JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1994 VOL. 9 84 1 Effect of Colloidal Stability of Ceramic Suspensions on Nebulization of Slurries for Inductively Coupled Plasma Atomic Emission Spectrometry* Juan C. Fariiias and Rodrigo Moreno lnstituto de Cerarnica y Vidrio (C. S. 1. C.) 28500 Arganda del Rey Madrid Spain Jean-Michel Mermet Laboratoire des Sciences Analytiques Universite de Lyon I 69622 Villeurbanne Cedex France The direct solid analysis of ceramic powders can be carried out by inductively coupled plasma atomic emission spectrometry (ICP-AES) using slurry sample introduction. However a highly stable suspension is needed in order to obtain a representative aerosol for introduction into the ICP. In this work the importance of the effect of the rheology and the stability of the ceramic suspensions on the analytical results provided by slurry nebulization ICP-AES is demonstrated.The basic concepts involved in the stabilization and homogenization of ceramic slurries are discussed. A general overview of the stabilizing mechanisms (electro- static steric and electrosteric) and the role of the different stabilizing additives and the most adequate use of them is described. Alumina (A1,03) slurries as a case study are discussed. The rheological parameters such as zeta potential viscosity and sedimentation have been studied by changing the pH of the slurry and by introducing different dispersing additives (Dolapix PC-33 Darvan-7 Darvan-C sodium hexametaphosph- ate glycerol plus Kodak photoflow Triton X-1 00 and Produkt PKV-5088). Their effect on stability is discussed as well as the relationship between the stability of the slurry and the intensity and precision of the measurements provided by ICP-AES.It is clearly demonstrated that higher intensities and lower relative standard deviation values are obtained for a well-dispersed stable slurry. Keywords Inductively coupled plasma atomic emission spectrometry; signal enhancement; slurry nebuliz- ation; ceramic suspension; colloidal stability Inductively coupled plasma atomic emission spectrometry (ICP-AES) is nowadays the most appropriate technique for the chemical analysis of ceramic materials.'-5 However this technique requires that the sample be in solution prior to the analysis. The process of dissolution can lead to problems including losses of volatile elements contamination increase in the analysis time dilution of the analyte etc.In order to avoid these problems direct solid analysis by ICP-AES using slurry sample introduction techniques has received increasing interest over the last few years. The technique has been applied to different inorganic and ceramic materials such as k a ~ l i n ~ . ~ soils,1o slags," geological materials,12-16 firebrick,l7 alumina (A1203),18-20 silica ( SiO,)," titania ( Ti02),21 zirconia (Zr02),22 silicon carbide ( SiC)'9,20 and silicon nitride (Si3N4).23 The two main factors determining the capability of the technique are the particle size of the starting p o ~ d e r s ~ * " * ' ~ ~ ~ and the stability and homogeneity of the s l ~ r r i e s . ~ ~ - ~ ~ The effect of particle size has been studied by different workers either by using several particle size fractions of one sample (e.g.k a ~ l i n ~ clays,' slags," whole or by using materials with different particle sizes (e.g. kaolin,6 A1203,18'19 Si3N4',). In these investigations a definite relationship between particle size and transport and atomization efficiencies is demonstrated. It has been proved that large particles are excluded by the spray chamber and do not reach the plasma,6~18.19*24 and that only small particles contribute sig- nificantly to the analyte atomic emission signal. Different workers recommend working with particle sizes lower than around 10,23,26 8' and 5 pm.I8 The effect of the stability and homogeneity of the slurries has not been studied thoroughly in the literature.However a highly stable and homogeneous slurry is required in order to achieve a homogeneous reliable aerosol for introduction into the plasma in order to provide precise and accurate analytical results. The literature reports only on the introduction of * Presented at the 1993 European Winter Conference on Plasma Spectrochemistry Granada Spain January 10-1 5 1993. different additives to assure these characteristics of the slurry. Table 1 shows the different stabilizing methods that have been used for several types of inorganic materials. The following key aspects can be noted in this table. ( i ) In samples of the same type a great variety of different additives has been employed. For silicate-based materials (soils clays kaolins etc.) the following additives have been used glycerol + Kodak photoflow tetrasodium pyrophosphate (Na,P207) Triton X-100 NH solution HCl and HN03.For advanced ceramic materials (A120 TiO ZrO Sic Si,N4) glycerol sodium hexametaphosphate (NaHMP so-called Calgon) alone or mixed with monoisopropanolamine and ethanol have been used. In some cases no additives were employed. (ii) The same additives have been used by several workers with large differ- ences in their concentration varying by up to a factor of 100 (e.g. Triton X-100 and NaHMP); and (iii) in many cases no additives have been used. In these cases a simple agitation either mechanical or ultrasonic was used to disperse the powders. In any instance it seems that all the additives and stabilizing procedures have been used empirically and without a real understanding of their role in the suspension and stability mechanisms. Most of them may not be the most suitable and can be modified in order to enhance the stability properties of the slurry.Nevertheless some workers have shown an interest in some of the parameters involved in the stability of the slurries to be analysed by ICP-AES. Ebdon and Collier6 have compared the effect of three dispersants Calgon Dispex (sodium polyacryl- ate) and NH solution for kaolin slurries showing that NH solution is the most adequate; they studied the influence of NH concentration on the atomization efficiency and the influence of the slurry viscosity on the nebulization efficiency. Fernandez et a/." have also studied the effect of three disper- sants (Triton X-100 Na4P,07 and NH solution) in slag slurries; they also found that the most effective was NH solution.Laird et al.' calculated the percentage recovery in slurries of clay minerals when different concentrations of HNO,Table 1 Additives used in the literature for stabilizing suspensions of inorganic materials Concentration (%) 40 40 40 0.01 0.04 0.1 0.5 1 0.05 0.1 1 0.006 0.01 0.02 0.1 0.5 0.1 10 10 90 0.35 0.35 3.1 1 mol I-' 2 moll-' - - - - - - - Main additive Glycerol Other additives HC1 (0.5 mol 1 - I ) Kodak photoflow (0.2%) Comments 90 rnin ultrasonic agitation As an antifoaming agent and a dispersant. 10 rnin ultrasonic agitation 5 min magnetic agitation and 30 min ultrasonic bath to complete the dispersion 10 s ultrasonic probe Protective colloid.2 10 min magnetic agitation - - Sample ZrO Geological materials Technique ICP-AES ICP-AES Ref. 22 12 14 13 16 27 28 29 30 24 17 15 10 31 21 32 i 8 23 33 34 31 35 1 1 6 7 36 16 9 8 37 38 39 40 19 41 Kodak photoflow (2%) - Geological materials Silicate materials ICP-AES ICP-AES Triton X-100 HN03 (5%) - Coal and coal fly ash Iron oxide pigments Bituminous and sub-bituminous coal Whole coal Firebrick ETAAS * ETAAS* ICP-AES Wetting and dispersing agent As a dispersant ICP-AES ICP-AES Tetrasodium p y rophosphate ( Na4Pz07 1 As a dispersing agent Dispersant solution Thixotropic slurry Geochemical materials Soils Whole coal ICP-AES FI-ICP-AESt ETAAS* - - H2 antifoam agent (0.2%) and Viscalex HV30 thixotropic thickening agent (2%) Monoisopropanolamine (0.020/,) pH = 10 70 "C - Sodium hexametaphosphate [(NaPO3)J Ti02 ICP-AES > 10 min magnetic agitation Iron oxide and titanium oxide Refractory materiais Si3N Coal Whole coal Whole coal pigments a-AI203 CV-AAS f ICP-AES ICP-AES ICP- MS§ ETAAS* ETAAS* ETAAS* Dispersing agent As a dispersant. 1 min ultrasonic agitation As a dispersant Wetting agent.5 min magnetic agitation Wetting agent. 5 min magnetic agitation As a dispersant. 8 rnin ultrasonic agitation and 2 3 rnin 15 min ultrasonic bath and magnetic agitation As a dispersant. 30 rnin ultrasonic bath6 and magnetic magnetic agitation agitation6s7 - 5 min magnetic agitation and 30 rnin ultrasonic bath 15 s ultrasonic probe Sonicated for 30 s at 40 W and magnetic agitation 5 min magnetic stirring 3 5 min magnetic stirring 3 min magnetic stirring Vortex mixing - Aerosol OT Ethanol - HNO (0.5%) HNO ( 5 % ) - NH solution Slags Kaolin ICP-AES ICP-AES Kaolin Silicate materials Clays Clay minerals A1203 Soils Soil River sediment Refractory oxide powders (Al,O SiO and Sic) Sic DCP-AESI ICP-AES ICP-AES ICP-AES ETAAS* ETAAS* ETAAS* ETAAS* ICP-AES HNO HC1 None ETAAS * 15 rnin ultrasonic bath and magnetic agitation * Electrothermal atomization atomic absorption spectrometry.t Flow injection inductively coupled plasma atomic emission spectrometry. 1 Cold vapour atomic absorption spectrometry. 3 Inductively coupled plasma mass spectrometry. 1 Direct current plasma atomic emission spectrometry.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1994 VOL. 9 843 were added.Lopez Garcia and Hernandez Cordoba2* studied the stabilization provided by Triton X-100 at different concen- trations on iron oxide pigments. Van Borm et aL2’ focused on the importance of pH adjustment in stabilizing A1203 and Sic slurries. In most instances however a lack of understanding about the processes involved in the colloid chemistry of the slurries can be observed. For this reason the aim of the present work deals with two main objectives (1) to clarify the most important funda- mentals on which the theory of slurry stabilization is based showing the different types of stabilizing adhtives and the most adequate use of them; and (2) to demonstrate the influence of the stability on the intensity and precision of the ICP-AES measurements and consequently on the analytical results.Alumina slurries under different pH conditions and with different stabilization agents are reported as a case study. Theoretical Considerations Stabilization of Ceramic Slurries The stability of a suspension depends on the interaction among the particles. There is an attractive interaction due to the London-Van der Waals forces which tend to link the particles to each other. On the other hand when the particles are immersed in a polar liquid they develop an electrical double layer around them. This double layer provides a repulsive force among the particles. The slurry stability is only achieved when the repulsive interaction dominates over the attractive London-Van der Waals forces. This constitutes the electrostatic stabilizing r n e ~ h a n i s m ~ ~ ’ ~ ~ which can be provided by the following agents pH electrolytes and surfactants.Another possible kind of stabilization is the polymeric mechanism in which the stability is provided by means of long-chain polymers adsorbed onto the surface of the particles preventing contact between them. This mechanism usually known as the steric stabilizing mechanism is preferred in the case of high solids loading suspensions in non-aqueous media. A third possible mechanism is the so-called electrosteric stabilizing mechanism which results from a combination of both electrostatic and steric mechanisms. Electrostatic stabilizing mechanism When a particle is immersed in a polar liquid a potential gradient is originated between the particle surface and the liquid medium which can be described by the Nernst equation RT ZF E =Eo + - In a where Eo is the standard potential when the activity of the potential determining ion (a) is unity Z is its valence F is the Faraday constant R is the gas constant and T is the absolute temperature.In ceramic suspensions the ions that determine the potential are H+ and OH-. For oxides like A1203 the surface charge is negative and the protons are attracted to the surface thus creating a gradient of concentration from the surface to the liquid (Fig. 1). On the other hand the OH- ion concentration decreases near the surface. These gradients of concentration give rise to a charged layer from the surface to the liquid the so-called electrical double layer. This follows the Stern model which supposes a mono- layer of counter-ions strongly adsorbed to the particle surface and a diffuse layer in which the concentration of counter-ions decreases as the distance increases (Fig.2). When the concentration of the potential determining ion is altered the relative adsorption of ions onto the surface and the surface potential change according to the Nernst equation. There is a certain concentration of potential determining ions Polar liquid Fig. 1 Concentration gradient of H+ and OH- ions for a ceramic oxide particle dispersed in a polar liquid Particle surface /- I- Stern plane d 1/K Distance + Fig.2 Stern model of the electrical double layer. The potential changes from t+ho (surface potential) to t+ha (Stern potential) in the Stern plane and decreases up to 0 in the diffuse double layer 6 is the distance from the particle surface to the Stern plane; 1 / ~ is the Debye length that is the double layer thickness; and z is the potential at the Stern plane at which the positive and the negative surface activities are equal and the net surface potential is zero.This concentration defines the so-called isoelectric point; at this point there is no double layer and the particles flocculate. When a particle surrounded by a double layer is moving through the liquid the Stern layer and part of the diffuse layer move also. The potential at the shear plane is called the zeta potential (4‘).44 The zeta potential value is zero at the isoelectric point. These concepts are schematically illustrated in Fig. 3. The surface potential and the isoelectric point are not affected by the potential determining ions that is the pH.However the counter-ions (those with opposite charge to that of the surface) have a strong effect on the diffuse double layer.844 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1994 VOL. 9 1 Shear plane \ Extension of the diffuse layer of counter-ions Concentration of cations Stern layer Concentration of anions Fig. 3 Double layer model for an electronegative particle immersed in a polar liquid. The potential at the shear plane is the zeta potential An increasing concentration of counter-ions reduces the thick- ness of the double layer and thus the zeta potential because the distance between the surface and the shear plane does not change as plotted in Fig. 4.In practice a slurry becomes stable when the zeta potential value is high far away from the isoelectric point. This fact usually makes necessary the addition of one of the following dispersants (i) a potential determining ion (that is a pH modification); or (ii) a counter-ion. The effect of a counter-ion is given by the flocculation value (the concentration required to produce flocculation). This value decreases when the coun- ter-ion charge increases and for a series of ions of the same charge decreases when the relative atomic mass increases. For this reason monovalent ions (e.g. Na' NH4+) are preferred to disperse ceramic powders. The most usual dispersants for ceramic slurries are sodium orthophosphate (Na3P04) sodium acid pyrophosphate ( Na,H2P20,) Na4P,0 sodium tripoly- phosphate ( Na5P3OIo) and NaHMP (a cyclic structure with 6- 13 phosphate groups).In non-aqueous media the electrostatic repulsion is less Distance t Shear plane Fig. 4 Effect of concentration of counter-ions on the zeta potential. As the shear plane is located at a fixed distance the zeta potential I(/sl (ix. the potential energy at the shear plane) is higher for lower counter- ion concentration A. For higher concentrations B the zeta potential I(/s2 decreases effective due to the lower ionic concentration and the lower dielectric constant of organic liquids. However electrostatic stabilization can be achieved by means of surface active agents surfa act ant^).^' These agents have a hydrocarbon chain (lyo- phobic) and a polar group (lyophilic). The hydrocarbon part is soluble in oils while the head group is soluble in water.The surfactants may be non-ionic anionic or cationic depending on the charge of the head group. Usually the surfactants contain groups such as OH- -COOH -SO3- -NH2- -OS03- NH4+ etc. the most commonly employed being the following. Non-ionic Triton X (polyethyleneoxy- ethanol) and alkalonamides; anionic phosphate ester sulfosuc- cinate and sodium dodecylsulfate; and cationic quaternary ammonium. Steric stabilizing mechanism This mechanism is based on the adsorption of macromolecules onto the particle surface.46 Steric stabilization is preferred in organic media. However in order to provide the required stability the steric stabilizers must have relative molecular masses higher than 10000.On the other hand they must anchor strongly to the surface. The steric model is shown schematically in Fig. 5 (a). Poly(methy1 methacrylate) poly- acrylamide polystyrene poly(viny1 acetate) etc. are commonly used to provide steric stability which is advantageous for both aqueous and non-aqueous media since sterically stabilized particles can be thermodynamically stable whereas electro- statically stabilized particles are only metastable. The magni- tude of the steric repulsion depends on the surface coverage the configuration of the adsorbed polymers and the thickness of the adsorbed layer. Electrosteric stabilizing mechanism Under some conditions a combination of electrostatic and sferic mechanisms can be useful to enhance stability. This combination is referred to as the electrosteric mechanism which generally originates from charges associated with the anchored polymer that is a polyelectrolyte [Fig.5(b)].47 The use of polyelectrolytes is common practice in dispersing Fig.5 Stabilization of slurries by means of the adsorption of long- chain polymers onto the surface (a) steric mechanism in which the stability is achieved by steric hinderance and (b) electrosteric mechan- ism in which the steric effect is reinforced by the presence of charges along the chain providing an electrostatic repulsionJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1994 VOL. 9 845 ceramics the most usually used being the salts of organic tannic acids humic acids and polycarbonic acids. Rheology of Ceramic Slurries As reported by other workers,6 the viscosity influences both the aspiration rate and the nebulization process.The basic concepts of rheology4* are discussed below. The viscosity of a suspension is a measure of the internal resistance of its parts to the relative movement. The viscosity is Newtonian when the shear force per unit of area z between two parallel planes of moving liquid is proportional to the rate gradient duldx T = ~ I du/dx where y~ is the viscosity whose units are Pa s. This is the ideal behaviour of liquids but usually ceramic slurries do not show this ideal behaviour and the viscosity value depends on the shear rate. Furthermore the rheological behaviour can show a time dependency. Steady-state behuviour When the particles tend to agglomerate an increasing shear rate breaks the contacts between particles and the viscosity decreases.This phenomenon is known as pseudoplasticity or shear-thinning. Plasticity is very similar but in this instance the suspension does not flow until a certain shear stress limit is achieved. This limit is known as the yield point. A third type of behaviour is the so-called dilatancy or shear-thickening in which the viscosity increases with the shear rate. These three types of behaviour are plotted in Fig. 6(u). Time-dependent behaviour Two general types of behaviour exist. Thyxotropic behaviour is similar to plastic or pseudoplastic behaviour but depends on time. When a thyxotropic slurry is sheared at a constant rate the viscosity decreases; when the shear stops the system acquires its starting structure thus giving rise to a thyxotropic cycle.This behaviour is shown in Fig. 6(b). Shear rate - Shear rate -+ Shear rate - Fig. 6 Non-Newtonian rheological behaviour of slips. There are two general kinds of behaviour those showing no time dependency (a) and those varying with time thyxotropy (b) and rheopexy (c) Fig. 7 Structure of slurries of plate-like particles and possible aggre- gation mechanisms. When the system is relaxed the plates are parallel (a); when a disturbing force is applied the plates become disorientated forming structure by means of face to face interactions (b) giving place to card-pack aggregation (c) or by means of edge to face interactions card-house aggregation ( d ) The time-dependent effect in which the viscosity increases with time at a constant shear rate is known as rheopexy [Fig.6(c)]. The rheology of a slurry is very dependent on the shape of the particles. In general spherical particles show no time dependency. But in the case of platelets or rod-like particles some aggregation exists especially for highly concentrated slurries. This case can be illustrated as in Fig. 7(a) and (b) for a mass of platelets suspended in a liquid.49 When there is no external disturbing force the platelets arrange themselves with their long axes horizontal and parallel to each other [Fig. 7(a)] when this suspension is disturbed the particles become dis- orientated since they have to re-disperse in a new plane [Fig. 7(b)]. This is the case with highly concentrated slips of clays in which the surface has a negative net charge while the edges have a positive net charge.When the solids content increases the platelet particles attract each other to form aggregates [Fig. 7(c) and (43. This fact is responsible for the time dependent behaviour of concentrated clay slurries. Experimental Instrumentation The pH measurements were performed by using a Metrohm Model 691 pH meter. The zeta potential and the viscosity were measured with a mass transport analyser (Micromeritics USA) and with a rotational viscometer (Haake Rotovisco RV20 Germany) respectively. The ICP-AES measurements were performed with a Spectroflame-ICP D system from Spectro Analytical Instruments (Germany). This system is equipped with a 27 MHz generator and a dual monochromator. A high power of 1600 W and a low carrier gas flow rate of 0.6 1 min-' were used in order to minimize the possible problems due to incomplete volatilization of the particles.The i.d. of the injector was 1.7 mm. A conespray nebulizer and a single-pass spray chamber were used. Measurements of the line intensities were carried out based on the use of the standard direct peaking mode procedure. Both this procedure and a long integration time 10 s resulted in a lowering of the observed fluctuations of the signal due to the random injection of the particles and consequently in a decrease of the relative standard deviation (RSD) values. The elements and lines studied were the follow- ing the macro-constituent (A1 308.215 nm) and three minor elements (Ca 393.366 nm Si 288.158 nm and Fe 238.204 nm).846 JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1994 VOL.9 .. Reagents The desired pH was attained by adding either HCl or NH40H (Merck Germany); NaHMP (Carlo Erba Italy) was used as the inorganic electrolyte. Different surfactants and polyelectro- lytes were used to stabilize the slurries Triton X-100 (Fisher USA) and glycerol (BDH UK) plus Kodak photoflow liquid (Eastman Kodak USA) were used as surfactants. The polyelec- trolytes were Dolapix PC-33 (Schimmer-Schwarz Germany) Darvan-7 and Darvan-C (Vanderbilt USA) and Produkt PKV-5088 (Schimmer-Schwarz Germany). Samples A commercial A1203 powder prepared in two different milling steps was used (Al,03 Alcoa CT-3000 and CT-3000 SG). The first has aggregates with a mean size above 20 pm. The superground (SG) powder is a milled powder with a mean particle size of 0.5 pm in which the presence of soft agglomer- ates (about 4 pm) can be observed.Procedure The coarse powder was studied by using sedimentation tests. The fine powder was studied by means of zeta potential viscosity and sedimentation tests. The zeta potential was measured for concentrated aqueous slips as a function of pH and viscosity measurements were performed for concentrated slurries as a function of both pH and NaHMP concentration. The sedimentation tests were performed in 25 ml graduated test tubes for different de-flocculation conditions. The slurries were prepared with a solids content of 1% m/m followed by 15 min of agitation in an ultrasonic bath prior to the test. In order to achieve the best stability and homogeneity of the slurries different kinds of dispersants were used (1) changes in pH by adding HC1 or NH40H; (2) inorganic dispersants such as NaHMP; and (3) organic dispersants either surfactants or polyelectrolytes at different concentrations.For ICP-AES 20 consecutive emission intensity measurements from 2% m/m slurries were performed for each slurry and for each element. Results and Discussion Influence of pH on the Slurry Stability The zeta potential versus pH of A1203 fine powder is plotted in Fig. 8. As can be seen the isoelectric point occurs at pH 9. At this point the slurry is not stable. According to this plot stability could be achieved at either acidic or basic pH values because in these pH zones maximum zeta potential values are obtained.However in the basic zone very high viscosities are obtained when the pH is increased with NH40H.” This can be seen in Fig. 9 which shows the viscosity of alumina slurries at different pH values where the basic zone is not included since the viscosity is very high (out of scale of the viscometer). 40 m 35 t C O I I I 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 PH Fig. 9 Apparent viscosity versus pH for concentrated A1,0 slurries (shear rate= 100 s - l ) . The minimum viscosity is achieved at pH 4-5 As expected the minimum viscosity occurs at those pH values where the zeta potential is at a maximum i.e. pH4. This result is in good agreement with the expected result as pointed out before. Therefore the most appropriate conditions for the preparation of a homogeneous slurry is pH 4-5.The measurements mentioned above can also be verified with sedimentation tests. Thus the sedimentation curve of alumina slurries at different pH values is plotted in Fig. 10. From this curve it can be seen again that the best stability is achieved in the acidic zone. The slurry at pH 9 exhibits a very high sedimentation rate as expected taking into account that this pH corresponds to the isoelectric point. At pH 10 par- tial sedimentation occurs as predicted by the viscosity measurements. On the other hand the behaviour of the coarser alumina powder is governed by the presence of large hard agglomerates. In this instance the mixing of the powder is very difficult as the agglomerates settle when agitation stops. The sedimen- tation curve for this powder is shown in Fig.10; the pH control is not sufficient itself to provide the required stability. This stability can only be achieved by milling the starting agglomer- ated powder in a ball mill. In this instance as the agglomerates may be broken down dispersion of the powder is most likely to occur. All these data clearly demonstrate the determining role of the pH as an electrostatic stabilizer. The H + or OH- ions known as potential determining ions act on the particle modifying the electrical double layer thickness as discussed under Theoretical Considerations. Some workers as can be seen in Table 1 have added acidsg.16 or b a ~ e ~ ~ ~ ~ ~ but 30 I 0 2 4 6 8 10 12 14 Ti m e/m i n 6 5 6 7 8 9 10 1 1 12 13 PH Fig. 8 Zeta potential oersus pH of alumina slurries showing that the isoelectric point occurs at pH x 9 Fig.10 Sedimentation behaviour of A1,0 slurries fine grained CT-3000 SG at A pH 2.3 4 and 5.9; B pH 10.9; and C pH 9 and coarse agglomerated CT-3000 at D pH 2.3 4 5.9 9 and 10.9. The sedimentation height was defined as the height of the sedimentation front in a graduated test tubeJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1994 VOL. 9 847 160 2 E 120 8 80 v) -- c > v) v) .- 5 40 surprisingly they do not report the pH which is the parameter that actually controls the stability. - - - - Influence of Dispersants on Slurry Stability Different kinds of dispersants were studied by using sedimen- tation tests. In all cases the final pH values provided by the addition of these dispersants are between 8.8 and 9.2 (ie.the pH of minimum stability). The results are shown in Fig. 11. Firstly an inorganic dispersant usually employed in cer- amics such as NaHMP," is considered. Slips de-flocculated with this dispersant are stable with time and no sedimentation occurs as reported el~ewhere.~~ The use of NaHMP has also been reported for slurries of ceramic powders such as A1203,'* Ti0221 and Si3N4,23 at concentrations of 0.1 0.01 and 0.5% respectively using ICP-AES. However the effect of the disper- sant must also be controlled because an excess can be deletere- ous to the stability of the slurry. This fact is demonstrated in the present work for A1203 slurries. Fig. 12 shows the effect of the concentration of NaHMP on the viscosity of A1203 slurries. As can be seen the viscosity first decreases and then above a concentration of 0.1% m/m of NaHMP increases i.e.there is a concentration of NaHMP for which the viscosity is at a minimum. Thus in order to disperse a slurry properly adequate control of the additives is required; not only the kind of additive and its properties but also its most effective concentration. A typical surfactant such as Triton X-100 has also been studied. As can be seen from sedimentation behaviour this surfactant does not provide the expected stability in alumina slurries although it has been used for slurry ICP-AES. Other workers have used Kodak photoflow in glycerol to 30 I I A I Fig. 11 Effect of different dispersants on the sedimentation behaviour of the Al,03 slurry CT-3000 SG A Dolapix PC-33 Darvan-7 Darvan-C and NaHMP; B glycerol + Kodak photoflow; C Triton X-100 and D PKV-5088 2oo ; I I I I I I 0 0.2 0.4 0.6 0.8 1.0 [NaHMPl(% m/m) Fig.12 Viscosity of concentrated A1203 slurries as a function of NaH M P concentration disperse geological rnaterial~~'-'~ and ZrO powders.22 The results obtained in the present work indicate that no real dispersion is achieved. A few minutes are sufficient for the sedimentation. However the very high viscosity of glycerol makes sedimentation difficult. Lastly four common polyelectrolytes for dispersing alumina slurries were considered. PKV-5088 is a common dispersant for non-oxide ceramics but does not prevent the sedimentation of A1203. The other three dispersants (Dolapix PC-33 Darvan-C and Darvan-7) are generally used in ceramic slurries of oxides and are obviously able to keep the particles perfectly dispersed.52 As previously indicated all these dispersants modify the pH of the slurry to values of between 8.8 and 9.2.Although this pH corresponds to the isoelectric point the addition of these polyelectrolytes at concentrations of 0.5% m/m with respect to the solids is sufficient to provide the required stability. This fact can be explained by an electrosteric mechanism in which adsorption of the molecules onto the particle surface takes place. Influence of Fine and Coarse Powders on the Intensity and Precision of the ICP-AES Measurements Fig. 13 shows the values of the intensity and RSD for slurries prepared from both A1203 starting powders at the maximum stability pH of 4.The results for the macroconstituent (Al) and those obtained for the three minor components (Ca Fe and Si) are shown in the figure. As can be seen the finest powder gives the highest intensity values and the lowest RSD values. This is particularly important in the case of Al where the intensity increases in a 65 1 ratio although the RSD decreases by a factor of 20. For minor elements particularly Ca this variation is not so clear because the treatment of the slurries at acidic pH in an ultrasonic bath partially dissolves these elements. These differences are related to the stability of the slurry as slurries prepared from the coarse powder do not become stable as coarse agglomerates are present and tend to settle quickly as observed from the sedimentation curves.These agglomerates are rejected during transport and do not reach the plasma. Furthermore the atomization efficiency is lower. There is therefore a diminution of the homogeneity of the aerosol resulting in a lower precision in the measurements. Influence of pH on the Intensity and Precision of the ICP-AES Measurements The effect of the slurry pH on the analytical measurements can be seen in Fig. 14 where the values of intensity and RSD ta Intensity Al ' Ca I Fe I Si Al Ca ' Fe ' Si Element 18 16 14 12 10 * 5 a 6 4 2 0 Fig. 13 Intensity and RSD values of all the elements studied for the starting A1,0 powders at maximum stability conditions (pH = 4) (a) CT-3000; and (b) CT-3000 SGJOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1994 VOL. 9 14 El Intensity m RSD 12 10 8 - s n - 6 4 2 n 2.3 4.0 5.9 9.0 10.9 PH intensity and RSD values of A1 for CT-3000 SG under )H conditions nding to A1 from slurries of fine-grained A120 at pH values are plotted.The higher the stability of the t pH 4) the higher the intensity results and the lower values. Once again the worst results are obtained for st stability that is at the isoelectric point (pH 9). The btained at pH 11 are worse than those corresponding idic conditions which is consistent with the previously results. itensity and RSD values corresponding to the minor (Ca Fe and Si) are shown in Figs. 15 and 16 rely. The behaviour of these minor elements is quite o that of the macroconstituent. ! of the Dispersants on the Intensity and Precision of the S Measurements ct of the dispersant on the intensity and precision is related to the capability of the dispersant to provide ility to the system.urries prepared with the additives NaHMP Dolapix Darvan-7 and Darvan-C present a high stability as :d in Fig. 11. For these additives very high intensities r low RSD values are obtained for A1 as can be seen in These results are very similar to those obtained for prepared at pH 4 (Fig. 14). In contrast the Triton nd PKV-5088 additives do not provide the required A 1 a 2.3 4.0 5.9 9.0 10.9 PH Intensity values of Ca Fe (x40) and Si ( x 30) for CT-3000 r different pH conditions 16 14 12 10 a 6 4 2 n A I I I I I 2.3 4.0 5.9 9.0 10.9 PH Fig. 16 pH conditions RSD values of Ca Fe and Si for CT-3000 SG under different 550 16 I Intensity 14 A v) 4- .- 5 500 12 2 2 9 10 1 +-’ .- 450 8 n v) z CT X > 1 6 u .- 400 4 u - 2 350 0 A B C D E F G Additive Fig.17 Intensity and RSD values of A1 for CT-3000 SG with different dispersing additives A Triton X- 100; B PKV-5088; C glycerol + Kodak photoflow; D Dolapix PC-33; E Darvan-7; F Darvan-C and G NaHMP stability to the system. Consequently the slurries prepared with these additives show very low intensity and high RSD values similar to the case of slurries prepared at pH 9 where no stabilization occurs. The glycerol-Kodak photoflow mixture also presents low intensities and high RSD values which dem- onstrates the inability of this mixture to disperse the powder. In Figs. 18 and 19 intensity and RSD values respectively of the minor elements are plotted showing the same behaviour as A1 for both intensity and RSD.Conclusions Some efforts to facilitate the analysis of inorganic materials by using the slurry ICP-AES technique have been described in the literature. However the study of the stability of the slurries to be analysed has not been considered sufficiently. In this work the basic concepts involved in the stabilization of slurries are clarified. Besides transport and volatilization problems it has been demonstrated that the stability and the homogeneity of slurries play a significant role in the obtainment of adequate analytical performance. High emission intensity and high pre- cision are only achieved if the slurry is stable homogeneous and reliable as a more representative sample delivery to the plasma is achieved. The higher the stability of the slurry the higher the intensity and the lower RSD values obtained for both macroconstituents and minor elements.JOURNAL OF ANALYTICAL ATOMIC SPECTROMETRY AUGUST 1994 VOL.9 849 A B C D E F G Additive Ca Fig. 18 Intensity values of Ca Fe ( x 40) and Si ( x 30) for CT-3000 SG with different dispersing additives A Triton X-100; B PKV-5088; C glycerol+Kodak photoflow; D Dolapix PC-33; E Darvan-7; F Darvan-C; and G NaHMP 16 14 12 10 A 8 v) u 6 4 2 n 2 ; 8 Ca " A B C D E F G Additive Fig. 19 RSD values of Ca Fe and Si for CT-3000 SG with different dispersing additives A Triton X-100; B PKV-5088; C glycerol + Kodak photoflow; D Dolapix PC-33; E Darvan-7; F Darvan-C; and G NaHMP Different dispersant additives may be useful to provide the required stability to the slurry (1) potential determining ions (pH) (2) inorganic electrolytes or counter-ions (3) surfactants and (4) polyelectrolytes.The most usual methods of determin- ing the stability characteristics of the slurry are the zeta potential and the viscosity measurements. Sedimentation tests can also be performed to estimate the stability of the slurries these having the advantage that no equipment is required. This work has been financially supported by a BRITE- EURAM I1 project BE-5168. J. C. F. and J.-M. M. gratefully acknowledge Spectro Analytical Instruments for the loan of the Spectroflame-ICP D system. References 1 Morikawa H. and Ishizuka T. Analyst 1987 112 999. 2 Ishizuka T. Uwamino Y. Tsuge A. and Kamiyanagy T. Anal. Chim.Acta 1984 161 285. 3 Fariiias J. C. and Barba M. F. Mikrochim. Acta 1989 111 299. 4 Fariiias J. C. and Barba M. F. J. Anal. At. Spectrom. 1992,7,869. 5 Fariiias J. C. and Barba M. F. J. Anal. At. Spectrom. 1992,7,877. 6 Ebdon L. and Collier A. R. J . Anal. At. Spectrom. 1988 3 557. 7 Ebdon L. and Collier A. R. Spectrochim. Acta Part B 1988 43 355. 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 Laird D. A. Dowdy R. H. and Munter R. C. J. Anal. At. Spectrom. 1990 5 515. Spiers G. A. Dudas M. J. and Hodgins L. W. Clays Clay Miner. 1983 31 397. Ambrose A. J. Ebdon L. Foulkes M. E. and Jones P. J. Anal. At. Spectrom. 1989 4 219. Fernandez Sanchez M. L. Fairman B. and Sanz-Medel A. J.Anal. At. Spectrom. 1991 6 397. Verbeek A. A. and Brenner I. B. J. Anal. At. Spectrom. 1989 4 23. Long G. L. and Brenner I. B. J . Anal. At. Spectrom. 1990,5,495. Halicz L. and Brenner I. B. Spectrochim. Acta Part B 1987 42 207. Darke S. A. Long S. E. Pickford C. J. and Tyson J. F. Fresenius' J. Anal. Chem. 1990 337 284. Gervais L. S. and Salin E. D. J. Anal. At. Spectrom. 1991 6,41. Ebdon L. and Goodall P. Spectrochim. Acta Part B 1992 47 1247. Ebdon L. Foulkes M. E. and Hill S. J. Anal. At. Spectrom. 1990 5 67. Raeymaekers B. Graule T. Broekaert J. A. C. Adams F. and Tschopel P. Spectrochim. Acta Part B 1988 43 923. Van Borm W. A. H. Broekaert J. A. C. Klockenkamper R. Tschopel P. and Adams F. C. Spectrochim. Acta Part B 1991 46 1033. Broekaert J. A. C. Leis F.Raeymaekers B. and Zaray G. Spectrochim. Acta Part B 1988 43 339. Huang M. and Shen X.-e. Spectrochim. Acta Part B 1989 44 957. Isozaki A. Ogawa M. Shibagaki M. and Morita Y. Anal. Sci. (Suppl.) 1991 7 1249. Ebdon L. and Wilkinson J. R. J. Anal. At. Spectrom. 1987,2 39. Bendicho C. and de Loos-Vollebregt M. T. C. J. Anal. At. Spectrom. 1991 6 353. Willis J. B. Anal. Chem. 1975 47 1752. Bradshaw D. and Slavin W. Spectrochim. Acta Part B 1989 44 1245. Lopez Garcia I. and Hernandez Cordoba M. J . Anal. At. Spectrom. 1989 4 701. Lopez Garcia I. and Hernandez Cordoba M. J. Anal. At. Spectrom. 1990 5 647. McCurdy D. L. and Fry R. C. Anal. Chem. 1986 58 3126. Ebdon L. and Parry H. G. M. J. Anal. At. Spectrom. 1987,2 131. Lopez Garcia I. Vizcaino Martinez M. J. and Hernandez Cordoba M.1. Anal. At. Spectrom. 1991 6 627. Ebdon L. Foulkes M. E. Parry H. G. M. and Tye C. T. J. Anal. At. Spectrom. 1988 3 753. Ebdon L. and Parry H. G. M. J. Anal. At. Spectrom. 1988,3 131. MareEek J. and Synek V. J. Anal. At. Spectrom. 1990 5 385. Sparkes S. T. and Ebdon L. J. Anal. At. Spectrom. 1988 3 563. Karwowska R. and Jackson K. W. J. Anal. At. Spectrom. 1987 2 125. Hinds M. W. Katyal M. and Jackson K. W. J. Anal. At. Spectrom. 1988 3 83. Hinds M. W. Jackson K. W. and Newman A. P. Analyst 1985 110 947. Qiao H. and Jackson K. W. Spectrochim. Acta Part B 1992 47 1267. Docekal B. and Krivan V. J. Anal. At. Spectrom. 1992 7 521. Shaw D. J. Introduction to Colloid and Surface Chemistry Butterworths Boston MA 1980. Parfitt G. D. Dispersion of Powders in Liquids Applied Science Publishers New York 1981. Hunter R. J. Zeta Potential in Colloid Chemistry Academic Press New York 1981. Tadros T. F. Surfactants Academic Press London 1984. Napper D. H. J. Colloid Interface Sci. 1977 58 390. Moreno R. Am. Ceram. SOC. Bull. 1992 71 1521. Moreno R. Moya J. S. and Requena J. Bol. SOC. Esp. Ceram. Vidrio 1986 25 3. Michaels A. S. in Ceramic Fabrication Processes ed. Kingery W. D. Wiley New York 1958 pp. 23-31. Requena J. Moreno R. and Moya J. S. J. Am. Ceram. SOC. 1989 72 1511. Faisson J. and Haber R. A. Ceram. Eng. Sci. Proc. 1991 12 106. Moreno R. Requena J. and Moya J. S. J. Am. Ceram. SOC. 1988 71 1036. Paper 3/05699F Received September 21 1993 Accepted January 27 1994