J Mater sci(2006041:8093-8100 DoI10.1007/s10853-006-0646-y The preparation of LPs SiC-fibre -reinforced sic ceramics using electrophoretic deposition S. Novak·K. Mejak·G. Drazic Received: 12 January 2006/Accepted: 6 July 2006/Published online: 28 October 2006 o Springer Science+Business Media, LLC 2006 Abstract With the aim of determining the possibility main role of the fibres is to cause the deflection of any of producing Sic-based ceramic-matrix composi initiated crack and hence to dissipate its energy using the electrophoretic deposition(EPD)technique, order to protect the material from catastrophic frac the effect of the composition of Sic-based suspensions ture. Due to their outstanding high-temperature on the deposition was studied. Ethanol suspensions of behaviour, Sic-fibre-reinforced Sic ceramics are used two different grades of Sic powders, with and without in various applications that are subject to extreme the addition of a sintering aid, were used for depositing conditions, for example, in aerospace. Lately, Sicf/Sic on steel electrodes or on SiC-fibres. The pH of the composites have also been considered to be one of the suspensions, the solids loading and the particle size best candidates for use in future fusion reactors, where were shown to have a strong influence on the deposi- in addition to the high-temperature mechanical and tion process and on the properties of the fresh deposits. physical properties of the SiC, a low neutron activation The overall results demonstrate that by using appro- is one of the critical requirements for the material 3, priate conditions for the EPD, a firm SiC-based deposit 4. At present, the main limitation when it comes to can be collected at the Sic-fibres and after suitable producing a material to meet the requirements of this ermal treatment, a pore-free Sic-matrix, well highly demanding application is a lack of suitable adhered to the sic fibres can be achieved manufacturing technique. Various methods of Sic fibre-preform infiltration have been considered, such as chemical vapour infiltration(CVi), polymer infiltration and pyrolysis(PIp), nanopowder infiltration and tran sient eutectoid(NITE)as well as the simple immersion of ceramic slurries and combinations of these methods Introduction 5-10] From among these methods the most attractive for producing a material for fusion applications Continuous-fibre-reinforced ceramics are a group of appears to be CVI; however, this is an extremely slow materials with significantly better mechanical proper- and costly process and seems to be unable to produce ties than monolithic or particulate composites. In the material without open pores, and consequently one of past decade various composites with fibres were the important requirements-gas impermeability-is developed in which continuous fibres of carbon, difficult to achieve. In other processes, the high alumina or silicon carbide were used to provide the temperatures needed to produce sufficient density reinforcement for various ceramic matrixes [1, 2]. The and a large shrinkage of the material present serious drawbacks S. Novak(國)·K. Mejak.G. Drazic In our investigations we looked at the potential of Department of Nanostructured Materials, Jozef Stefan electrophoretic deposition(EPD)as a possible tech Institute, Jamova 39, Ljubljana SI-1000, Slovenia nique for the infiltration of a Sic-fibre preform to e-mail: sasa. novak@ijs produce a SiC/SiC composite, providing that a suitable 2 Springer
Abstract With the aim of determining the possibility of producing SiC-based ceramic-matrix composites using the electrophoretic deposition (EPD) technique, the effect of the composition of SiC-based suspensions on the deposition was studied. Ethanol suspensions of two different grades of SiC powders, with and without the addition of a sintering aid, were used for depositing on steel electrodes or on SiC-fibres. The pH of the suspensions, the solids loading and the particle size were shown to have a strong influence on the deposition process and on the properties of the fresh deposits. The overall results demonstrate that by using appropriate conditions for the EPD, a firm SiC-based deposit can be collected at the SiC-fibres and after suitable thermal treatment, a pore-free SiC-matrix, well adhered to the SiC fibres, can be achieved. Introduction Continuous-fibre-reinforced ceramics are a group of materials with significantly better mechanical properties than monolithic or particulate composites. In the past decade various composites with fibres were developed in which continuous fibres of carbon, alumina or silicon carbide were used to provide the reinforcement for various ceramic matrixes [1, 2]. The main role of the fibres is to cause the deflection of any initiated crack and hence to dissipate its energy in order to protect the material from catastrophic fracture. Due to their outstanding high-temperature behaviour, SiC-fibre-reinforced SiC ceramics are used in various applications that are subject to extreme conditions, for example, in aerospace. Lately, SiCf/SiC composites have also been considered to be one of the best candidates for use in future fusion reactors, where, in addition to the high-temperature mechanical and physical properties of the SiC, a low neutron activation is one of the critical requirements for the material [3, 4]. At present, the main limitation when it comes to producing a material to meet the requirements of this highly demanding application is a lack of suitable manufacturing technique. Various methods of SiC- fibre-preform infiltration have been considered, such as chemical vapour infiltration (CVI), polymer infiltration and pyrolysis (PIP), nanopowder infiltration and transient eutectoid (NITE) as well as the simple immersion of ceramic slurries and combinations of these methods [5–10]. From among these methods the most attractive for producing a material for fusion applications appears to be CVI; however, this is an extremely slow and costly process and seems to be unable to produce material without open pores, and consequently one of the important requirements—gas impermeability—is difficult to achieve. In other processes, the high temperatures needed to produce sufficient density and a large shrinkage of the material present serious drawbacks. In our investigations we looked at the potential of electrophoretic deposition (EPD) as a possible technique for the infiltration of a SiC-fibre preform to produce a SiCf/SiC composite, providing that a suitable S. Novak (&) K. Mejak G. Drazˇic´ Department of Nanostructured Materials, Jozef Stefan Institute, Jamova 39, Ljubljana SI-1000, Slovenia e-mail: sasa.novak@ijs.si J Mater Sci (2006) 41:8093–8100 DOI 10.1007/s10853-006-0646-y 123 The preparation of LPS SiC-fibre-reinforced SiC ceramics using electrophoretic deposition S. Novak Æ K. Mejak Æ G. Drazˇic´ Received: 12 January 2006 / Accepted: 6 July 2006 / Published online: 28 October 2006 Springer Science+Business Media, LLC 2006
8094 J Mater Sci(2006)41:8093-8100 dditive for densification is used. Various reports show the deposit were monitored continuously. The densi- that EPD has been successful for preparing thick and ties of the fresh(wet) deposits were calculated from thin films and various free-standing materials [11-13, the buoyancy in the suspension, and the solids content for example, bulk SiC-based ceramics with boron, in the deposits was determined by 4s were fir measuring the arbon, Al2O or AIN as sintering additives [14-16. weight change during drying. The deposits were fired at Another example of using EPD is the infiltration of temperatures from 1300 to 1500C in an Ar atmo- [17, 18]. For the material to be used in the first wall of a TEM and ED ered samples were inspected by SEM, lumina or carbon fibres woven with a ceramic matrix sphere. The sint fusion reactor suitable low-neutron -activation sinter ing additives must be used. Such additives would allow low-temperature densification without any detrimental Results and discussion effects on the fibres This excludes the conventional additives for SiC, such as B, C, and additives from the Figure 1 shows the zeta-potential for the used powders system Al2 O3(AIN)-Y2O3 that require high sintering in an ethanol suspension as a function of pH. It is clear temperatures, typically above 1800C. In our study that the point of zero charge(PzC)for the um-Sic two grades of the Sic powder was used and powder appears at about pH 6.0. At lower values the ed as the additive: this particles carry a relatively high positive net-charge, enables densification below the temperature limit of while the zP in the alkaline region are negative. The 1500 .C at close-to-zero shrinkage. The conditions for addition of aluminium phosphate(AP) shifted the the EPD of a SiC-based matrix material on SiC-fibres PZc to a lower value, suggesting that it is adsorbed on were investigated the surface of the Sic particles. The measured ZP values for the nano-Sic powder were much lower than those for the um-SiC across the whole range of PH Experimental The PZC appears at a lower pH value, suggesting that the fine Sic particles have an oxidised surface layer Two grades of Sic powder were used in this investi- Figures 2a and b show the morphology of the um-SIC gation:"H-SiC", a submicron B-SiC BF-12(H Starck, and nano-SiC powders, while Figs 2c and d show Goslar, Germany) with an average particle size of HRTEM images of the oxygen-containing amorphous 0.5 um; and"nano-SiC", a powder with an average surface layers on the particles of both powders. In particle size of 50 nm(Hefei KiIn Nanom Technol. Figs 2e and f EDXS spectra of the amorphous layer Dev. Co. Ltd, China). Aluminium phosphate, "AP", and bulk Sic particle of um-SiC are shown. It is (TKI doo, Slovenia)was used as an additive for the demonstrated, that the thickness of the layers is similar in a ball-mill for at leae e pe s of the powder and the for both powders(1-2 nm). The XPS analysis revealed 2 h. The characteri istics of the powder, while in the nano-SiC powder, beside oxyca particles in the suspensions were quantified by mea- bides also SiOz was found suring the zeta-potential(ZP) using a ZetaPals zeta- meter(Brookhaven, USA). The operational ph was adjusted using HCl, citric acid or NH,OH, and measured with a pH meter(Metron Ltd, Switzerland) For simplicity, the measured values are designated as 00四 Hin this investigation. Surface chemistry of the powders was analysed by X-ray troscopy(XPS) using TFA spectrometer(Pysical N Electronics, USA) The EPD experiments were performed at a constant voltage of 60 V. The electrodes were either square stainless steel plates, 2 x 2 cm2, placed vertically at distance of 2 cm or a bundle of the sic fibres 0123456789101112 Nippon Carbon Co, Ltd). The solids conte uspensions for the EPD varied from Fig. 1 Zeta-potential of um-SiC (without and with addition sition the electrical current and the weight change of function of pl phate)and nano-SiC powders in eth (4.3 vol %)to 70 wt%(37.3 vol %) During the depo- aluminium pho 2 Springer
additive for densification is used. Various reports show that EPD has been successful for preparing thick and thin films and various free-standing materials [11–13], for example, bulk SiC-based ceramics with boron, carbon, Al2O3 or AlN as sintering additives [14–16]. Another example of using EPD is the infiltration of alumina or carbon fibres woven with a ceramic matrix [17, 18]. For the material to be used in the first wall of a fusion reactor, suitable low-neutron-activation sintering additives must be used. Such additives would allow low-temperature densification without any detrimental effects on the fibres. This excludes the conventional additives for SiC, such as B, C, and additives from the system Al2O3(AlN)-Y2O3 that require high sintering temperatures, typically above 1800 C. In our study two grades of the SiC powder was used and a phosphate glass was employed as the additive; this enables densification below the temperature limit of 1500 C at close-to-zero shrinkage. The conditions for the EPD of a SiC-based matrix material on SiC-fibres were investigated. Experimental Two grades of SiC powder were used in this investigation: ‘‘l-SiC’’, a submicron b-SiC BF-12 (H. Starck, Goslar, Germany) with an average particle size of 0.5 lm; and ‘‘nano-SiC’’, a powder with an average particle size of 50 nm (Hefei Kiln Nanom.Technol. Dev. Co. Ltd, China). Aluminium phosphate, ‘‘AP’’, (TKI doo, Slovenia) was used as an additive for the densification. The suspensions of the powder and the additive in ethanol were prepared by homogenisation in a ball-mill for at least 2 h. The characteristics of the particles in the suspensions were quantified by measuring the zeta-potential (ZP) using a ZetaPals zetameter (Brookhaven, USA). The operational pH was adjusted using HCl, citric acid or NH4OH, and measured with a pH meter (Metron Ltd, Switzerland). For simplicity, the measured values are designated as ‘‘pH’’ in this investigation. Surface chemistry of the powders was analysed by X-ray photoelectron spectroscopy (XPS) using TFA spectrometer (Pysical Electronics, USA). The EPD experiments were performed at a constant voltage of 60 V. The electrodes were either square stainless steel plates, 2 · 2 cm2 , placed vertically at a distance of 2 cm, or a bundle of the SiC fibres (Nicalon, Nippon Carbon Co., Ltd). The solids content in the suspensions for the EPD varied from 15 wt.% (4.3 vol.%) to 70 wt.% (37.3 vol.%). During the deposition the electrical current and the weight change of the deposit were monitored continuously. The densities of the fresh (wet) deposits were calculated from the buoyancy in the suspension, and the solids content in the deposits was determined by measuring the weight change during drying. The deposits were fired at temperatures from 1300 to 1500 C in an Ar atmosphere. The sintered samples were inspected by SEM, TEM and EDS. Results and discussion Figure 1 shows the zeta-potential for the used powders in an ethanol suspension as a function of pH. It is clear that the point of zero charge (PZC) for the lm-SiC powder appears at about pH 6.0. At lower values the particles carry a relatively high positive net-charge, while the ZP in the alkaline region are negative. The addition of aluminium phosphate (AP) shifted the PZC to a lower value, suggesting that it is adsorbed on the surface of the SiC particles. The measured ZP values for the nano-SiC powder were much lower than those for the lm-SiC across the whole range of pH. The PZC appears at a lower pH value, suggesting that the fine SiC particles have an oxidised surface layer. Figures 2a and b show the morphology of the lm-SiC and nano-SiC powders, while Figs. 2c and d show HRTEM images of the oxygen-containing amorphous surface layers on the particles of both powders. In Figs. 2e and f EDXS spectra of the amorphous layer and bulk SiC particle of lm-SiC are shown. It is demonstrated, that the thickness of the layers is similar for both powders (1–2 nm). The XPS analysis revealed presence of oxycarbides at the surface of the lm-SiC powder, while in the nano-SiC powder, beside oxycarbides also SiO2 was found. -60 -40 -20 0 20 40 60 pH ZP (mV) µm-SiC µm-SiC + AP nano-SiC 0 1 2 3 4 5 6 7 8 9 10 11 12 Fig. 1 Zeta-potential of lm-SiC (without and with addition of aluminium phosphate) and nano-SiC powders in ethanol as a function of pH 8094 J Mater Sci (2006) 41:8093–8100 123
J Mater sci(2006041:8093-8100 8095 Fig 2(a)-SEM micrograph of um-SiC, (b)-TEM powder, (c),(d)-HRTEM particle in um-SiC and nano- SiC powders;(e),(f-EDX and bulk particle in um-SiC 650 nm 2n In order to better understand the process, suspen- deposit (i.e, before drying) was found to contain only sions of um-SiC powder were deposited without any 59.1 wt %(26.8 vol %)solids( Fig 3a). Due to the low sintering additive on steel electrodes. The pH values solids content, such a visco-elastic deposit is not able to were selected in accordance with the characteristic resist its own weight: when the force of gravity points in the ZP VS pH graph in Fig. 1: the natural pH overcomes the weak interparticle forces, the viscous of the suspension (pH 5.0)and the pH values where flow behaviour predominates over the elastic beha the ZP appears highly positive or negative(pH 2.8 and iour and the deposit returns to the state of a viscous 11, respectively). The solids content in the starting liquid. uspension was 50 wt. and the density was 1.26 g/ As expected, at pH values close to the Pzc no cm. In another set of experiments, the solids content deposit was formed. At pH ll the deposit formed in the suspensions was varied from 30 to 70 wt %, the anode in accordance with the ZP vS. pH keeping the pH value constant. The solids content in relationship presented in Fig. 1. It containe the as-formed(wet)deposits, the deposition rate and 67.8 wt %(34.9 vol %) of solids(Fig 3a), which is the current change during the depositions are pre- higher than that obtained from the suspension sented in Figs. 3a-f natural pH. The rate of deposition was also higher The EPD of the suspension with the natural pH (i.e, than at pH 2.8 and pH 5(Fig 3c): this could be pH 5.0) resulted in a very loose deposit being formed related to the larger amount of electrolyte needed for at the cathode. After approximately 10 min, when its the pH adjustment, which is also reflected in the weight exceeded approximately 10 g, it slipped from higher initial current(Fig. 3e). Resulting lower solids the electrode. Obviously, under the given conditions, content and a lower density of the deposit formed the initially low-viscosity suspension flocculated at the pH ll can be explained by the lower ZP than at electrode and hence its viscosity increased. The fresh pH 2.8
In order to better understand the process, suspensions of lm-SiC powder were deposited without any sintering additive on steel electrodes. The pH values were selected in accordance with the characteristic points in the ZP vs. pH graph in Fig. 1: the natural pH of the suspension (pH 5.0) and the pH values where the ZP appears highly positive or negative (pH 2.8 and 11, respectively). The solids content in the starting suspension was 50 wt.% and the density was 1.26 g/ cm3 . In another set of experiments, the solids content in the suspensions was varied from 30 to 70 wt.%, keeping the pH value constant. The solids content in the as-formed (wet) deposits, the deposition rate and the current change during the depositions are presented in Figs. 3a–f. The EPD of the suspension with the natural pH (i.e., pH 5.0) resulted in a very loose deposit being formed at the cathode. After approximately 10 min, when its weight exceeded approximately 10 g, it slipped from the electrode. Obviously, under the given conditions, the initially low-viscosity suspension flocculated at the electrode and hence its viscosity increased. The fresh deposit (i.e., before drying) was found to contain only 59.1 wt.% (26.8 vol.%) solids (Fig. 3a). Due to the low solids content, such a visco-elastic deposit is not able to resist its own weight: when the force of gravity overcomes the weak interparticle forces, the viscous- flow behaviour predominates over the elastic behaviour and the deposit returns to the state of a viscous liquid. As expected, at pH values close to the PZC no deposit was formed. At pH 11 the deposit formed at the anode in accordance with the ZP vs. pH relationship presented in Fig. 1. It contained 67.8 wt.% (34.9 vol.%) of solids (Fig. 3a), which is higher than that obtained from the suspension at natural pH. The rate of deposition was also higher than at pH 2.8 and pH 5 (Fig. 3c); this could be related to the larger amount of electrolyte needed for the pH adjustment, which is also reflected in the higher initial current (Fig. 3e). Resulting lower solids content and a lower density of the deposit formed at pH 11 can be explained by the lower ZP than at pH 2.8. Fig. 2 (a) - SEM micrograph of lm-SiC, (b) - TEM micrograph of nano-SiC powder, (c), (d) - HRTEM micrographs of the amorphous layer on SiC particle in lm-SiC and nanoSiC powders; (e), (f) - EDX spectra of the surface layer and bulk particle in lm -SiC sample J Mater Sci (2006) 41:8093–8100 8095 123
8096 J Mater Sci(2006)41:8093-8100 Fig 3 Solids content in fresh b90 deposit of um-siC as a solids content in pH 2.8 function of(a)-PH and(b) suspensions: 50 wt% solids content in starting ns;deposition rate as a function of (c)-pH and (d)-solids content in the uspensions;electrical current change during the deposition as a function of (e)-pH and (f-solids content(60V,steel starting electrode; suspension in(a), (e),(e): 50 wt. solids: suspensions in(b),(d),(e) 56789101112 60 Solids content in suspension(wt c25 E巴E 23456789101112 5060 solids content (%) o30% o pH5 △60% 70% 可 time [min time [min From Fig 3c it is obvious that the rate of deposition In the next set of experiments, the effect of the from the suspension with pH 2.8 was slightly lower solids content in the starting suspension was verified than that at the natural pH 5. The deposit formed at for the deposition of suspensions with 30-70 wt. of the cathode and appeared to be much firmer in the um-SiC powder at a constant pH value of 2.8 comparison to the deposit formed at the natural ph. Figures 3b and d reveal that the density of the deposits Its density was 1.78 g/cm, which is significantly higher as well as the deposition rate increased with ar than the starting suspension, i.e., 1.26 g/cm. The solids increase in the powder content in the starting suspen content in the deposit was determined to be 72.3 wt % sion On the other hand, the effect of the solids content i.e., 40.1 vol %(Fig. 3a). on the initial current in the suspensions, reflecting their 2 Springer
From Fig. 3c it is obvious that the rate of deposition from the suspension with pH 2.8 was slightly lower than that at the natural pH 5. The deposit formed at the cathode and appeared to be much firmer in comparison to the deposit formed at the natural pH. Its density was 1.78 g/cm3 , which is significantly higher than the starting suspension, i.e., 1.26 g/cm3 . The solids content in the deposit was determined to be 72.3 wt.%, i.e., 40.1 vol.% (Fig. 3a). In the next set of experiments, the effect of the solids content in the starting suspension was verified for the deposition of suspensions with 30–70 wt.% of the lm-SiC powder at a constant pH value of 2.8. Figures 3b and d reveal that the density of the deposits as well as the deposition rate increased with an increase in the powder content in the starting suspension. On the other hand, the effect of the solids content on the initial current in the suspensions, reflecting their 30 40 50 60 70 80 90 1 2 3 4 5 7 9 10 11 12 pH Solids content in deposit (%) solids content in suspensions: 50 wt. % starting suspensions 30 40 50 60 70 80 90 30 40 50 60 70 80 Solids content in suspension (wt.%) starting suspensions pH 2.8 Solis content in deposit (%) 0 0.5 1 1.5 2 2.5 1 2 3 4 5 7 9 10 11 12 pH Deposition rate (g/min) 0 0.5 1 1.5 2 2.5 0 10 20 30 40 50 60 70 80 solids content (%) Deposition rate (g/min) 6 8 6 8 0 1 2 3 4 5 0 5 10 15 time [min] pH2.8 pH5 pH11 current (mA) 0 1 2 3 4 5 0 5 10 15 time [min] 30 % 50 % 60 % 70 % current (mA) a c e f d b Fig. 3 Solids content in fresh deposit of lm-SiC as a function of (a) - pH and (b) - solids content in starting suspensions; deposition rate as a function of (c) - pH and (d) - solids content in the suspensions; electrical current change during the deposition as a function of (e)- pH and (f) - solids content (60 V, steel electrode; suspension in (a), (c), (e): 50 wt. % solids; suspensions in (b), (d), (e): pH 2.8) 8096 J Mater Sci (2006) 41:8093–8100 123
J Mater sci(2006041:8093-8100 conductivity, was negligible(Fig. 3f). This suggests a 18 minor contribution of the charged um-SiC particles to H17 wt the overall conductivity. Furthermore, no current drop during the deposition was observed 8wt.% The above results reveal that the consistency of the deposit, evaluated in this study on the basis of its density and the solids content, is governed primarily by the surface charge on the particles. The deposits with the highest density were obtained by depositing from the acidic suspension in which the Sic particles carried F limited value by keeping the concentration of ions low c the highest charge, while the current was kept at a (small electrolyte additions ). The density as well as the deposition rate increased with the solids content in the starting suspensions. It should also be pointed out that a minor current drop during the epd was observed when a deposit with high solids content (72.3 wt % was formed, i.e. at pH 2.8. In contrast, a significant current drop was observed at pH 5 and 11 where deposits with lower densities were formed. This is not time [min] in agreement with the assumption that the current drop caused by the inhibited transport of charge carriers Fig. 4 Electrical current during the deposition of nano-SiC ue to the accumulation of the deposit at the electrode. suspension as a function of solids content in the starting Further,according to the results presented in Figs. 3e suspensions(60 V, steel electrode) and f, the major charge carrier in the analysed suspensions was the ions, while the charged particles contribute to the conductivity of the suspensions make a minor contribution hence. it is reasonable to however a lower concentration of silanol ions in the assume that a large particle separation in the deposits diluted suspensions might also have the same effect. containing up to 81 wt. solids, i.e. 52 vol %(see Furthermore, during the deposition a significant cur- Fig 3b, f)allows the ions, as the proposed major rent drop was also observed harge carriers, to move towards the electrode through The addition of the AP changed not only the zP vs the deposit. pH relationship, but also the course of the deposition. Additional EPD experiments were performed with At the natural pH of the um-SiC suspension with the the nano-SiC powder. Due to a high apparent viscosity AP addition, i.e. pH 2, the particles carry a low of the suspensions, resulting from a high speci ositive net charge, as presented in Fig. 1. and there surface area of the powder, a suspension with a fore they collect at the cathode. Figure 5 shows that maximum of 17 wt. was used for the EPD. As a the addition of the AP significantly increased the initial result, in accordance with the above-presented depen- current, most probably due to a high concentration of dence of deposition rate on the solids content in the free phosphate ions in the suspensions. The deposition suspensions, the deposition rate was only 0. 25 g/min. rate was 0.3 g/min, which was much lower than for the The deposit was very loose, and after 10 min of suspension with a similar ph but without the additive deposition it slipped from the electrode. As presented (see Fig. 3c). As presented in Fig. 5, increasing the pl in Fig 4, the initial current was much higher than that to 10, where the ZP was observed to have high observed for all the analysed um-SiC suspensions This negative value, resulted in a decreased initial current could be explained by the presence of surface silanol and an increase in the deposition rate to 0.8 g/min. The groups losing the proton in water containing solutions decreased conductivity suggests a decreased concen- that reflects also in the natural acidity of the suspension tration of free ions due to a chemical reaction of (pH <2). Furthermore, in contrast to the um-Sic phosphate ions with the added ammonium hydroxide powder, where a negligible effect of the particles In contrast, the deposition from the suspension of concentration on the conductivity was observed (see nano-Sic powder with the AP additive was only Fig 3f), the dilution of the suspension caused a successful in acidic suspensions. However, even in this decrease in the initial current, suggesting that the case the deposition was slow, and the deposits small and numerous nano-SiC particles might also tained bubbles. The initial current was very 2 Spring
conductivity, was negligible (Fig. 3f). This suggests a minor contribution of the charged lm-SiC particles to the overall conductivity. Furthermore, no current drop during the deposition was observed. The above results reveal that the consistency of the deposit, evaluated in this study on the basis of its density and the solids content, is governed primarily by the surface charge on the particles. The deposits with the highest density were obtained by depositing from the acidic suspension in which the SiC particles carried the highest charge, while the current was kept at a limited value by keeping the concentration of ions low (small electrolyte additions). The density as well as the deposition rate increased with the solids content in the starting suspensions. It should also be pointed out that a minor current drop during the EPD was observed when a deposit with high solids content (72.3 wt.%) was formed, i.e. at pH 2.8. In contrast, a significant current drop was observed at pH 5 and 11 where deposits with lower densities were formed. This is not in agreement with the assumption that the current drop is caused by the inhibited transport of charge carriers due to the accumulation of the deposit at the electrode. Further, according to the results presented in Figs. 3e and f, the major charge carrier in the analysed suspensions was the ions, while the charged particles make a minor contribution. Hence, it is reasonable to assume that a large particle separation in the deposits containing up to 81 wt.% solids, i.e. 52 vol.% (see Fig. 3b, f) allows the ions, as the proposed major charge carriers, to move towards the electrode through the deposit. Additional EPD experiments were performed with the nano-SiC powder. Due to a high apparent viscosity of the suspensions, resulting from a high specific surface area of the powder, a suspension with a maximum of 17 wt.% was used for the EPD. As a result, in accordance with the above-presented dependence of deposition rate on the solids content in the suspensions, the deposition rate was only 0.25 g/min. The deposit was very loose, and after 10 min of deposition it slipped from the electrode. As presented in Fig. 4, the initial current was much higher than that observed for all the analysed lm-SiC suspensions. This could be explained by the presence of surface silanol groups losing the proton in water containing solutions that reflects also in the natural acidity of the suspension (pH < 2). Furthermore, in contrast to the lm-SiC powder, where a negligible effect of the particles concentration on the conductivity was observed (see Fig. 3f), the dilution of the suspension caused a decrease in the initial current, suggesting that the small and numerous nano-SiC particles might also contribute to the conductivity of the suspensions. However, a lower concentration of silanol ions in the diluted suspensions might also have the same effect. Furthermore, during the deposition a significant current drop was also observed. The addition of the AP changed not only the ZP vs. pH relationship, but also the course of the deposition. At the natural pH of the lm-SiC suspension with the AP addition, i.e. pH 2, the particles carry a low positive net charge, as presented in Fig. 1, and therefore they collect at the cathode. Figure 5 shows that the addition of the AP significantly increased the initial current, most probably due to a high concentration of free phosphate ions in the suspensions. The deposition rate was 0.3 g/min, which was much lower than for the suspension with a similar pH but without the additive (see Fig. 3c). As presented in Fig. 5, increasing the pH to 10, where the ZP was observed to have high negative value, resulted in a decreased initial current and an increase in the deposition rate to 0.8 g/min. The decreased conductivity suggests a decreased concentration of free ions due to a chemical reaction of phosphate ions with the added ammonium hydroxide. In contrast, the deposition from the suspension of nano-SiC powder with the AP additive was only successful in acidic suspensions. However, even in this case the deposition was slow, and the deposits contained bubbles. The initial current was very high, 0 2 4 6 8 10 12 14 16 18 0 1 3 4 6 7 8 91 2 5 0 time [min] Current [mA] 17 wt. % 8 wt. % Fig. 4 Electrical current during the deposition of nano-SiC suspension as a function of solids content in the starting suspensions (60 V, steel electrode) J Mater Sci (2006) 41:8093–8100 8097 123
8098 J Mater Sci(2006)41:8093-8100 As the main aim of the present study was the EPD of sic with the ap addition on Sic-fibres in further experiments a bundle of Sic-fibres was placed in the 10 cell as the deposition electrode. Due to its small △pH2 effective surface area, the initial current was only 3 mA. The ph of the 口pH10 with the addition of ap was 2. 8 and the solids content X pH2.8 was 35 wt %. Figure 7 presents the current change during the deposition and the image of the bundle with the deposit. Due to the small quantity of the deposit or the fibres. the weight was not monitored. a visual inspection revealed an apparently relatively firm deposit on the Sic-bundle The Sic bundle with the deposit, composed of nano- Sic powder and AP, was sintered in an inert atmo- HILLH sphere at 1450C. The cross-section of the sintered sample was polished and inspected by SEM, TEM and without AP EDS. The microstructures of the matrix were rather dense with a small amount of closed porosity, and good cohesion between the matrix and the fibres is evident ( Figs. &a and b). Due to a low firing temperature, the 0 1 2 3 4 56 789 10 SiC-particle size remained practically unchanged, i time [min 50 nm(Fig. &c). The EDS spectrum reveals a certain Fig.5 The current during the deposition of um-SiC suspension amount of oxygen in the matrix material, probably with the addition of aluminium phosphate as a function of solids resulting primarily from the oxygen-containing layer content in the starting suspensions(60 V, steel electrode) on the nano-SiC particles. 70 mA, and a significant current drop was observed Conclusions during the deposition (Fig 6a). This could explained by the reduced active surface of the elec- In the present investigation the effect of the compo- trode due to the presence of bubbles. As shown in sition of Sic-based suspensions on EPD was studied Fig 6b, the bubbles were concentrated by the elec- with the aim of establishing the conditions for depo- trode, while the bulk deposit appears rather dense. sition on SiC-fibres. Ethanol suspensions of two Dilution of the suspension with ethanol diminished the different grades of Sic powders, with and without the current, as was also observed for the nano-Sic addition of AP (as a sintering aid), were used for suspension without the AP addition. depositing on steel electrodes Fig. 6(a)-The current change during the depositie a b of the suspension of nano-SiC hosphate as a function of solids content; (b)-image of the inner side of the deposi bubble 1 cm me [min 2 Springer
70 mA, and a significant current drop was observed during the deposition (Fig. 6a). This could be explained by the reduced active surface of the electrode due to the presence of bubbles. As shown in Fig. 6b, the bubbles were concentrated by the electrode, while the bulk deposit appears rather dense. Dilution of the suspension with ethanol diminished the current, as was also observed for the nano-SiC suspension without the AP addition. As the main aim of the present study was the EPD of SiC with the AP addition on SiC-fibres, in further experiments a bundle of SiC-fibres was placed in the cell as the deposition electrode. Due to its small effective surface area, the initial current was only 3 mA. The pH of the suspension of nano-SiC powder with the addition of AP was 2.8 and the solids content was 35 wt.%. Figure 7 presents the current change during the deposition and the image of the bundle with the deposit. Due to the small quantity of the deposit on the fibres, the weight was not monitored. A visual inspection revealed an apparently relatively firm deposit on the SiC-bundle. The SiC bundle with the deposit, composed of nanoSiC powder and AP, was sintered in an inert atmosphere at 1450 C. The cross-section of the sintered sample was polished and inspected by SEM, TEM and EDS. The microstructures of the matrix were rather dense with a small amount of closed porosity, and good cohesion between the matrix and the fibres is evident (Figs. 8a and b). Due to a low firing temperature, the SiC-particle size remained practically unchanged, i.e., 50 nm (Fig. 8c). The EDS spectrum reveals a certain amount of oxygen in the matrix material, probably resulting primarily from the oxygen-containing layer on the nano-SiC particles. Conclusions In the present investigation the effect of the composition of SiC-based suspensions on EPD was studied with the aim of establishing the conditions for deposition on SiC-fibres. Ethanol suspensions of two different grades of SiC powders, with and without the addition of AP (as a sintering aid), were used for depositing on steel electrodes. 0 1 2 3 4 5 6 7 8 9 10 11 12 02 57 10 time [min] pH 2 pH10 pH2.8 current (mA) without AP 1 3 4 6 8 9 Fig. 5 The current during the deposition of lm-SiC suspension with the addition of aluminium phosphate as a function of solids content in the starting suspensions (60 V, steel electrode) 0 10 20 30 40 50 60 70 80 024 time [min] Current [mA] 50 w.t % 25 wt.% 15 wt. % 1 3 5 1 cm a b Fig. 6 (a) - The current change during the deposition of the suspension of nano-SiC powder and aluminiumphosphate as a function of solids content; (b) - image of the inner side of the deposit with bubbles 8098 J Mater Sci (2006) 41:8093–8100 123
J Mater sci(2006041:8093-8100 7 The current change during the deposition of a suspension of nano-SiC powder with the addition of aluminum phosphate on a bundle of Sic-fibres(solids content: 10 wt %);(b)-an E: den se of the bi 012345678910 time(min Fig 8(a),(b)-SEM micrographs of the sintered material with SiC-fibres;(c) and(d) TEM micrograph and EDS spectrum of the SiC 500 sr 00 The results showed that the conductivity of the The deposits with the highest packing density were suspensions is strongly dependent on the amount of obtained from highly loaded acidic suspensions of free ions. In the case of submicron Sic powder, the submicron Sic powder. The EPD of the nano-SiC major charge carriers were ions, while the contribution powder was much less effective than that for the of the nano-sized Sic powder is probably not negligi micron SiC. This is presumably connected to the ble. The solids content in the fresh as-formed deposit oxide layer on the particles, however, a full expla increased with the surface charge on the particles. The nation will require further detailed studies. Alumin solids content was also observed to have a beneficial ium phosphate, added as a sintering aid for Sic,as effect in the analysed range, i. e, up to 70 wt % During well as the use of nano-Sic powder, significantly the depositions, a current drop was normally observed changed the behaviour of the suspension. The large for porous deposits containing bubbles. This suggests current at the used potential of 60 V and the that the current drop was not caused by the inhibited formation of bubbles on the steel electrode were transport of the charge carriers due to the accumula- observed. In contrast, when the suspension with the tion of the deposit, but rather by the reduced active nano-Sic powder and the sintering additive was urface of the electrode deposited on the bundle of Sic fibres, the current 2 Springer
The results showed that the conductivity of the suspensions is strongly dependent on the amount of free ions. In the case of submicron SiC powder, the major charge carriers were ions, while the contribution of the nano-sized SiC powder is probably not negligible. The solids content in the fresh as-formed deposit increased with the surface charge on the particles. The solids content was also observed to have a beneficial effect in the analysed range, i.e., up to 70 wt.%. During the depositions, a current drop was normally observed for porous deposits containing bubbles. This suggests that the current drop was not caused by the inhibited transport of the charge carriers due to the accumulation of the deposit, but rather by the reduced active surface of the electrode. The deposits with the highest packing density were obtained from highly loaded acidic suspensions of submicron SiC powder. The EPD of the nano-SiC powder was much less effective than that for the submicron SiC. This is presumably connected to the oxide layer on the particles, however, a full explanation will require further detailed studies. Aluminium phosphate, added as a sintering aid for SiC, as well as the use of nano-SiC powder, significantly changed the behaviour of the suspension. The large current at the used potential of 60 V and the formation of bubbles on the steel electrode were observed. In contrast, when the suspension with the nano-SiC powder and the sintering additive was deposited on the bundle of SiC fibres, the current Fig. 8 (a), (b) - SEM micrographs of the sintered material with SiC-fibres; (c) and (d) TEM micrograph and EDS spectrum of the SiCmatrix, respectively 0 1 2 3 4 5 0 2 4 5 6 7 8 9 10 time [min] Current [mA] 1 mm 1 3 a b Fig. 7 (a) - The current change during the deposition of a suspension of nano-SiC powder with the addition of aluminum phosphate on a bundle of SiC-fibres (solids content: 10 wt. %); (b) - an image of the bundle with deposit J Mater Sci (2006) 41:8093–8100 8099 123
8100 J Mater Sci(2006)41:8093-8100 was low and no bubbles were observed in the 7. Nannetti CA Riccardi B Ortona A La Barbera A Scafe e d t Vekinis G(2002)J Nucl Mater 307-311: 1196 8. Katoh Y, Kohyama A, Nozawa T et al (2004)J Nucl Mater 329:587 cially supported by the and by the ministry of 9. Nannetti CA, Ortona A, de Pinto Da et al(2004)JAm Ceram Soc 87: 1205 lovenia. The XPS analysis of the powders were performed by 10. Novak S, Drazic G(2 Slip-infiltration of SiC-fiber Dr.J. Kovac preforms for production of sicf/sic composites, 12th Int. conf on fusion reactor materials. Santa Barbara. CA. USA 11. Boccaccini AR, Zhitomirsky I (2002)Curr Opin Solid State References 12. Van der Biest O, Put S, Anne G, vleugels J(2004)J Mater sci39:779 1. Krenkel W(2005)In: Bansal NP(ed)Handbook of ceramic 13. Fukada Y, Nagarajan N, Mekky W, Bao Y, Kim HS Nicholson PS (2004)J Mater Sci 39: 787 2. Keller KA, Jefferson G, Kerans R(2005)In: Bansal NP(ed) 14. Vandeperre L, Van der Biest o, Bouyer F, Persello J, Foissy Handbook of ceramic composites. Kluwer, Boston, P 377 3. Tavassoli AAF (2002)J Nucl Mater 302: 73 5 ouyer F, Foissy A(1999)J Am Ceram Soc 82(8): 00 4. Jones RH, Giancarli L, Hasegawa A, Katoh Y, Kohyama A, 16. Moritz K, Muller E(2002) Key Eng Mat 206(2):193 Riccardi B, Snead LL, Weber WJ(2002) J Nucl Mater 17. Boccaccini AR, Kaya C, Chawla KK(2001)Composites A 307:1057 ppl S 5. Naslain R(2004)Comp Sci Technol 64(2): 155 18. Boccaccini AR(2004) Mater Sci Forum 221 6. Igawa N, Taguchi T, Nozawa T et al(2005)J Phys Chem Solids 66: 551 2 Springer
was low and no bubbles were observed in the deposit. Acknowledgement The work was financially supported by the European Commission Euratom/Fusion and by the Ministry of High Education, Science and Technology of Republic of Slovenia. The XPS analysis of the powders were performed by Dr. J. Kovac. References 1. Krenkel W (2005) In: Bansal NP (ed) Handbook of ceramic composites. Kluwer, Boston, p 117 2. Keller KA, Jefferson G, Kerans R (2005) In: Bansal NP (ed) Handbook of ceramic composites. Kluwer, Boston, p 377 3. Tavassoli AAF (2002) J Nucl Mater 302:73 4. Jones RH, Giancarli L, Hasegawa A, Katoh Y, Kohyama A, Riccardi B, Snead LL, Weber WJ (2002) J Nucl Mater 307:1057 5. Naslain R (2004) Comp Sci Technol 64(2):155 6. Igawa N, Taguchi T, Nozawa T et al (2005) J Phys Chem Solids 66:551 7. Nannetti CA, Riccardi B, Ortona A, La Barbera A, Scafe E, Vekinis G (2002) J Nucl Mater 307–311:1196 8. Katoh Y, Kohyama A, Nozawa T et al (2004) J Nucl Mater 329:587 9. Nannetti CA, Ortona A, de Pinto DA et al (2004) J Am Ceram Soc 87:1205 10. Novak S, Drazic G (2005) Slip-infiltration of SiC-fiber preforms for production of sicf/sic composites, 12th Int. conf. on fusion reactor materials, Santa Barbara, CA, USA 11. Boccaccini AR, Zhitomirsky I (2002) Curr Opin Solid State Mater Sci 6:251 12. Van der Biest O, Put S, Anne G, Vleugels J (2004) J Mater Sci 39:779 13. Fukada Y, Nagarajan N, Mekky W, Bao Y, Kim HS, Nicholson PS (2004) J Mater Sci 39:787 14. Vandeperre L, Van der Biest O, Bouyer F, Persello J, Foissy A (1997) J Eur Ceram Soc 17:373 15. Bouyer F, Foissy A (1999) J Am Ceram Soc 82(8):001 16. Moritz K, Muller E (2002) Key Eng Mat 206(2):193 17. Boccaccini AR, Kaya C, Chawla KK (2001) Composites A – Appl Sci Manuf 32(8):997 18. Boccaccini AR (2004) Mater Sci Forum 221 8100 J Mater Sci (2006) 41:8093–8100 123
Copyright of Journal of Materials Science is the property of Springer Science Business Media B.V. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holders express written permission. However, users may print, download or email articles for individual use