractions due to extensive double layer overlap, it is therefore necessary to adsorb polymers or polyelectrolytes onto the particles to stabilise the particles via steric or electrosteric mechanisms, respectively. Electrosteric stabilisation describes the combined stabilisation mechanisms of electrostatic and steric stabilisation. Figs. 1 and 2 show typical examples of the effect of altering the pH of an alumina suspension (SDK 160, volume fraction 0.11, stirrer speed 300 rpm) at different

Fig. 2. Effect of KCl concentration on the iso electric point of SDK160 alumina. esDouble distilled water, ns1 mm KCl, ss50 mm KCl, Open symbolssdecreasing pH, Closed symbolssincreasing pH.

66

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

electrolyte strengths. Fig. 1 demonstrates the effect of varying the background electrolyte strength from that of double distilled water to 1 and 50 mM NaCl, whilst Fig. 2 demonstrates the same effect but for KCl. At low pH the suspensions are positively charged, whilst at high pH the suspensions are negatively charged. The i.e.p. can be identified from the pH value at which the zeta potential is zero. Fig. 1 shows that the i.e.p. with no salt present is approximately 7.4, which is shifted to 7.9 and 8.2 when the background electrolyte is changed to 1 mM NaCl and 50 mM NaCl, respectively. The same effect is seen for the KCl, the i.e.p. is shifted from 7.4 to 7.9 and 8.1, respectively. This shift in i.e.p. is due to specific adsorption of cations onto the alumina particles. If the electrolyte was inert it would be expected that the i.e.p. would be independent of the salt concentration. The same effect has been noted by Burke w110x for a Sumitomo alumina powder. The type of acid y base that is used to alter the pH of a suspension can also have a major effect, Greenwood and Kendall w111x noted that the i.e.p. of an alumina powder was independent of the base used, but the degree of hystersis varied depending on the polarisability of the cation. This can also be seen in Figs. 1 and 2 in that NaCl gives a slightly larger change than that obtained with KCl, plus the degree of hystersis is non existent for the suspensions prepared with double distilled water, but does occur for the higher salt concentrations. There is normally a wide range of reported values for the i.e.p. for any powder as the surface chemistry will vary according to how it has been manufactured. This alumina powder will be electrostatically stabilised if dispersed at pH values greater than 11 and less than 5. 3.2. Selecting a dispersant and its optimum amount Traditionally in the ceramics industry, polyelectrolytes have been utilised to prevent flocculation of particles. Due to the charged nature of the polyelectrolyte they impart stability to the particles via an electrosteric mechanism. Hence, adsorption of these charged molecules onto a particle surface will alter the surface charge and hence the zeta potential. So using electroacoustics, it is possible to follow the changes in the zeta potential with increasing amounts of polyelectrolyte. This is extremely useful in determining the optimum amount of polyelectrolyte required to stabilise the particles under different conditions. If too little is added then some flocs will persist in suspension and if too much is added then destabilisation may occur. A polyelectrolyte behaves in a similar manner to an ordinary electrolyte, the effects of which are well known, i.e. addition of too much salt causes flocculation. Hence, it is crucial from a colloid stability argument to have the correct amount of dispersant present in the system. Additionally, there is an economic argument to overdosing especially when working on the large scale, e.g. grinding of minerals w112x. As previously mentioned various methods exist for establishing the optimum amount of dispersants. These include rheology, adsorption isotherms and sedimentation. The latter two are time consuming and tedious. Electroacoustics is, however, a quick and simple technique to allow the study of numerous dispersants onto a suspension and then selection of the most suitable one for the system in question

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

67

Fig. 3. Adsorption of cationic and anionic dispersants onto a silicon nitride powder.

and to determine its optimum amount. A suitable dispersant is one that imparts a large zeta potential to the particles when only a small amount is added w113,114x. The optimum amount of dispersant as determined by electroacoustics, adsorption isotherms and rheology experiments show excellent agreement w102x. Burke et al. w109x also demonstrated that electroacoustic results confirm optimum dispersant dosages as measured by sedimentation heights and particle sizing. However, as mentioned previously polyelectrolytes impart stability by electrosteric means so there will be a steric contribution to the stabilisation, which will not be reflected in the zeta potential measurements, but would be detected by atomic force microscopy or the surface force apparatus. Fig. 3 shows the adsorption of a cationic polyelectrolyte (6569 Ciba Chemicals, Bradford, UK) and an anionic polyelectrolyte (Dispex N40, Ciba Chemicals, Bradford, UK) onto a silicon nitride powder. The volume fraction in both cases was 0.12. The initial zeta potential of the suspension before addition of the anionic polyelectrolyte was y6.4 mV at a pH of 5.96 and a conductivity of 0.298 S y m. These zeta potentials are too low for electrostatic stabilisation to be effective; therefore, the particles are expected to be flocculated. The initial zeta potential of the suspension before addition of the cationic polyelectrolyte was y4.3 mV at a pH of 5.93 and a conductivity of 0.312 S y m, so in reasonable agreement with one another. Both curves show the same trends in that as the polyelectrolyte adsorbs onto the particle the zeta potential increases in magnitude. At a certain concentration of dispersant the zeta potential no longer increases as the particles are now completely covered in dispersant. This then is taken to be the optimum amount of dispersant required to stabilise the powder. Any excess dispersant added after this point does not affect the zeta potential, as it remains unadsorbed in solution; hence the zeta potential remains as a plateau. From Fig. 3, the plateau region begins at 3 mg y g for the cationic polyelectrolyte and 1.8 mg y g for the anionic dispersant. The plateau region is not exactly level because as more polyelectrolyte is added to the system it behaves like a simple salt, i.e. it reduces the zeta potential, so as more

68

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

Fig. 4. Effect of five commercial dispersants on Alcoa A16 powder. Background electrolyte strength 10 mM KCl. ssDolapix CA, nsDolapix PC21, hsDolapix PC67, Open esDolapix PC33 and closed diamondssDolapix CE64.

polyelectrolyte is unadsorbed in solution the zeta potential will gradually decrease. Unadsorbed polyelectrolyte in solution may potentially cause destabilisation of the particles via depletion flocculation. Fig. 4 shows how the zeta potential of an alumina suspension (Alcoa A16, background electrolyte 10 mM KCl) can be altered by the addition of five commercially available polyelectrolytes. In this case, the dispersants utilised were from Zschimmer and Schwarz. Initially, the zeta potential of the suspension was approximately y29 mV. It can be clearly seen that addition of polyelectrolytes increases the zeta potential of the suspension such that the suspension becomes more stable. The trend for all five dispersants is very similar in that initially the zeta potential changes strongly with small amounts of dispersant and then after a certain concentration the zeta potential begins to plateau out as no more dispersant is adsorbed on the surface. However, each different dispersant imparts a different final zeta potential and requires a different amount of dispersant to cover the particles. The effect of dilution of the dispersants has been taken into account so the dispersants can be compared. Fig. 4 shows that three dispersants plateau out at a dispersant concentration of 0.2 mg y m2, i.e. Dolapix CA, Dolapix PC21 and Dolapix PC67. Of these Dolapix CA imparts the greatest final zeta potential so this would be an excellent dispersant for the alumina. However, it must be noted that the likely stabilisation mechanism for polyelectrolyte dispersants is electrosteric stabilisation, therefore, there may well be a steric contribution to the stabilisation mechanism depending on how the dispersant adsorbs. Dolapix CE 64 imparts a final zeta potential of y45 mV and requires approximately twice as much dispersant to do so, making it a poor candidate in comparision. The zeta potentials from the suspension stabilised with Dolapix PC 33 do not appear to level out and further data points would be required to determine the optimum amount.

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581 Table 1 Effect of volume fraction on the optimum adsorbed amount Dispersant Tiron Volume fraction 0.111 0.214 0.269 0.360 0.111 0.219 0.298 0.359 0.111 0.221 0.296 0.358 Optimum amount of dispersant (mgym2) 0.14 0.13 0.13 0.14 0.19 0.22 0.23 0.22 0.28 0.28 0.27 0.29

69

Final zeta potential (mV) y53.0 y48.5 y52.0 y50.5 y53.0 y55.5 y59.5 y55.5 y51.0 y52.0 y51.0 y54.0

Duramax D3021

Aluminon

3.3. Effect of volume fraction on the zeta potential, the adsorbed amount and the i.e.p. In the previous section, it was shown how electroacoustics could be used to determine the optimum amount of dispersant for a given powder. However, it is important to know that these are independent of the volume fraction at which the experiment is carried out. Table 1 shows the optimum amount of three commercially available dispersants required to cover an alumina powder at four different volume fractions. It can be seen that within error the adsorbed amounts are all similar indicating that the adsorbed amount was independent of the volume fraction at which the experiment was carried out. In the case of expensive powders, it is then prudent to use the lowest volume fraction that generates a strong electroacoustic signal. In this case, the powder studied was Alcoa A16SG and the dispersants were Aluminon (Fluka) Tiron (Fluka) and Duramax D3021 and the background electrolyte was 10 mM KCl. As all the final zeta potentials are large and negative this indicates that all three dispersants are suitable to stabilise the powder against flocculation. Fig. 5 shows another plot of zeta potential against pH for an Alcoa A16 powder suspended in 10 mM NaCl. The experiment is repeated at three different volume fractions using 1 M HCl and 1 M NaOH to adjust the pH. The i.e.p. for all three powders occurred at a pH of 8.2 and was independent of the volume fraction. The zeta potential of a suspension should be independent of the volume fraction (provided there is no substantial electrical double layer overlap) at which the experiment is carried out. To investigate if this is true, a series of titania suspensions was prepared ranging from 12.2 to 50% by weight in double distilled water. The suspensions were allowed to stand for 15 min before being placed into the Acoustosizer II. Five measurements were taken and the mean average and standard deviation was calculated (see Table 2). The results are shown in Fig. 6a and b. Fig.

70

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

Fig. 5. Effect of volume fraction on the iso electric point of Alcoa A16 alumina. Background electrolyte strength 10 mM NaCl. es0.142, hs0.186 and ss0.237.

6a clearly shows a decrease in the zeta potential with increasing weight fraction, however, this can be easily explained by the corresponding increase in conductivity and increase in pH seen in Fig. 6b. Therefore, in order to truly test the effect of volume fraction the pH and ionic conductivity must be the same for all samples. This could be obtained by centrifuging down the most concentrated sample and using the supernatant to disperse the remaining powder. 3.4. Effect of dispersant on the iso electric point The degree to which polyelectrolytes are dissociated depends strongly on the pH. For example, polyacrylic acid is fully dissociated at approximately pH 9, so at lower pH values there are fewer dissociated groups. The degree of dissociation controls the configuration of the polyelectrolyte in solution and hence, how it adsorbs onto a particle surface. In this experiment a titanium dioxide (Tioxide) suspension at 20% by weight was prepared in double distilled water and the optimum amount of Dolapix PC33 added as a dispersant. The zeta potential of the suspension was then recorded as a function of the amount of the pH. This was compared with the iso electric point of the titanium dioxide suspension without the dispersant present. The results in Fig. 7 shows how the addition of the dispersant
Table 2 Average mean values of pH, conductivity and zeta potential for different weight fractions of titania Weight fraction 12.2 18.2 25.0 32.0 38.5 44.8 50.0 pH 7.15"0.05 7.30"0.02 7.30"0.01 7.43"0.01 7.46"0.02 7.49"0.02 7.48"0.03 Conductivity Sym 0.0141"0.0003 0.0197"0.0005 0.0190"0.0004 0.0321"0.0008 0.0387"0.0007 0.0479"0.0010 0.0547"0.0018 Zeta potential mV 10.9"1.0 10.0"0.2 6.9"0.1 6.2"0.1 5.2"0.1 4.0"0.1 3.0"0.1

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

71

Fig. 6. (a) Effect of weight fraction on the zeta potential of a titanium dioxide suspension. (b) Corresponding conductivities and pH values as a function of weight fraction.

moves the effective i.e.p. from approximately 7.8 to 5.9 when the particles are completely covered in dispersant. 3.5. Effect of powder surface area on the optimum adsorbed amount Another factor that may effect the adsorbed amount of polyelectrolyte is the size of the particles. Unfortunately in the processing of ceramic powders, it is extremely

Fig. 7. Effect of adding a dispersant on the iso electric point of a titanium dioxide suspension.

72

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

Fig. 8. Effect of salt concentration on the addition of a cationic dispersant. Reprinted from Powder Technology, 113, Greenwood and Kendall, Effect of ionic strength on the adsorption of cationic polyelectrolytes onto alumina studied using electroacoustic measurements.

difficult to obtain monodisperse particles, however it is straight-forward to obtain powders of the same purity, but with different surface areas. In a previously published paper Greenwood and Kendall w110x investigated the optimum amount of Darvan 821A required to cover a series of Sumitomo alumina powders of varying surface area. All the powders were 99.9% pure alumina. Within experimental error, it was seen that the adsorbed amount of dispersant was independent of the surface area of the powders i.e. 0.59 mg y m2, in other words the adsorbed amount was independent of the powder surface area. This agrees with the results of Kayes and Rawlins w115x, Baker et al. w116x, Luckham and Faers, w117x who all investigated the adsorption of block copolymers onto different sized polystyrene lattices. It is known that the conformation of polymers changes as the particle size changes at constant adsorbed amount, i.e. thicker layers are found on larger particles w118 121x. 3.6. Effect of ionic strength on adsorption of dispersants Another factor that affects the adsorption of polyelectrolytes from solution onto ceramic particles is the ionic strength of the suspension, see Greenwood and Kendall, w111x. Fig. 8 shows the adsorption of a cationic polyelectrolyte called polyDADMAC (PC20 HV) onto an alumina powder as a function of different background electrolyte strengths. The three curves follow the same trend in that the particles are initially negatively charged and the addition of the polyelectrolyte reduces the magnitude of the zeta potential. At a certain concentration zero zeta potential is achieved, but flocculation does not occur due to the steric component of the stabilisation. Addition of further polyelectrolyte now causes charge reversal until a point is reached where no more polyelectrolyte adsorbs and the plateau region is reached. In double distilled water, the polyelectrolyte has an extended configuration and when it adsorbs onto a particle it does so with the conformation of a train, i.e. it

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

73

Fig. 9. Dissolution of material from two zirconia powders. Reprinted from two papers in (a) Journal of the European Ceramic Society, 19, Greenwood and Kendall, Selection of suitable dispersants for aqueous suspensions of zirconia and titania powders using electroacoustics. (b) Journal of the European Ceramic Society, 20, Greenwood and Kendall, Acoustophoretic studies of aqueous suspensions of alumina and 8 mol.% yttria stabilised zirconia powders.

runs along the surface. Hence, very little adsorbs onto the particles and complete coverage is obtained at 0.10 mg y m2. However, increasing the ionic strength to 10 mM such that the charges on the polyelectrolyte are screened slightly (as well as the particle charge). Thus, the polyelectrolytes configuration in solution is now more coiled and it adsorbs on a particle with more loops and tails giving a thicker layer and a four-fold increase in the adsorption per unit surface area, i.e. approximately 0.40 mg y m2. Increasing the salt concentration further to 50 mM now causes the coil to collapse further and the adsorbed amount becomes lower (0.20 mg y m2), until eventually a critical salt concentration will be reached at which no adsorption occurs. The magnitude of the zeta potential on complete dispersant adsorption is constant with screening by the salt, i.e. the zeta potential falls from q50 mV to q25 mV to q20 mV as the salt concentration increases. A trend that is also reflected in the initial zeta potentials. The thickness of the layer is important in terms of stabilisation as it determines how close two particles can approach one another. The effect of ionic strength on anionic dispersants will be discussed in Section 3.8. 3.7. Dissolution effects Insoluble powders such as alumina and titania when dispersed in water show a constant zeta potential with time. However, other powders may demonstrate a variation in zeta potential with time. Thus, it is important, as one of the initial tests on a powder, to check whether there is some dissolution of material from the powder that may effect the surface chemistry of the particles and hence the zeta potential. This involves measuring the zeta potential every 15 min or so over a suitable period of 46 h. Fig. 9 shows the increase in zeta potential with time for two yttria doped zirconia powders w100,103x. Both powders reveal the same trend;

74

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

Table 3 Physical properties of the dispersants Code PAA 1 PAA 2 PAA 3 PAA 4 PAA 5 PAA 6 PAA 7 PMAA PMAA PMAA PMAA
*

Description Darvan 821A* Alcosperse EN Dispex A40* ASCN I Polyscience I DP2 7621 Polyscience II Aldrich I Aldrich II Darvan 7 Darvan C*

Molecular weight (Daltons) 3500 4000 10 000 21 000 60 000 99 200 225 000 6500 9500 1016 000 1016 000

Activity (%) 40 26.1 40 24.4 35 26.4 20 30 30 25 25

1 2 3 4

Ammonium salt. Rest sodium salts.

they are initially positively charged and the zeta potential becomes more positive over time. This increase with time may be attributed to the dissolution of yttria from the powder readsorbing onto the powder surface. The rate of increase for the 8 mol.% yttria doped powder was greater than the 3 mol.% one, because there is more yttria present to dissolve out. Shojai et al. w123x also reported on the dissolution of yttria from an yttria doped zirconia under acidic conditions. They concluded that the major species leached out in acidic aqueous media was Y3q. However, it was not clear if the equilibrium limit of the dissolved ions was reached or if the ions were readsorbed onto the surface. Similar results have been noted for spinel w122x powders, but the increase in zeta potential was this time attributed to sodium ions adsorbing onto the powder surface. 3.8. Molecular weight of the dispersant Another factor that can be investigated by electroacoustics is the molecular weight of the polyelectrolyte. In this work a series of polyacrylic acids and polymethacrylic acids were adsorbed onto an alumina powder (SDK160) at three different electrolyte concentrations namely no added salt, 1 mM and 5 mM KCl. The characteristics of the polyelectrolytes are given in Table 3. As can be seen from the table the dispersants are supplied as diluted down aqueous solutions of polyelectrolytes, so it is vital to allow for this dilution factor when comparing results. So all results are quoted in terms of mg of active dispersant per unit surface area of powder. All water used was double distilled and the experiments were carried out at 25 8C. 400 ml of suspension was prepared at a volume fraction of 0.086 (150 g in 400 ml of water). Due to the viscosity of some of the higher molecular weight dispersants these were diluted down further with a known amount of distilled water. The other dispersants were added as supplied using the automatic titration software. The optimum adsorbed amount was recorded from the usual shape of the curve and

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

75

Fig. 10. Optimum amount of adsorbed polyacrylic acid as a function of molecular weight. Diamondss double distilled water, Squaress1 mM KCl and triangless5 mM KCl.

plotted against the molecular weight of the polyacrylic acid for the three different situations. The results are shown in Fig. 10 for the polyacrylic acids and in Table 7 for the polymethacrylic acid. Table 4 to Table 6 show the initial characteristics of the suspensions before addition of the dispersants. With no dispersant present the initial zeta potential of the suspension in double distilled water was y5.4"1.2 mV at a pH of 8.58"0.09 and a conductivity of 0.042 S y m (Table 4). Increasing the salt concentration slightly to 1 mM KCl (Table 5) reduced the initial zeta potential by a small amount to y5.2"1.6 mV, whereas the conductivity rose slightly (as would be expected) to 0.050 S y m and the pH fell slightly to 8.49"0.14. Increasing the salt concentration further to 5 mM (Table 6), again reduced the zeta potential slightly to y4.4"0.7 mV, increased the conductivity to 0.096 S y m, but increased the pH slightly to
Table 4 Initial characteristics of SDK 160 alumina powder suspensions in double distilled water Dispersant type PAA 1 PAA 2 PAA 3 PAA 4 PAA 5 PAA 6 PAA 7 PMAA 1 PMAA 2 PMMA 3 PMAA 4 Initial pH 8.51 8.63 8.56 8.52 8.71 8.80 8.50 8.53 8.54 8.56 8.57 Initial conductivity (Sym) 0.042 0.032 0.047 0.037 0.039 0.054 0.042 0.039 0.044 0.040 0.045 Initial zeta potential (mV) y4.4 y6.0 y7.2 y5.3 y4.5 y7.7 y5.0 y3.3 y6.1 y4.9 y4.6

76

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

Table 5 Initial characteristics of SDK 160 alumina powder suspensions in 1 mM KCl Dispersant type PAA 1 PAA 2 PAA 3 PAA 4 PAA 5 PAA 6 PAA 7 PMAA 1 PMAA 2 PMMA 3 PMAA 4 Initial pH 8.53 8.47 8.77 8.52 8.65 8.53 8.53 8.32 8.30 8.34 8.51 Initial conductivity (Sym) 0.050 0.053 0.052 0.054 0.043 0.054 0.049 0.047 0.048 0.048 0.055 Initial zeta potential (mV) y6.4 y4.5 y8.0 y5.3 y3.8 y4.7 y4.7 y4.3 y3.5 y3.6 y7.6

8.61"0.16. Given the error bars it can be seen that the pH stays constant with increasing salt concentration, whilst the zeta potential decreases slightly due to screening by the salt ions. With no salt present the optimum adsorbed amount was 0.42 mg y m2 independent of the molecular weight. Adding a small amount of salt (1 mM KCl) gives an increased adsorbed amount of 0.60 mg y m2 again independent of the molecular weight. Increasing the salt concentration further to 5 mM KCl reduces the adsorbed amount to 0.31 mg y m2, once again independent of the molecular weight. Once again the results can be explained by considering the configuration of the polyelectrolye and how it adsorbs on the surface. With no salt present the polyacrylic acid adsorbs onto the surface as a train and due to the highly charged nature of the polyelectrolyte the chains arrange themselves as far away from one another as
Table 6 Initial characteristics of SDK 160 alumina powder suspensions in 5 mM KCl Dispersant type PAA 1 PAA 2 PAA 3 PAA 4 PAA 5 PAA 6 PAA 7 PMAA 1 PMAA 2 PMMA 3 PMAA 4 Initial pH 8.43 8.81 8.41 8.57 8.40 8.79 8.77 8.79 8.56 8.63 8.58 Initial conductivity (Sym) 0.097 0.078 0.104 0.121 0.100 0.077 0.078 0.101 0.102 0.100 0.097 Initial zeta potential (mV) y5.0 y4.8 y4.2 y6.2 y4.3 y3.9 y3.8 y5.0 y3.6 y4.5 y3.9

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581 Table 7 Optimum amount of PMAA dispersant against background electrolyte strength Dispersant code PMAA 1 PMAA 2 PMAA3 PMAA4 Optimum amount in pure water (mg ym2) 0.57 0.62 0.58 0.56 Optimum amount in 1 mM KCl (mgym2) 0.39 0.44 0.46 0.42

77

Optimum amount in 5 mM KCl (mgym2) 0.39 0.44 0.42 0.47

possible. Increasing the salt concentration allows the trains to pack closer together due to charge screening, such that the chains can now pack closer together and hence the adsorbed amount is increased. The independence from molecular weight is due the dispersant adsorbing as a train. If the polyelectrolyte adsorbed with loops and tails instead then a strong dependence on molecular weight would be expected. Increasing the salt concentration to 5 mM has a dramatic difference as it affects the polyelectrolyte configuration such that the adsorbed amount is much less than the previous two cases, but without further experiments it is not possible to suggest what exactly is occurring at this point. The results from the polymethacrylic dispersants are not as clear cut due to the limited number of samples and molecular weights. The highest adsorbed amount occurs with the double distilled water (0.60 mg y m2) and adding salt reduces the adsorbed amount to approximately 0.42 mg y m2 in both cases. All three sets of results show an independence of molecular weight within error, so again suggesting that the polymethacrylic acid also adsorbs in the train configuration due to strong electrostatic attraction. Similar work has been performed by Santhiya and co workers w69x studying the adsorption of polyacrylic acid onto alumina using electroacoustics and rheology. The adsorption density decreased with increasing pH but increased with increasing molecular weight. This work differs to the current work in that they used low ionic strength and low pH suspensions. Close examination of their Fig. 2 data reveals that at high pH there is no molecular weight dependence as noted here. 4. Conclusions Electroacoustics is a powerful technique for investigating the zeta potential of concentrated aqueous suspensions. The technique is extremely useful for investigating the colloidal processing of ceramic powders. By optimising the processing conditions the interparticle forces can be manipulated to prevent flocculation and flaws eliminated in the final sintered body. Variables that can be altered include the pH and powder surface area along with the molecular weight of the dispersant. The technique is also suitable for detecting dissolution of material out of powders. The background electrolyte strength has been high lighted as being extremely important as it determines the conformation of the polyelectrolyte in solution, hence controlling the manner in which it adsorbs at the interface. How the polyelectrolyte

78

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

adsorbs then determines how much is adsorbed and whether this is sufficient to impart stability to the particles. The nature of the acids and bases used to adjust the pH of the suspension can also play an important role. References
w1x R.J. Hunter, Review recent developments in the electroacoustic characterisation of colloidal suspensions and emulsions, Colloids Surf. A 141 (1998) 3765. w2x R.W. OBrien, The electroacoustic equations for a colloidal suspension, J. Fluid Mech. 212 (1990) 8193. w3x R.W. OBrien, B.R. Midmore, A. Lamb, R.J. Hunter, Electroacoustic studies of moderately concentrated colloidal suspensions, Faraday Disc. 90 (1990) 301312. w4x R.W. OBrien, J. Colloid Interface Sci. 171 (1995) 495504. w5x A.A. Griffiths, Philos. Trans. R. Soc. Lond. A221 (1920) 163. w6x K. Kendall, N. Alford, J.D. Birchall, Naure 330 (1987) 5153. w7x P. Debye, J. Chem. Phys. 1 (1933) 13. w8x E. Yaeger, J. Bugosh, F. Hovoraka, J. McCarthy, J. Chem. Phys. 17 (1949) 411. w9x J. Hermans, Phil. Mag. 25 (1938) 426. w10x J. Hermans, Phil. Mag. 26 (1938) 674. w11x A. Rutgers, Nature 157 (1946) 74. w12x J.A. Enderby, Proc. R. Soc. (Lond) A 207 (1951) 329342. w13x F. Booth, J.A. Enderby, Proc. Phys. Soc. 45 (1952) 321324. w14x R. Zana, E.B. Yaeger, J. Phys. Chem. 71 (1967) 3502. w15x R. Zana, E.B. Yaeger Ultrasonic vibration potentials in: J.OM Bockris, B.E. Conway, R.E. White (Eds.), Modern Aspects of Electrochemistry, vol. 14, 1982, 161. w16x B.J. Marlow, D. Fairhurst, H.P. Pendse, Langmuir 4 (1988) 611626. w17x T. Oja, D. Cannon, G.L. Petersen, US Patent 4497, 208. w18x A. Babchin, R.S. Chow, R.P. Sawatsky, Adv. Coll. Inter. Sci. 30, 1989. w19x A.S. Dukhin, P.J. Goetz, Langmuir 12 (1996) 43354344. w20x J.L. Valdez, in: S.G. Malghan (Ed.), Electroacoustics for characterisation of particulates and suspensions, US Department of Commerce, Washington, DC, 1993, pp. 111128, NIST Special publication 856. w21x R.W. OBrien, J. Fluid Mech. 190 (1988) 7186. w22x P. Mazur, J.Th.G. Overbeek, Rec. Trav. Chim (Pays-Bas) 70 (1951) 83. w23x R.W. OBrien, P. Garside, R.J. Hunter, Langmuir 10 (1994) 931935. w24x R.J. Hunter, Zeta Potential in Colloid Science, Academic Press, New York, 1981. w25x R.W. OBrien, D.W. Cannon, W.N. Rowlands, Electroacoustic determination of particle size and zeta potential, J. Coll. Inter. Sci. 173 (1995) 406418, Chapter 3. w26x A.S. Dukhin, H. Ohshima, V.N. Shilov, P.J. Goetz, Langmuir 15 (1999) 34453451. w27x A.S. Dukhin, V.N. Shilov, H. Ohshima, P.J. Goetz, Langmuir 15 (1999) 66926706. w28x A.S. Dukhin, V.N. Shilov, H. Ohshima, P.J. Goetz, Langmuir 16 (2000) 26152620. w29x A.S. Dukhin, P.J. Goetz, S. Truesdail, Langmuir 17 (2001) 964968. w30x A.S. Dukhin, P.J. Goetz, Coll. Surf. A 144 (1998) 4958. w31x R.O. James, J. Texter, P.J. Scales, Langmuir 7 (1991) 19931997. w32x R.O. James, R.J. Hunter, R.W. OBrien, Langmuir 8 (1992) 420423. w33x P.J. Scales, E. Jones, Langmuir 8 (2) (1992) 385389. w34x W.N. Rowlands, R.J. Hunter, Clays Clay Miner. 40 (1992) 287291. w35x R.J. Hunter, M. James, Clays Clay Miner. 40 (1992) 644649.

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

79

w36x R.O. James, T.W. Healy, J. Colloid Interface Sci. 40 (1972) 4281. w37x S.G. Malghan, in: S.G. Malghan (Ed.), Electroacoustics for Characterisation of Particulates and Suspensions, US Department of Commerce, Washington DC, 1993, pp. 111128, NIST Special publication 856. w38x C. Feldmann, J. Merikhi, J. Coll. Inter. Sci. 223 (2000) 229234. w39x N. Ushifusa, M.J. Cima, J. Am. Cer. Soc. 74 (10) (1991) 24432447. w40x P.T. Pei, J.F. Kelly, S.G. Malghan, Coll. Surf. A 70 (3) (1993) 277287. w41x T.S.B. Sayers, Coll. Surf. 77 (10) (1993) 3947. w42x N.D. Sanders, K.J. Roth, Tappi J. 76 (11) (1993) 209214. w43x N.D. Sanders, K.J. Roth, Nord. Pulp Pap. Res. 8 (1) (1993) 148152. w44x J. Gron, P. Dahlvik, Ind. Carta 33 (8) (1995) 395398, (in Italian). w45x U. Paik, V.A. Hackley, J. Am. Cer. Soc. 83(10) 23812384. w46x J. Texter, Langmuir 8 (1) (1992) 291298. w47x V.A. Hackley, S.G. Malghan, J. Mater. Sci. 29 (17) (1994) 44204430. w48x H.M. Wang, R.A. McCauley, Ceram. Trans. 62 (1996) 149156. w49x (a) V.A. Hackley, R. Premachandran, S.G. Malghan, Key Eng. Mater. 8991 (Silicon nitride 93) (1994) 67968294 (b) V.A. Hackley, J. Am. Cer. Soc. 80 (9) (1997) 23152325. w50x V.A. Hackley, U. Paik, B.H. Kim, S.G. Malghan, J. Am. Cer. Soc. 80 (7) (1997) 17811788. w51x S.G. Malghan, P.S. Wang, A. Sivakumar, P. Somasundaran, Compos. Inter. 1 (3) (1993) 193210. w52x S.G. Malghan, R. Premachandran, P.T. Pei, Powder Tech. 79 (1) (1994) 4352. w53x U. Paik, V.A. Hackley, H.W. Lee, J. Am. Cer. Soc. 82 (4) (1999) 833840. w54x C. Walldal, J. Colloid Interface Sci. 217 (1) (1999) 4959. w55x C.A.M. Baltar, J.F. Oliveira, Miner. Eng. 11 (5) (1998) 463467. w56x P.A. Smith, R.A Haber, J. Am. Cer. Soc. 78 (7) (1995) 17371744. w57x V.E. Subhin, R.J. Hunter, R.W. OBrien, J. Colloid Interface Sci. 159 (1993) 174183. w58x R.T. Klingbiel, H. Coll, R.O. James, J. Texter, Coll. Surf. 68 (12) (1992) 103109. w59x R.W. OBrien, W.N. Rowlands, R.J. Hunter, in: S.B. Malghan (Ed.), NIST Special publication 856 Electroacoustics for charcterisation of particulates and suspensions. Washington DC, (1993), pp 122. w60x R.W. OBrien, W.N. Rowlands, R.J. Hunter, Characterization of ceramic materials using electroacoustics, in: J.H. Adair, J.A. Casey, C.A. Randall, S. Venigall (Eds.), Scientific and Technological Applications of Colloidal Suspensions. Ceramic Transactions, 54, American Ceramics Society, Westerville, OH, 1994, pp. 5366. w61x R.W. OBrien, J. Colloid Interface Sci. 171 (1995) 495504. w62x A. Pettersson, G. Marino, A. Pursiheimo, J.B. Rosenholm, J. Colloid Interface Sci. 228 (2000) 7381. w63x J.K. Beattie, A. Djerdjev, J. Am. Cer. Soc. 83 (10) (2000) 23602364. w64x C. Knosche, H. Friedrich, M. Stintz, Part. Part. Syst. Characterisation 14 (4) (1997) 175180. w65x M. Rasmusson, S. Wall, Coll. Surf. A 122 (13) (1997) 169181. w66x C.J. Shih, B.H. Lung, M.H. Hon, Mater. Chem. Phys 60 (2) (1999) 150157. w67x J. Cessarano III, I.A. Aksay, A. Bleier, J. Am. Cer. Soc. 71 (4) (1988) 250255. w68x J. Cessarano III, I.A. Aksay, J. Am. Cer. Soc. 71 (12) (1988) 10621067. w69x D. Santhiya, G. Nandini, S. Subramanian, K.A. Natarajan, S.G. Malghan, Coll. Surf. A 133 (1988) 157163. w70x M. Rasmusson, W. Rowlands, R.W. OBrien, R.J. Hunter, J. Colloid Interface Sci. 189 (1997) 92100. w71x C.A. Prestidge, W.N. Rowlands, Miner. Eng. 10 (10) (1997) 11071118. w72x R.J. Hunter, R.W. OBrien, Coll. Surf A 126 (1997) 123128. w73x R.W. OBrien, W.N. Rowlands, J. Colloid Interface Sci 159 (1993) 471476. w74x W.N. Rowlands, R.W. OBrien, J. Colloid Interface Sci. 175 (1995) 190200.

80

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581

w75x W.N. Rowlands, R.W. OBrien, R.J. Hunter, V. Patrick, J. Colloid Interface Sci. 188 (1997) 325335. w76x M. Kosmulski, J.B. Rosenholm, J. Phys. Chem. 100 (28) (1996) 11 68111 687. w77x M. Kosmulski, J.B. Rosenholm, Langmuir 15 (1999) 8934. w78x S.B. Johnson, P.J. Scales, T.W. Healy, Langmuir 15 (1999) 89358936. w79x S.B. Johnson, P.J. Scales, T.W. Healy, Langmuir 15 (1999) 28362843. w80x S.B. Johnson, A.S. Russel, P.J. Scales, Coll. Surf. 141 (1998) 119130. w81x Y.G. Berube, P.L. de Bruyn, J. Colloid Interface Sci. 28 (1968) 92. w82x L. Gierst, L. Vandenberghen, E. Nicholas, A. Fraboni, J. Electrochem. Soc. 113 (1996) 1025. w83x G.V. Franks, S.B. Johnson, P.J. Scales, D.V. Boger, T.W. Healy, Langmuir 15 (1999) 44114420. w84x S.B. Johnson, G.V. Franks, P.J. Scales, T.W. Healy, Langmuir 15 (1999) 28362843. w85x M. DeBoer, R.G. Leliveld, J.W. Geus, H.G. Bruil, Appl. Catalysis A 102 (1) (1993) 3551. w86x F.N. Desai, H.R. Hammad, K.F. Hayes, Langmuir 9 (11) (1993) 28882894. w87x A. Djerdjev, Coagulation of alpha alumina suspensions observed by electroacoustics. B.Sc Thesis University of Sydney 1996. w88x V.A. Hackley, S.G. Malghan, Ceram Trans. 62 (1996) 117124. w89x S.B. Johnson, D.R. Dixon, P.J. Scales, Coll. Surf. 146 (1999) 281291. w90x T. Ukigai, H. Sugawara, N. Tobori, Chem. Lett. 5 (1995) 371372. w91x A.L. Costa, C. Galassi, R. Greenwood, J. Colloid Interface Sci. 212 (1999) 350356. w92x R. Pradip, R.S. Premachandran, S.G. Malghan, Bull. Mater. Sci. 17 (6) (1994) 5360. w93x R.S. Premachandran, S.G. Malghan, Powder Technol. 79 (1) (1994) 5360. w94x V.A. Hackley, J. Am. Cer. Soc. 80 (9) (1997) 23152325. w95x V.A. Hackley, P.S. Wang, S.G. Malghan, Mater. Chem. Phys. 36 (12) (1993) 112118. w96x C. Galassi, F. Bertoni, S. Ardizzone, C.L. Bianchi, J. Mater. Res 15 (1) (2000) 155163. w97x C. Galassi, F. Bertoni, S. Ardizzone, C.L. Bianchi, J. Mater. Res 15 (1) (2000) 164169. w98x L. Bergstrom, E. Laarz, R. Greenwood Proceedings of Ultrasonic and Dielectric Characterisation Techniques for Suspended Particles. NIST (Aug 1997). w99x E. Laarz, L. Bergstrom, J. Eur. Cer. Soc. 20 (4) (2000) 431440. w100x R. Greenwood, K. Kendall, J. Eur. Cer. Soc. 19 (1999) 479488. w101x E. Roncari, C. Galassi, C. Bassarello, J. Am. Cer. Soc. 83 (12) (2000) 29932998. w102x R. Greenwood, L. Bergstrom, J. Eur. Cer. Soc. 17 (2) (1997) 537548. w103x R. Greenwood, K Kendall, J. Eur. Cer. Soc. 20 (2000) 7784. w104x H. Maier, J.A. Baker, J.C. Berg, J. Colloid Interface Sci. 119 (1987) 512517. w105x L.K. Koopal, J. Lyklema, Disc. Faraday Soc. 59 (1975) 230. w106x M.L. Carasso, W.N. Rowlands, R.W. OBrien, J. Colloid Interface Sci. 193 (1997) 200214. w107x M. Carasso, The effect of polymer adsorption on the electroacoustic signals of colloidal silica,. Ph.D. Thesis, University of Sydney, 1996. w108x N.P. Miller, J.C. Berg, Coll. Surf. A 59 (1991) 119128. w109x H. Hodne, A. Saasen, Cem. Concr. Res. 30 (2000) 17671772. w110x M. Burke, R. Greenwood, K. Kendall, J. Mater. Sci. 33 (1998) 51495156. w111x R. Greenwood, K. Kendall, Powder Technol. 113 (2000) 148157. w112x R. Greenwood, S.W. Kingman, N.A. Rowson, G. Brown, Powder Technol. 123 (2002) 199207. w113x C. Galassi, E. Roncari, R. Greenwood, A. Piancastelli, 5th EuroCeramic Conference, Key Engineering Materials, Vol. 132136, Trans Tech Publications Switzerland, (1997), p. 329332. w114x R. Greenwood, K. Kendall. Paper Number 5. World Congress on Particle Technology, Brighton 1998. CD Rom I. Chem. E. w115x J.B. Kayes, D.A. Rawlins, Coll. Polymer Sci. 257 (1979) 622. w116x J.A. Baker, J.C. Berg, R.A. Pearson, Langmuir 5 (1989) 339. w117x M. Faers, P.F. Luckam, Coll. Surf. A 86 (1994) 317. w118x M.S. Ahmed, M.S.El - Aasser, J.W. Vanderhoff in: Goddard, B. Vincent (Eds.), Polymer adsorption and dispersion stability A.C.S Symposium Series (1984), p. 77.

R. Greenwood / Advances in Colloid and Interface Science 106 (2003) 5581 w119x w120x w121x w122x w123x

81

M.J. Garvey, Th.F. Tadros, B. Vincent, J. Colloid Interface Sci. 49 (1974) 57. M.J. Garvey, Th.F. Tadros, B. Vincent, J. Colloid Interface Sci. 55 (1976) 440. J.A. Baker, H. Maier, E. Killman, Coll. Surf. A 3 (1988) (1988) 51. R. Greenwood, K. Kendall, Brit. Ceram. Trans. 97 (4) (1998) 174179. F. Shojai, A.B.A. Pettersson, T. Mantyla, J.B. Rosenholm, J. Eur. Cer. Soc. 20 (3) (2000) 277283.

You might also like