- Email: [email protected]
Rheological characterisation of water-based AlN slurries for the tape casting process
Journal of Materials Processing Technology 169 (2005) 206–213
Rheological characterisation of water-based AlN slurries for the tape casting process S.M. Olhero, J.M.F. Ferreira ∗ Department of Ceramic and Glass Engineering, CICECO, University of Aveiro, 3810-193 Aveiro, Portugal Received 15 March 2004; received in revised form 30 November 2004; accepted 9 March 2005
Abstract In the present work rheological properties of aqueous concentrated AlN suspensions have been investigated in the presence of a sintering aid, deffloculant, binder and plasticizers, in order to screen the most suitable experimental conditions to obtain a good rheological behaviour for tape casting thick and non-cracked tapes with good flexibility. Suspensions exhibiting the desired shear thinning behaviour could be prepared. Adding binder and plasticizers did not change the shear thinning behaviour, but the viscosity revealed an increasing trend with increasing added amounts of binder. The flexibility was improved for the tapes derived from the more viscous suspensions containing higher amounts of binder. Crack free tapes having a maximum thickness of 1.5 mm, and with binder and plasticizer contents in the ranges of 10–15 wt.% and 5–10 wt.%, respectively, could be obtained. © 2005 Elsevier B.V. All rights reserved. Keywords: Rheology; AlN; Aqueous processing; Tape casting
1. Introduction Tape casting is a low cost process for making high quality laminated materials for which an adequate thickness control and good surface finish are required. This is one of the most widely used techniques for producing thin ceramic sheets that are subsequently laminated to fabricate products such as capacitors, substrates for integrated circuits for uses in electronic applications [1,2]. In this process, a well-mixed slurry consisting of a suspension of ceramic particles along with other additives, such as dispersants to assure the stabilisation, and binders and plasticizers to confer adequate strength and flexibility to the tape [3–5]. Once the suspension has been prepared it is cast onto a surface through the action of a blade that levels the slurry. The cast is then dried until the solvent has evaporated. The organic components, i.e., binder, plasticizer and dispersant, remain in the tape after casting, but they are generally eliminated by burning them out, thus generating an open porosity in the ∗
Corresponding author. Tel.: +351 234 370242; fax: +351 234 425300. E-mail address: [email protected] (J.M.F. Ferreira).
0924-0136/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2005.03.007
green body. Obviously, a homogeneous and uniform product can only be obtained if the starting suspension itself has a high homogeneity and stability. This homogeneity must be preserved during all the processing steps of casting, drying, burning out and sintering, and requires a careful selection and accurate control of the processing additives in the slurry. As in other forming methods, the arrangement and packing of the particles in the green body influences the sintering behaviour and the final properties. The green microstructure depends on the system to be consolidated and the forming technique employed. Assuming well-dispersed starting slurry, the microstructure of the casting tapes will be determined by two key processing factors: (i) particles’ arrangement during the casting process and the shrinkage during drying; (ii) the shear stress generated when the slurry passes under the blade. Due to all of these reasons, the rheological behaviour of the suspensions is of paramount importance in the tape casting process. The rheology determines the flow behaviour in the casting unit, which is dependent on the type and concentration of powder, binder, solvent and other organic additives such as dispersants and wetting agents.
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
Tape casting has traditionally been performed using organic solvents as dispersing liquid media but there is now a trend to move away from organic solvents and an expected transition towards water-based systems [6–8]. The main advantages for switching from organic to a water-based system are reduced health and environmental hazards coupled with a lower cost. However, the known reactivity of the AlN powders towards water makes the processing of this powder in aqueous media very difficult. Many efforts have been made to protect AlN powders against hydrolysis in order to obtain aqueous suspensions [9–12]. In a previous work, aqueous stable suspensions with 50 vol.% of aluminium nitride have been successfully prepared in the presence of different surface active agents, namely H3 PO4 and an anionic surfactant having carboxylic functional groups [13]. Nevertheless, the replacement of organic solvents by water has a strong influence on rheological behaviour of slurries due to the poor solvency of water towards organic binders [14]. This replacement also affects the drying behaviour of the tapes and the critical crackling thickness due to the higher surface tension of water that gives rise to the development of stronger capillary forces. A wide range of water soluble binders exists and several of these have been already evaluated for tape casting, including derivates of cellulose ethers such as hydroxyl-ethyl-cellulose, and hydroxyl-propylmethyl-cellulose, polyvinyl alcohols, acrylic polymers, etc. However, water soluble binders tend to increase the viscosity of suspensions and, because of that, the preference usually goes towards acrylic polymer emulsions, more commonly referred as latex binders [4,5,7]. In the present work rheological properties of aqueous AlN suspensions in presence of different amounts of dispersant, binder and plasticizers, have been investigated. The compatibility between binder and plasticizers, in presence of all the components in the formulation of the aqueous AlN suspensions (protective agent, dispersant, sintering additive) was evaluated, by viscosity measurements. Time dependent behaviour was evaluated and dynamic measurements were performed in order to select the most suitable processing conditions like the casting speed for a given tape thickness. Non-cracked green tapes with high thickness and good flexibility could be obtained.
2. Experimental procedure
207
glass transition temperature of the binder (approximately −10 ◦ C), it is flexible at room temperature. To increase the elasticity of the green tapes, two commercial polymeric products based on polyvinyl alcohol with different average molecular weights, 200 and 400 g mol−1 (P200 and P400, ´ AgoraMat Ldt, Portugal) were used as plasticizers. 2.2. Preparation and characterisation of the aqueous suspensions Stable aqueous AlN-SHS suspensions containing 40 and 50 vol.% (95 wt.% AlN-SHS + 5 wt.% CaF2 ) solids could be easily prepared following a method described elsewhere [15]. The method comprises the simultaneous adding of 0.4 wt.% H3 PO4 as a protective agent against hydrolysis, and a suitable amount of a dispersing agent (D), an anionic surface active agent. Rheological properties of the suspensions were determined using a rotational Rheometer (Bohlin C-VOR Instruments, UK). The measuring configuration adopted was a cone and plate (4◦ , 40 mm, and gap 150 m) and stress sweep measurements were conducted between 0.1 and 500 s−1 . A first set of rheological measurements (flow curves) was performed using suspensions with 40 vol.% total solids loading dispersed with several amounts of the surface active agent, for a given added amount of binder, in order to gather data about the effect of the added amount of dispersant and select the most appropriate conditions for dispersing the powders’ mixture. After determining the required amount of the anionic surface active agent, different proportions of binder and plasticizers 5, 10, and 15 wt.%, based on the mass of inorganic solids, were added and the mixtures let to stir in the polyethylene bottle in a rotating system, for 1 h. The flow curves of the as-obtained suspensions were measured to enable selection of the most appropriate amounts of the processing aids. Based on the experimental results, suspensions containing 50 vol.% solids could be prepared and further characterised concerning time dependent- and viscoelastic-properties. The time dependent behaviour was evaluated from the recovery time of the viscosity after a steep decrease of the shear rate from 50 (after an equilibration time of 60 s) to 1 s−1 . The suspensions were let at this shear rate during 5 min. To evaluate dynamic properties and obtain information on the behaviour of the suspensions in the linear viscoelastic region, stress sweeps with amplitude from 0.01 to 10 Pa at a constant frequency of 1 Hz were performed.
2.1. Materials 2.3. Preparation and consolidation of the green tapes An aluminium nitride powder obtained by self-propagating high temperature synthesis, hereafter designated by AlN-SHS, with a mean particle size of 2.5 m was used in this work. A commercial polymeric emulsion (MDM2, ´ AgoraMat Ldt, Portugal) with a pH value of about 4, was used as binder. It consists of an aqueous dispersion of small polymer particles with a diameter range of 0.3–2 m, with a solid content of about 53 wt.%. Due to the sufficiently low
The green tapes were prepared by casting the as-prepared suspensions onto a plastic film (Polypropylene (PP) Western Wallis, USA) with a laboratory tape caster (Elmetherm, Oradom Sur Vayres, France). A gap of 2 mm under the blade and a fixed casting speed of ≈3 cm s−1 were selected. The processing was carried out at room temperature and humidity. The green tapes were quantitatively characterised by
208
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
measuring the thickness and the flexibility. According to Greenwood [16], the flexibility of the final dried tapes was judged by placing the sample on a flat surface and bending one side around a glass rod. If the tape could be bent through and angle greater than 135◦ , about 10 times, it was judged to be flexible enough. SEM micrographs (SEM, Hitachi S-4100, Tokyo, Japan) of the tapes surface prepared from suspensions with 50 vol.% solids loading with different combinations of binder and plasticizer were obtained.
3. Results and discussion 3.1. Effects of different added amounts of dispersant on rheology of suspensions The added amounts of processing additives (dispersants, binders, plasticizers) and their possible mutual interactions play important roles in determining their compatibility and the rheological properties of the suspensions for the tape casting process [17]. Therefore, it is of paramount importance to understand the influence of each processing additive on the rheological properties and casting performance of the tape casting slurries. Since binders and plasticizers are commonly added to well-dispersed slurries, it is usual practice to make the detailed investigations according to the following hierarchy: dispersant → binder → plasticizers. Fig. 1 shows the steady shear viscosity curves of aqueous suspensions containing 40 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) dispersed with increasing amounts of dispersant from 0 to 0.5 wt.% relative to the dry mass of solids in absence, Fig. 1a, and in the presence of 10 wt.% binder (Fig. 1b). From Fig. 1a, it can be seen that the viscosity level decreases as the added amount of dispersant increases and that the rheological behaviour changes from shear thinning along the whole shear rate range to a more complex behaviour, especially when passing from 0.1 to 0.2 wt.%. In fact, for added amounts of dispersant ≥0.2 wt.%, the viscosity curves present a first shear thinning branch in the low shear rate range (up to about 50 s−1 ) followed by a very short Newtonian plateau and then by a shear thickening behaviour up to shear rates of about 200 s−1 where the viscosity reaches maximum values, which are kept almost constant with shear rate further increasing. The shear thinning behaviour is usually associated with the slurry structure. Under near rest conditions the interfacial forces dominate the particulate system and at low shear rates, liquid might be immobilised in void spaces within flocks and the flock network. As the shear rate increases, the flocks and flock network breaks down and the entrapped liquid is released and a more ordered structure in the flow direction is formed offering less resistance to flow [18]. Contrarily, at high shear rates the hydrodynamic interactions between particles become dominant and the number of collisions per unit time increases. For particles (or agglomerates) to slide over
Fig. 1. Viscosity curves of the aqueous suspensions containing 40 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) with different amounts of dispersant: (a) without binder and (b) in the presence of a 10 wt.% binder.
each other they have to increase their average separation distance, causing an apparent increase in solid volume fraction which is responsible for the increased viscosity, i.e., the shear thickening [19,20]. The shear thickening branch of the flow curves suggests the presence of particle agglomerates. This is not surprising in AlN powders prepared by SHS, since the combustion reaction generates temperatures >1600 ◦ C [21]. Fig. 1b shows that adding 10 wt.% of binder to the previous suspensions attenuates the differences among the viscosity curves with different additions of dispersant and smoothes the rheological behaviour. The shear thickening effects, although still present for shear rates ≥50 s−1 , are much less perceptible. Therefore, besides the smoothening effect on viscosity curves, the binder thickened the most fluid suspensions and thinned the higher viscosity ones. These results suggest that (i) the smaller sized latex particles of the binder might fit in the interstices formed by the coarser AlN particles and act as a lubricant [3]; (ii) there should be a certain competition of dispersant and binder for the available adsorption sites at the surface of the particles [4]. From the results shown in Fig. 1, it seems possible to conclude that: (i) the lower viscosity suspensions are obtained in the presence of 0.5 wt.% of deffloculant, especially for the formulations including the binder; (ii) although the
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
209
apparent competition between the dispersant and the binder for the available sites at particles’ surface, there is no evident incompatibility between these two processing additives [4]. 3.2. Effects of different added amounts of binder and plasticizer on rheology of suspensions Based on the results presented above, the dispersant concentration was fixed at 0.5 wt.%, while the added amounts of binder varied in the range of 5–15 wt.%. The concentration of solids was also kept at 40 vol.%. Fig. 2 shows the effect of the added amounts of binder on viscosity of the suspensions. An increasing trend of viscosity is observed with increasing added amounts of binder. Furthermore, all the viscosity curves presented a first shear thinning branch within the same shear rate range mentioned before (up to about 50 s−1 ) followed by a perturbation attributed to particles’ agglomerates. The shear thinning behaviour is desired for the tape casting process. This enables structural decomposition when the suspension passes under the blade and its level out, and the structural regeneration after passing the blade, avoiding particles segregation and unwanted post-casting flows. According to the experimental procedure described above, a gap of 2 mm under the blade and a fixed casting speed of ≈3 cm s−1 were used, corresponding to an applied shear rate of 15 s−1 during the tape casting process. This means that we are on the safe side and no shear thickening effects are experienced by the suspensions. In order to increase the flexibility and consequently easy handling of the green tapes, two plasticizers with different molecular weight, were tested. Fig. 3a and b show the viscosity curves of the suspensions containing 40 vol.% solids loading with a fixed amount of binder (10 wt.%) and different amounts of the two plasticizers, P200 and P400, respectively. The shear thinning behaviour of the suspensions was kept and no outstanding differences can be observed between the two plasticizers. Increasing the added amounts of both plasticizers enhance the fluidity of suspensions, an opposite
Fig. 2. Viscosity curves of the aqueous suspensions containing 40 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ), 0.5 wt.% of dispersant and different amounts of binder (5, 10, and 15 wt.%).
Fig. 3. Viscosity curves of the aqueous suspensions containing 40 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ), 10 wt.% of binder and different amounts of plasticizers: (a) P200; (b) P400.
effect in comparison with the binder. This is according to the low viscosity of the plasticizers and to their specific role in the tape casting process, improving the viscous character (flexibility) of the green tapes, even after drying, in detriment of the elastic properties. Once more, there is no evident incompatibility between dispersant, binder and plasticizers. From a practical point of view, the properties of the tapes were improved when derived from the higher viscosity suspensions containing higher amounts of binder. Based on these observations, combinations of 10 or 15 wt.% of binder with 5 or 10 wt.% of plasticizer were selected for the next experiments. Furthermore, these proper dispersion conditions also enabled to increase the solids loading of the suspensions to 50 vol.%. Maximising the solids loading is essential to obtain non-cracked tapes with high thickness. Cracks mostly originate from the slow drying of the tapes when higher volumes of water are present in suspension leading to high drying shrinkage values of the tapes. Fig. 4 presents the viscosity curves of the suspensions containing 50 vol.% solids in the presence of 10 and 15 wt.% binder combined with 5 and 10 wt.% of the two plasticizers (P200 and P400). In comparison with the suspensions with 40 vol.% solids, these suspensions with high solids loading (50 vol.%) also presented a shear thinning behaviour within the same low shear rate range and became smoother (lower tendencies to
210
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
Fig. 4. Viscosity curves of the aqueous suspensions containing 50 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ), and different combinations of binder and plasticizers.
shear thickening for higher shear rate values). As expected, the suspensions containing the higher molecular weight plasticizer (P400), exhibit higher viscosity values for a given added amount. Again, it is clear that viscosity accentuates with increasing the added amount of binder and decreases with increasing added amounts of plasticizers. From these results it is possible to extract information about the casting speed more adequate to these suspensions. As referred above, during tape casting the slurry undergoes a certain structural decomposition depending mostly on the casting velocity and gap height. According to Bitterlich and Heinrich [8], the casting speed should not be at the lower shear rate values, because in this region small variation of shear rate results in higher changes in viscosity, and therefore in changes of the fluid mechanic conditions in the casting head. The figures show that for all the suspensions the shear rate must be in the range of about 10–40 s−1 which corresponds to casting speeds of 2–8 cm s−1 for a gap height of 2 mm, therefore, comprising the casting speed of 3 cm s−1 used in the present work.
Fig. 5. Viscosity vs. time of the aqueous suspensions containing 50 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) and selected combinations of binder and plasticizers.
3.4. Viscoelastic properties of the suspensions Dynamic measurements were also performed in order to access information about the internal structure of the suspensions. Fig. 6a shows the dependence of the storage modulus G, and loss modulus G on shear stress for suspensions
3.3. Structural regeneration of the suspensions To check if the slurries exhibit any time dependent effects a shear rate step function was preset. First, the suspensions were submitted to a 50 s−1 shear rate during 60 s and then the shear rate was immediately decreased to 1 s−1 . Fig. 5 presents the viscosity versus time of the suspensions containing 50 vol.% solids and the same combinations of binder and plasticizers already used (Fig. 4). The viscosity increases continuously after the step decrease of the shear rate, attaining a steady viscosity value after few seconds. These results were confirmed by additional experiments and no evident hysterese effects in the viscosity curves could be detected after increasing and decreasing shear rates. From these results, it can be concluded that the higher molecular weight plasticizer (P400) enables a faster structural regeneration when compared with the other one (P200), and that the binder also plays an active role in regenerating the structure, which accentuates with increasing the added amount.
Fig. 6. G and G vs. shear stress of the aqueous suspensions containing 50 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) with different additions of processing additives: (a) 10 wt.% binder and 5 wt.% of each of the two plasticizers (P200 and P400); (b) 15 wt.% binder and 10 wt.% of each of the two plasticizers (P200 and P400).
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
211
Table 1 Green tape thickness and flexibility results for tapes obtained from suspensions containing 50 vol.% solids and different combinations of binder and plasticizers Sample
BI (wt.%)
P200 (wt.%)
P400 (wt.%)
Green thickness (mm)
Flexibilitya
A B C D
10 10 15 15
5 – 10 –
– 5 – 10
1.67 1.38 1.37 1.55
X X V V
Samples code: A 10 wt.% binder + 5 wt.% P200; B 10 wt.% binder + 5 wt.% P400; C 15 wt.% binder + 10 wt.% P200; D 15 wt.% binder + 10 wt.% P400. a X: non-flexible; V: flexible.
Fig. 7. Pictures of the green tapes obtained by tape casting from the aqueous suspensions.
containing 50 vol.% solids and combinations of 10 wt.% of binder + 5 wt.% of each plasticizer (P200 or P400) and Fig. 6b shows the same rheological parameters for the suspensions but with combinations of 15 wt.% of binder + 10 wt.% of each plasticizer (P200 or P400).
At low shear stresses, the suspensions exhibit a gellike character (G > G ). Hence, the elastic behaviour dominates over the viscous behaviour and the structure shows some rigidity. The crossover point where G = G (tan δ = G /G = 1) occurs at different shear stress values according to the added amounts of the processing additives, namely at about 0.4 Pa for the lower content of additives (Fig. 6a) and at about 1 Pa for the higher added amounts of additives (Fig. 6b). After the crossover point, G > G and the viscous character predominates over the elastic one. This corresponds to the structural decomposition referred above where the internal structure of the suspension is destroyed. According to Bitterlich and Heinrich [8], above a critical shear stress the weak attractive forces between the powder particles and or the binder and plasticizer are broken down by the external shear stress, which destroys the internal network. These are important characteristics of tape casting slurries, which must spread over the carrier to accomplish the process. This means that a shear
Fig. 8. Microstructures of the tapes prepared from aqueous suspensions containing 50 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) and different selected combinations of binders and plasticizers: (1) 10 wt.% binder and 5 wt.% P200; (2) 10 wt.% binder and 5 wt.% P400; (3) 15 wt.% binder and 10 wt.% of P200; (4) 15 wt.% binder and 10 wt.%; 15 wt.% binder and 10 wt.% P400.
212
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
stress value greater than the critical shear stress must be applied. 3.5. Properties of the green tapes Table 1 presents the results of thickness and flexibility for the tapes obtained from the suspensions with 50 vol.% solids loading with different amounts of the binder and of the two plasticizers (P200 and P400). For all the suspensions tested, it was possible to obtain non-cracked green tapes with a thickness higher than 1 mm. The tapes produced from suspensions with 10 wt.% of added binder were not flexible enough to pass the flexibility test mentioned above, however they could be easily handled. The minimum amount of binder required to produce flexible tapes with thickness as high as 1.5 mm was 15 wt.%. All the slips produced good, smooth and uniform tapes without cracks as shown in Fig. 7. Fig. 8 shows the microstructures of the tapes obtained from the suspensions containing 50 vol.% and different combinations of binder and plasticizers: (1) 10 wt.% binder + 5 wt.% P200; (2) 10 wt.% binder + 5 wt.% P400; (3) 15 wt.% binder + 10 wt.% P200 and (4) 15 wt.% binder + 10 wt.% P400. It can be observed that all the tapes present homogeneous and uniform microstructures.
4. Conclusions The results presented in this work demonstrate the feasibility of producing AlN-based green tapes with thicknesses up to 1.5 mm from aqueous slurries. This possibility will have a tremendous impact in the industry at all levels, since the organic and volatile solvents can be completely replaced by the low cost and environment friendly water. Three essential key points need to be overcome for achieving this goal: 1. Protecting the AlN particles against the strong hydrolysis reactions, what has been achieved by adding H3 PO4 . 2. Find a suitable dispersing agent compatible with the protecting agent, what has been accomplished using a carboxylic anionic surfactant to increase the surface charge density and dispersing the AlN powder + the sintering aid. 3. Adding suitable combinations of an aqueous acrylic emulsion binder and plasticizers to set the required rheological properties of slurries to the specific demands of the tape casting process. Furthermore, it has been shown that rheology is an appropriate and important tool to evaluate the suitability of the slurries for the tape casting process and to decide about the necessary modifications in order to match the required rheological properties of slurries. The desired shear thinning behaviour and the smoothness of the viscosity curves are enhanced with added binder and plasticizers, although they impart opposite effects on viscosity level.
Acknowledgements The first author wishes to thank Fundac¸a˜ o para a Ciˆencia e a Tecnologia for the fellowship grant SFRH/BD/8754/2002. The authors are also in debt to PRAI-Centro—Programa Regional de Acc¸o˜ es Inovadoras do Centro de Portugal, for the financial support under the Project: “S´ıntese por combust˜ao de nitreto de alum´ınio (AlN) e processamento de substratos de AlN por tape casting em meio aquoso”. References [1] A. Roosen, Basic requirements for tape casting of ceramic powders, Ceram. Trans. 1 (1998) 675–692. [2] A.I.Y. Tok, F.Y.C. Boey, K.A. Khor, Tape casting of high dielectric ceramic composite substrates for microelectronics application, J. Mater. Process. Technol. 89–90 (1999) 508–512. [3] F. Doreau, G. Tari, C. Chartier, J.M.F. Ferreira, Processing of aqueous tape casting of alumina with acrylic emulsion binders, J. Eur. Ceram. Soc. 18 (1998) 311–321. [4] A. Kristoffersson, R. Lapsin, C. Galassi, Study of interactions between polyelectrolyte dispersants, alumina and latex binders by rheological characterisation, J. Eur. Ceram. Soc. 18 (1998) 2133–2140. [5] A. Kristoffersson, E. Roncari, C. Galassi, Comparison of different binders for water-based tape casting of alumina, J. Eur. Ceram. Soc. 18 (1998) 2123–2131. [6] Z. Yuping, J. Dongliang, P. Greil, Tape casting of aqueous Al2 O3 slurries, J. Eur. Ceram. Soc. 20 (2000) 1691–1697. [7] A. Kristofferson, Elis Carlstrom, Tape casting of alumina in water with an acrylic latex binder, J. Eur. Ceram. Soc. 17 (1997) 289– 297. [8] Bernd Bitterlich, Jurgen G. Heinrich, Aqueous tape casting of silicon nitride, J. Eur. Ceram. Soc. 22 (13) (2002) 2427–2434. [9] K. Krnel, T. Kosmac, Protection of AlN powder against hydrolysis using aluminium dihydrogen phosphate, J. Eur. Ceram. Soc. 21 (2001) 2075–2079. [10] T. Kosmac, K. Krnel, K. Kos, “Process for the protection of AlN powder against hydrolysis”, International Patent No. WO 99/12850, 18.03 (1999). [11] Y. Shimizu, J. Hatano, T. Hyodo, M. Egashira, Ion-exchange loading of yttrium acetate as a sintering aid on aluminium nitride powder via aqueous processing, J. Am. Ceram. Soc. 83 (11) (2000) 2793– 2797. [12] Y. Shimizu, K. Kawanabe, Y. Taky, Y. Takao, M. Egashira, AlN ceramics prepared by aqueous colloidal processing, in: H. Hausner, G.L. Messing, S. Hirano (Eds.), Ceramic Processing Science and Technology, Ceramic Transactions, vol. 51, American Ceramic Society, Westerville, OH, 1995, pp. 403–407. [13] M. Oliveira, S. Olhero, J. Rocha, J.M.F. Ferreira, Controlling hydrolysis and dispersion of AlN powders in aqueous media, J. Colloid Interface Sci. 261 (2003) 456–463. [14] B. Bitterlich, C. Lutz, A. Roosen, Rheological characterization of water-based slurries for the tape casting process, Ceram. Int. 28 (2002) 675–683. [15] M. Oliveira, S. Olhero, J.M.F. Ferreira, “M´etodos de passivac¸a˜ o do AlN em relac¸a˜ o a` hidr´olise, de processamento coloidal de cerˆamicos a` base de AlN em meio aquoso e de granulac¸a˜ o de p´os a partir dessas suspens˜oes”, Portuguese Patent No. 102852 (2002). [16] R. Greenwood, E. Roncari, C. Galassi, Preparation of concentrated aqueous alumina suspensions for tape casting, J. Eur. Ceram. Soc. 17 (12) (1997) 1393–1401. [17] Jennifer Lewis, Colloidal Processing of Ceramics, J. Am. Ceram. Soc. 83 (10) (2000) 2341–2359.
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213 [18] M.D. Sacks, Properties of silica suspensions and cast bodies, Am. Ceram. Soc. Bull. 63 (1984) 1510–1515. [19] L. Bergstrom, Rheology of concentrated suspensions, in: R.J. Pugh, L. Bergstrom (Eds.), Surface and Colloid Chemistry in Advanced Ceramic Processing, Surfactant Science Series, vol. 51, Marcel Dekker, New York, 1994, p. 193.
213
[20] H.A. Barnes, Dilatancy in suspensions of non aggregating solid particles dispersed in Newtonian liquids, J. Rheol. 33 (1989) 329– 366. [21] Renli Fu, Xin Xu, Jos´e M.F. Ferreira, Highly crystalline AlN particles synthesized by SHS, Mater. Lett., accepted for publication.
Rheological characterisation of water-based AlN slurries for the tape casting process S.M. Olhero, J.M.F. Ferreira ∗ Department of Ceramic and Glass Engineering, CICECO, University of Aveiro, 3810-193 Aveiro, Portugal Received 15 March 2004; received in revised form 30 November 2004; accepted 9 March 2005
Abstract In the present work rheological properties of aqueous concentrated AlN suspensions have been investigated in the presence of a sintering aid, deffloculant, binder and plasticizers, in order to screen the most suitable experimental conditions to obtain a good rheological behaviour for tape casting thick and non-cracked tapes with good flexibility. Suspensions exhibiting the desired shear thinning behaviour could be prepared. Adding binder and plasticizers did not change the shear thinning behaviour, but the viscosity revealed an increasing trend with increasing added amounts of binder. The flexibility was improved for the tapes derived from the more viscous suspensions containing higher amounts of binder. Crack free tapes having a maximum thickness of 1.5 mm, and with binder and plasticizer contents in the ranges of 10–15 wt.% and 5–10 wt.%, respectively, could be obtained. © 2005 Elsevier B.V. All rights reserved. Keywords: Rheology; AlN; Aqueous processing; Tape casting
1. Introduction Tape casting is a low cost process for making high quality laminated materials for which an adequate thickness control and good surface finish are required. This is one of the most widely used techniques for producing thin ceramic sheets that are subsequently laminated to fabricate products such as capacitors, substrates for integrated circuits for uses in electronic applications [1,2]. In this process, a well-mixed slurry consisting of a suspension of ceramic particles along with other additives, such as dispersants to assure the stabilisation, and binders and plasticizers to confer adequate strength and flexibility to the tape [3–5]. Once the suspension has been prepared it is cast onto a surface through the action of a blade that levels the slurry. The cast is then dried until the solvent has evaporated. The organic components, i.e., binder, plasticizer and dispersant, remain in the tape after casting, but they are generally eliminated by burning them out, thus generating an open porosity in the ∗
Corresponding author. Tel.: +351 234 370242; fax: +351 234 425300. E-mail address: [email protected] (J.M.F. Ferreira).
0924-0136/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jmatprotec.2005.03.007
green body. Obviously, a homogeneous and uniform product can only be obtained if the starting suspension itself has a high homogeneity and stability. This homogeneity must be preserved during all the processing steps of casting, drying, burning out and sintering, and requires a careful selection and accurate control of the processing additives in the slurry. As in other forming methods, the arrangement and packing of the particles in the green body influences the sintering behaviour and the final properties. The green microstructure depends on the system to be consolidated and the forming technique employed. Assuming well-dispersed starting slurry, the microstructure of the casting tapes will be determined by two key processing factors: (i) particles’ arrangement during the casting process and the shrinkage during drying; (ii) the shear stress generated when the slurry passes under the blade. Due to all of these reasons, the rheological behaviour of the suspensions is of paramount importance in the tape casting process. The rheology determines the flow behaviour in the casting unit, which is dependent on the type and concentration of powder, binder, solvent and other organic additives such as dispersants and wetting agents.
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
Tape casting has traditionally been performed using organic solvents as dispersing liquid media but there is now a trend to move away from organic solvents and an expected transition towards water-based systems [6–8]. The main advantages for switching from organic to a water-based system are reduced health and environmental hazards coupled with a lower cost. However, the known reactivity of the AlN powders towards water makes the processing of this powder in aqueous media very difficult. Many efforts have been made to protect AlN powders against hydrolysis in order to obtain aqueous suspensions [9–12]. In a previous work, aqueous stable suspensions with 50 vol.% of aluminium nitride have been successfully prepared in the presence of different surface active agents, namely H3 PO4 and an anionic surfactant having carboxylic functional groups [13]. Nevertheless, the replacement of organic solvents by water has a strong influence on rheological behaviour of slurries due to the poor solvency of water towards organic binders [14]. This replacement also affects the drying behaviour of the tapes and the critical crackling thickness due to the higher surface tension of water that gives rise to the development of stronger capillary forces. A wide range of water soluble binders exists and several of these have been already evaluated for tape casting, including derivates of cellulose ethers such as hydroxyl-ethyl-cellulose, and hydroxyl-propylmethyl-cellulose, polyvinyl alcohols, acrylic polymers, etc. However, water soluble binders tend to increase the viscosity of suspensions and, because of that, the preference usually goes towards acrylic polymer emulsions, more commonly referred as latex binders [4,5,7]. In the present work rheological properties of aqueous AlN suspensions in presence of different amounts of dispersant, binder and plasticizers, have been investigated. The compatibility between binder and plasticizers, in presence of all the components in the formulation of the aqueous AlN suspensions (protective agent, dispersant, sintering additive) was evaluated, by viscosity measurements. Time dependent behaviour was evaluated and dynamic measurements were performed in order to select the most suitable processing conditions like the casting speed for a given tape thickness. Non-cracked green tapes with high thickness and good flexibility could be obtained.
2. Experimental procedure
207
glass transition temperature of the binder (approximately −10 ◦ C), it is flexible at room temperature. To increase the elasticity of the green tapes, two commercial polymeric products based on polyvinyl alcohol with different average molecular weights, 200 and 400 g mol−1 (P200 and P400, ´ AgoraMat Ldt, Portugal) were used as plasticizers. 2.2. Preparation and characterisation of the aqueous suspensions Stable aqueous AlN-SHS suspensions containing 40 and 50 vol.% (95 wt.% AlN-SHS + 5 wt.% CaF2 ) solids could be easily prepared following a method described elsewhere [15]. The method comprises the simultaneous adding of 0.4 wt.% H3 PO4 as a protective agent against hydrolysis, and a suitable amount of a dispersing agent (D), an anionic surface active agent. Rheological properties of the suspensions were determined using a rotational Rheometer (Bohlin C-VOR Instruments, UK). The measuring configuration adopted was a cone and plate (4◦ , 40 mm, and gap 150 m) and stress sweep measurements were conducted between 0.1 and 500 s−1 . A first set of rheological measurements (flow curves) was performed using suspensions with 40 vol.% total solids loading dispersed with several amounts of the surface active agent, for a given added amount of binder, in order to gather data about the effect of the added amount of dispersant and select the most appropriate conditions for dispersing the powders’ mixture. After determining the required amount of the anionic surface active agent, different proportions of binder and plasticizers 5, 10, and 15 wt.%, based on the mass of inorganic solids, were added and the mixtures let to stir in the polyethylene bottle in a rotating system, for 1 h. The flow curves of the as-obtained suspensions were measured to enable selection of the most appropriate amounts of the processing aids. Based on the experimental results, suspensions containing 50 vol.% solids could be prepared and further characterised concerning time dependent- and viscoelastic-properties. The time dependent behaviour was evaluated from the recovery time of the viscosity after a steep decrease of the shear rate from 50 (after an equilibration time of 60 s) to 1 s−1 . The suspensions were let at this shear rate during 5 min. To evaluate dynamic properties and obtain information on the behaviour of the suspensions in the linear viscoelastic region, stress sweeps with amplitude from 0.01 to 10 Pa at a constant frequency of 1 Hz were performed.
2.1. Materials 2.3. Preparation and consolidation of the green tapes An aluminium nitride powder obtained by self-propagating high temperature synthesis, hereafter designated by AlN-SHS, with a mean particle size of 2.5 m was used in this work. A commercial polymeric emulsion (MDM2, ´ AgoraMat Ldt, Portugal) with a pH value of about 4, was used as binder. It consists of an aqueous dispersion of small polymer particles with a diameter range of 0.3–2 m, with a solid content of about 53 wt.%. Due to the sufficiently low
The green tapes were prepared by casting the as-prepared suspensions onto a plastic film (Polypropylene (PP) Western Wallis, USA) with a laboratory tape caster (Elmetherm, Oradom Sur Vayres, France). A gap of 2 mm under the blade and a fixed casting speed of ≈3 cm s−1 were selected. The processing was carried out at room temperature and humidity. The green tapes were quantitatively characterised by
208
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
measuring the thickness and the flexibility. According to Greenwood [16], the flexibility of the final dried tapes was judged by placing the sample on a flat surface and bending one side around a glass rod. If the tape could be bent through and angle greater than 135◦ , about 10 times, it was judged to be flexible enough. SEM micrographs (SEM, Hitachi S-4100, Tokyo, Japan) of the tapes surface prepared from suspensions with 50 vol.% solids loading with different combinations of binder and plasticizer were obtained.
3. Results and discussion 3.1. Effects of different added amounts of dispersant on rheology of suspensions The added amounts of processing additives (dispersants, binders, plasticizers) and their possible mutual interactions play important roles in determining their compatibility and the rheological properties of the suspensions for the tape casting process [17]. Therefore, it is of paramount importance to understand the influence of each processing additive on the rheological properties and casting performance of the tape casting slurries. Since binders and plasticizers are commonly added to well-dispersed slurries, it is usual practice to make the detailed investigations according to the following hierarchy: dispersant → binder → plasticizers. Fig. 1 shows the steady shear viscosity curves of aqueous suspensions containing 40 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) dispersed with increasing amounts of dispersant from 0 to 0.5 wt.% relative to the dry mass of solids in absence, Fig. 1a, and in the presence of 10 wt.% binder (Fig. 1b). From Fig. 1a, it can be seen that the viscosity level decreases as the added amount of dispersant increases and that the rheological behaviour changes from shear thinning along the whole shear rate range to a more complex behaviour, especially when passing from 0.1 to 0.2 wt.%. In fact, for added amounts of dispersant ≥0.2 wt.%, the viscosity curves present a first shear thinning branch in the low shear rate range (up to about 50 s−1 ) followed by a very short Newtonian plateau and then by a shear thickening behaviour up to shear rates of about 200 s−1 where the viscosity reaches maximum values, which are kept almost constant with shear rate further increasing. The shear thinning behaviour is usually associated with the slurry structure. Under near rest conditions the interfacial forces dominate the particulate system and at low shear rates, liquid might be immobilised in void spaces within flocks and the flock network. As the shear rate increases, the flocks and flock network breaks down and the entrapped liquid is released and a more ordered structure in the flow direction is formed offering less resistance to flow [18]. Contrarily, at high shear rates the hydrodynamic interactions between particles become dominant and the number of collisions per unit time increases. For particles (or agglomerates) to slide over
Fig. 1. Viscosity curves of the aqueous suspensions containing 40 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) with different amounts of dispersant: (a) without binder and (b) in the presence of a 10 wt.% binder.
each other they have to increase their average separation distance, causing an apparent increase in solid volume fraction which is responsible for the increased viscosity, i.e., the shear thickening [19,20]. The shear thickening branch of the flow curves suggests the presence of particle agglomerates. This is not surprising in AlN powders prepared by SHS, since the combustion reaction generates temperatures >1600 ◦ C [21]. Fig. 1b shows that adding 10 wt.% of binder to the previous suspensions attenuates the differences among the viscosity curves with different additions of dispersant and smoothes the rheological behaviour. The shear thickening effects, although still present for shear rates ≥50 s−1 , are much less perceptible. Therefore, besides the smoothening effect on viscosity curves, the binder thickened the most fluid suspensions and thinned the higher viscosity ones. These results suggest that (i) the smaller sized latex particles of the binder might fit in the interstices formed by the coarser AlN particles and act as a lubricant [3]; (ii) there should be a certain competition of dispersant and binder for the available adsorption sites at the surface of the particles [4]. From the results shown in Fig. 1, it seems possible to conclude that: (i) the lower viscosity suspensions are obtained in the presence of 0.5 wt.% of deffloculant, especially for the formulations including the binder; (ii) although the
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
209
apparent competition between the dispersant and the binder for the available sites at particles’ surface, there is no evident incompatibility between these two processing additives [4]. 3.2. Effects of different added amounts of binder and plasticizer on rheology of suspensions Based on the results presented above, the dispersant concentration was fixed at 0.5 wt.%, while the added amounts of binder varied in the range of 5–15 wt.%. The concentration of solids was also kept at 40 vol.%. Fig. 2 shows the effect of the added amounts of binder on viscosity of the suspensions. An increasing trend of viscosity is observed with increasing added amounts of binder. Furthermore, all the viscosity curves presented a first shear thinning branch within the same shear rate range mentioned before (up to about 50 s−1 ) followed by a perturbation attributed to particles’ agglomerates. The shear thinning behaviour is desired for the tape casting process. This enables structural decomposition when the suspension passes under the blade and its level out, and the structural regeneration after passing the blade, avoiding particles segregation and unwanted post-casting flows. According to the experimental procedure described above, a gap of 2 mm under the blade and a fixed casting speed of ≈3 cm s−1 were used, corresponding to an applied shear rate of 15 s−1 during the tape casting process. This means that we are on the safe side and no shear thickening effects are experienced by the suspensions. In order to increase the flexibility and consequently easy handling of the green tapes, two plasticizers with different molecular weight, were tested. Fig. 3a and b show the viscosity curves of the suspensions containing 40 vol.% solids loading with a fixed amount of binder (10 wt.%) and different amounts of the two plasticizers, P200 and P400, respectively. The shear thinning behaviour of the suspensions was kept and no outstanding differences can be observed between the two plasticizers. Increasing the added amounts of both plasticizers enhance the fluidity of suspensions, an opposite
Fig. 2. Viscosity curves of the aqueous suspensions containing 40 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ), 0.5 wt.% of dispersant and different amounts of binder (5, 10, and 15 wt.%).
Fig. 3. Viscosity curves of the aqueous suspensions containing 40 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ), 10 wt.% of binder and different amounts of plasticizers: (a) P200; (b) P400.
effect in comparison with the binder. This is according to the low viscosity of the plasticizers and to their specific role in the tape casting process, improving the viscous character (flexibility) of the green tapes, even after drying, in detriment of the elastic properties. Once more, there is no evident incompatibility between dispersant, binder and plasticizers. From a practical point of view, the properties of the tapes were improved when derived from the higher viscosity suspensions containing higher amounts of binder. Based on these observations, combinations of 10 or 15 wt.% of binder with 5 or 10 wt.% of plasticizer were selected for the next experiments. Furthermore, these proper dispersion conditions also enabled to increase the solids loading of the suspensions to 50 vol.%. Maximising the solids loading is essential to obtain non-cracked tapes with high thickness. Cracks mostly originate from the slow drying of the tapes when higher volumes of water are present in suspension leading to high drying shrinkage values of the tapes. Fig. 4 presents the viscosity curves of the suspensions containing 50 vol.% solids in the presence of 10 and 15 wt.% binder combined with 5 and 10 wt.% of the two plasticizers (P200 and P400). In comparison with the suspensions with 40 vol.% solids, these suspensions with high solids loading (50 vol.%) also presented a shear thinning behaviour within the same low shear rate range and became smoother (lower tendencies to
210
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
Fig. 4. Viscosity curves of the aqueous suspensions containing 50 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ), and different combinations of binder and plasticizers.
shear thickening for higher shear rate values). As expected, the suspensions containing the higher molecular weight plasticizer (P400), exhibit higher viscosity values for a given added amount. Again, it is clear that viscosity accentuates with increasing the added amount of binder and decreases with increasing added amounts of plasticizers. From these results it is possible to extract information about the casting speed more adequate to these suspensions. As referred above, during tape casting the slurry undergoes a certain structural decomposition depending mostly on the casting velocity and gap height. According to Bitterlich and Heinrich [8], the casting speed should not be at the lower shear rate values, because in this region small variation of shear rate results in higher changes in viscosity, and therefore in changes of the fluid mechanic conditions in the casting head. The figures show that for all the suspensions the shear rate must be in the range of about 10–40 s−1 which corresponds to casting speeds of 2–8 cm s−1 for a gap height of 2 mm, therefore, comprising the casting speed of 3 cm s−1 used in the present work.
Fig. 5. Viscosity vs. time of the aqueous suspensions containing 50 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) and selected combinations of binder and plasticizers.
3.4. Viscoelastic properties of the suspensions Dynamic measurements were also performed in order to access information about the internal structure of the suspensions. Fig. 6a shows the dependence of the storage modulus G, and loss modulus G on shear stress for suspensions
3.3. Structural regeneration of the suspensions To check if the slurries exhibit any time dependent effects a shear rate step function was preset. First, the suspensions were submitted to a 50 s−1 shear rate during 60 s and then the shear rate was immediately decreased to 1 s−1 . Fig. 5 presents the viscosity versus time of the suspensions containing 50 vol.% solids and the same combinations of binder and plasticizers already used (Fig. 4). The viscosity increases continuously after the step decrease of the shear rate, attaining a steady viscosity value after few seconds. These results were confirmed by additional experiments and no evident hysterese effects in the viscosity curves could be detected after increasing and decreasing shear rates. From these results, it can be concluded that the higher molecular weight plasticizer (P400) enables a faster structural regeneration when compared with the other one (P200), and that the binder also plays an active role in regenerating the structure, which accentuates with increasing the added amount.
Fig. 6. G and G vs. shear stress of the aqueous suspensions containing 50 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) with different additions of processing additives: (a) 10 wt.% binder and 5 wt.% of each of the two plasticizers (P200 and P400); (b) 15 wt.% binder and 10 wt.% of each of the two plasticizers (P200 and P400).
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
211
Table 1 Green tape thickness and flexibility results for tapes obtained from suspensions containing 50 vol.% solids and different combinations of binder and plasticizers Sample
BI (wt.%)
P200 (wt.%)
P400 (wt.%)
Green thickness (mm)
Flexibilitya
A B C D
10 10 15 15
5 – 10 –
– 5 – 10
1.67 1.38 1.37 1.55
X X V V
Samples code: A 10 wt.% binder + 5 wt.% P200; B 10 wt.% binder + 5 wt.% P400; C 15 wt.% binder + 10 wt.% P200; D 15 wt.% binder + 10 wt.% P400. a X: non-flexible; V: flexible.
Fig. 7. Pictures of the green tapes obtained by tape casting from the aqueous suspensions.
containing 50 vol.% solids and combinations of 10 wt.% of binder + 5 wt.% of each plasticizer (P200 or P400) and Fig. 6b shows the same rheological parameters for the suspensions but with combinations of 15 wt.% of binder + 10 wt.% of each plasticizer (P200 or P400).
At low shear stresses, the suspensions exhibit a gellike character (G > G ). Hence, the elastic behaviour dominates over the viscous behaviour and the structure shows some rigidity. The crossover point where G = G (tan δ = G /G = 1) occurs at different shear stress values according to the added amounts of the processing additives, namely at about 0.4 Pa for the lower content of additives (Fig. 6a) and at about 1 Pa for the higher added amounts of additives (Fig. 6b). After the crossover point, G > G and the viscous character predominates over the elastic one. This corresponds to the structural decomposition referred above where the internal structure of the suspension is destroyed. According to Bitterlich and Heinrich [8], above a critical shear stress the weak attractive forces between the powder particles and or the binder and plasticizer are broken down by the external shear stress, which destroys the internal network. These are important characteristics of tape casting slurries, which must spread over the carrier to accomplish the process. This means that a shear
Fig. 8. Microstructures of the tapes prepared from aqueous suspensions containing 50 vol.% of total solids loading (95 wt.% AlN-SHS + 5 wt.% CaF2 ) and different selected combinations of binders and plasticizers: (1) 10 wt.% binder and 5 wt.% P200; (2) 10 wt.% binder and 5 wt.% P400; (3) 15 wt.% binder and 10 wt.% of P200; (4) 15 wt.% binder and 10 wt.%; 15 wt.% binder and 10 wt.% P400.
212
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213
stress value greater than the critical shear stress must be applied. 3.5. Properties of the green tapes Table 1 presents the results of thickness and flexibility for the tapes obtained from the suspensions with 50 vol.% solids loading with different amounts of the binder and of the two plasticizers (P200 and P400). For all the suspensions tested, it was possible to obtain non-cracked green tapes with a thickness higher than 1 mm. The tapes produced from suspensions with 10 wt.% of added binder were not flexible enough to pass the flexibility test mentioned above, however they could be easily handled. The minimum amount of binder required to produce flexible tapes with thickness as high as 1.5 mm was 15 wt.%. All the slips produced good, smooth and uniform tapes without cracks as shown in Fig. 7. Fig. 8 shows the microstructures of the tapes obtained from the suspensions containing 50 vol.% and different combinations of binder and plasticizers: (1) 10 wt.% binder + 5 wt.% P200; (2) 10 wt.% binder + 5 wt.% P400; (3) 15 wt.% binder + 10 wt.% P200 and (4) 15 wt.% binder + 10 wt.% P400. It can be observed that all the tapes present homogeneous and uniform microstructures.
4. Conclusions The results presented in this work demonstrate the feasibility of producing AlN-based green tapes with thicknesses up to 1.5 mm from aqueous slurries. This possibility will have a tremendous impact in the industry at all levels, since the organic and volatile solvents can be completely replaced by the low cost and environment friendly water. Three essential key points need to be overcome for achieving this goal: 1. Protecting the AlN particles against the strong hydrolysis reactions, what has been achieved by adding H3 PO4 . 2. Find a suitable dispersing agent compatible with the protecting agent, what has been accomplished using a carboxylic anionic surfactant to increase the surface charge density and dispersing the AlN powder + the sintering aid. 3. Adding suitable combinations of an aqueous acrylic emulsion binder and plasticizers to set the required rheological properties of slurries to the specific demands of the tape casting process. Furthermore, it has been shown that rheology is an appropriate and important tool to evaluate the suitability of the slurries for the tape casting process and to decide about the necessary modifications in order to match the required rheological properties of slurries. The desired shear thinning behaviour and the smoothness of the viscosity curves are enhanced with added binder and plasticizers, although they impart opposite effects on viscosity level.
Acknowledgements The first author wishes to thank Fundac¸a˜ o para a Ciˆencia e a Tecnologia for the fellowship grant SFRH/BD/8754/2002. The authors are also in debt to PRAI-Centro—Programa Regional de Acc¸o˜ es Inovadoras do Centro de Portugal, for the financial support under the Project: “S´ıntese por combust˜ao de nitreto de alum´ınio (AlN) e processamento de substratos de AlN por tape casting em meio aquoso”. References [1] A. Roosen, Basic requirements for tape casting of ceramic powders, Ceram. Trans. 1 (1998) 675–692. [2] A.I.Y. Tok, F.Y.C. Boey, K.A. Khor, Tape casting of high dielectric ceramic composite substrates for microelectronics application, J. Mater. Process. Technol. 89–90 (1999) 508–512. [3] F. Doreau, G. Tari, C. Chartier, J.M.F. Ferreira, Processing of aqueous tape casting of alumina with acrylic emulsion binders, J. Eur. Ceram. Soc. 18 (1998) 311–321. [4] A. Kristoffersson, R. Lapsin, C. Galassi, Study of interactions between polyelectrolyte dispersants, alumina and latex binders by rheological characterisation, J. Eur. Ceram. Soc. 18 (1998) 2133–2140. [5] A. Kristoffersson, E. Roncari, C. Galassi, Comparison of different binders for water-based tape casting of alumina, J. Eur. Ceram. Soc. 18 (1998) 2123–2131. [6] Z. Yuping, J. Dongliang, P. Greil, Tape casting of aqueous Al2 O3 slurries, J. Eur. Ceram. Soc. 20 (2000) 1691–1697. [7] A. Kristofferson, Elis Carlstrom, Tape casting of alumina in water with an acrylic latex binder, J. Eur. Ceram. Soc. 17 (1997) 289– 297. [8] Bernd Bitterlich, Jurgen G. Heinrich, Aqueous tape casting of silicon nitride, J. Eur. Ceram. Soc. 22 (13) (2002) 2427–2434. [9] K. Krnel, T. Kosmac, Protection of AlN powder against hydrolysis using aluminium dihydrogen phosphate, J. Eur. Ceram. Soc. 21 (2001) 2075–2079. [10] T. Kosmac, K. Krnel, K. Kos, “Process for the protection of AlN powder against hydrolysis”, International Patent No. WO 99/12850, 18.03 (1999). [11] Y. Shimizu, J. Hatano, T. Hyodo, M. Egashira, Ion-exchange loading of yttrium acetate as a sintering aid on aluminium nitride powder via aqueous processing, J. Am. Ceram. Soc. 83 (11) (2000) 2793– 2797. [12] Y. Shimizu, K. Kawanabe, Y. Taky, Y. Takao, M. Egashira, AlN ceramics prepared by aqueous colloidal processing, in: H. Hausner, G.L. Messing, S. Hirano (Eds.), Ceramic Processing Science and Technology, Ceramic Transactions, vol. 51, American Ceramic Society, Westerville, OH, 1995, pp. 403–407. [13] M. Oliveira, S. Olhero, J. Rocha, J.M.F. Ferreira, Controlling hydrolysis and dispersion of AlN powders in aqueous media, J. Colloid Interface Sci. 261 (2003) 456–463. [14] B. Bitterlich, C. Lutz, A. Roosen, Rheological characterization of water-based slurries for the tape casting process, Ceram. Int. 28 (2002) 675–683. [15] M. Oliveira, S. Olhero, J.M.F. Ferreira, “M´etodos de passivac¸a˜ o do AlN em relac¸a˜ o a` hidr´olise, de processamento coloidal de cerˆamicos a` base de AlN em meio aquoso e de granulac¸a˜ o de p´os a partir dessas suspens˜oes”, Portuguese Patent No. 102852 (2002). [16] R. Greenwood, E. Roncari, C. Galassi, Preparation of concentrated aqueous alumina suspensions for tape casting, J. Eur. Ceram. Soc. 17 (12) (1997) 1393–1401. [17] Jennifer Lewis, Colloidal Processing of Ceramics, J. Am. Ceram. Soc. 83 (10) (2000) 2341–2359.
S.M. Olhero, J.M.F. Ferreira / Journal of Materials Processing Technology 169 (2005) 206–213 [18] M.D. Sacks, Properties of silica suspensions and cast bodies, Am. Ceram. Soc. Bull. 63 (1984) 1510–1515. [19] L. Bergstrom, Rheology of concentrated suspensions, in: R.J. Pugh, L. Bergstrom (Eds.), Surface and Colloid Chemistry in Advanced Ceramic Processing, Surfactant Science Series, vol. 51, Marcel Dekker, New York, 1994, p. 193.
213
[20] H.A. Barnes, Dilatancy in suspensions of non aggregating solid particles dispersed in Newtonian liquids, J. Rheol. 33 (1989) 329– 366. [21] Renli Fu, Xin Xu, Jos´e M.F. Ferreira, Highly crystalline AlN particles synthesized by SHS, Mater. Lett., accepted for publication.