J. Am. Ceram. Soc., 92 [1] 9–16 (2009) DOI: 10.1111/j.1551-2916.2008.02823.x r 2008 The American Ceramic Society
Journal
Aqueous Colloidal Processing of ZTA Composites
Susana M. Olhero,z,y Ibram Ganesh,z,z Paula M. C. Torres,z Fernando J. Alves,y and Jose´ M. F. Ferreiraw,z z
Department of Ceramics and Glass Engineering, CICECO, University of Aveiro, Aveiro, P-3810193, Portugal
y
Department of Mechanical Engineering and Industrial Management, FEUP, University of Porto, Porto, Portugal
z
Centre for Advanced Ceramics, International Advanced Research Centre for Powder Metallurgy and New Materials (ARCI), Hyderabad – 500 005, A.P., India
agglomerates by milling or ultrasonic treatments, or removing them and other flaw sources from the suspensions by filtration, or sedimentation are some advantages of wet processing routes. In this way, optimal particle packing and homogeneous green microstructures can be obtained. However, traditional slip casting (SC) in plaster molds has serious limitations in consolidating thick components due to its low driving force and the occurrence of density gradients, heterogeneous distribution of soluble species such as binder molecules or inorganic ions,16 and particle segregations upon liquid drainage due to clogging and/or gravity effects.17 Although improvements are possible by applying an external pressure,18,19 some difficulties still persist when solid complex shapes with varying thickness have to be formed. The thinner parts consolidate first, hindering the passage of the slip to feed the thicker ones. On the other hand, the versatility and capability of injection molding to consolidate complex shapes is strongly mitigated by the problematic binder removal step.20 In view of the above limitations, several other concepts have been developed recently for transforming fluid suspensions into rigid bodies without liquid removal. These forming processes include direct coagulation casting,21 hydrolysis-assisted solidification (HAS),11 starch consolidation,22,23 freezing of the suspension,24 aqueous gelcasting (GC) using organic monomers,9,13,14 the methylcellulose gelation on heating25, or the gelation of other polysaccharides on cooling26,27 to create a three-dimensional (3D) network. Among all these processing techniques, HAS emerges as the simplest and the most inexpensive one to consolidate several kinds of ceramics in which alumina is a minor or a major constituent.11,12 Consolidation is caused by the hydrolysis of AlN powder (AlN13H2OAl(OH)31NH3), which leads to a drastic change in the viscosity of the suspension. The resulting aluminum hydroxide acts as cement, conferring high stiffness to the consolidated parts.12 However, the as-consolidated parts are brittle and possess relatively poor strength, and certain components like crucibles, bushings, valve seats, pump components, etc., with thin walls and complex shapes are difficult to fabricate.12 In contrast, aqueous GC confers high green strength and enables green machining.9,13–15 Furthermore, it is an affordable, rapid, and near net-shape-forming process, suitable for producing complexshaped parts. In this study, ZTA ceramics containing 30 and 60 wt% ZrO2 were consolidated from aqueous suspensions (50 vol% solids loading) via the GC process. The same ceramic compositions were also consolidated by SC, HAS, and conventional die pressing of freeze-dried granules (FG) for comparison purposes. The dried consolidated parts were sintered for 1 h at 16001C and characterized for bulk density (BD), apparent porosity (AP), water absorption (WA) capacity, hardness, fracture toughness, three-point bend strength, X-ray diffraction (XRD) phase, and microstructural properties. The effects of different processing routes on the green and sintered properties of ZTA ceramics are presented.
Two different zirconia-alumina composites, ZTA-30 (70 wt% Al2O3130 wt% ZrO2) and ZTA-60 (40 wt% Al2O3160 wt% ZrO2), with potential for orthopedic applications, were processed in aqueous media and consolidated by slip casting (SC), hydrolysis-assisted solidification (HAS), and gelcasting (GC) from suspensions containing 50 vol% solids loading. For comparison purposes, the same ceramic compositions were also consolidated by die pressing of freeze-dried granules (FG). In the HAS process, 5 wt% of Al2O3 in the precursor mixture was replaced by equivalent amounts of AlN to promote the consolidation of the suspensions. Ceramics consolidated via GC exhibited higher green (three-point bend) strengths (B17 MPa) than those consolidated by other techniques. Further, these ceramics also exhibited superior fracture toughness and flexural strength properties after sintering for 1 h at 16001C in comparison with those consolidated by other techniques, including conventional die pressing (FG). I. Introduction
R
ECENTLY, zirconia-toughened-alumina (ZTA) ceramics have received considerable attention due to their attractive properties, including high-temperature mechanical strength, good thermal shock resistance, wear and oxidation resistance, low thermal conductivity, and the close match between their thermal expansion coefficients and those of metals.1,2 These properties make ZTA ceramics suitable for a variety of high demanding applications including dental screws, cutting blades, electrosurgical insulators, valve seals, body armor, pump components, oxygen sensors, dies, and prosthesis components such as hip joints.3–8 The zirconia grains embedded in an alumina matrix enhance the flexural strength, fracture toughness, and fatigue resistance by a stress-induced phase transformation mechanism of tetragonal zirconia.3–8 ZTA ceramics are usually consolidated by conventional dry pressing. This process has strong limitations particularly when components with a large size and complex shapes are to be produced, because they require extensive postsintering machining operations. Colloidal shaping methods that enable controlling and manipulating the forces between the particles dispersed in a liquid have attracted the attention of ceramists in the last decades.9–15 They enable to achieve higher microstructural homogeneity in green and sintered parts, while some offer near-net-shaping capabilities that reduce the postsintering machining operations and the production costs.9–15 The possibility of breaking particle
S. Danforth—contributing editor
Manuscript No. 24480. Received March 30, 2008; approved October 5, 2008. Financial support was provided under the grant SFRH/BPD/27013/2006 by Foundation for Science and Technology of Portugal. The financial support of CICECO is also acknowledged. w Author to whom correspondence should be addressed. e-mail:
[email protected]
9
10
Vol. 92, No. 1
Journal of the American Ceramic Society—Olhero et al. II. Experimental Procedure
(1) Raw Materials and Powder Processing An alumina powder (CT-3000SG, Alcoa-Chemie GmbH, Ludwigshafen, Germany, with an average particle size, D50 5 0.8 mm) and a zirconia powder (Tosoh-Zirconia, TZ-3YS, Yamaguchi, Japan, D50 5 0.43 mm) were used as the main components. A high-purity AlN powder (Grade AT, H.C. Stark, Goslar, Germany, D50 5 0.33 mm, oxygen content o1.3%) was also used as consolidating agent in the HAS process. The precursor compositions used to prepare the ZTA-30 (70 wt% Al2O3130 wt% ZrO2) and ZTA-60 (40 wt% Al2O3160 wt% ZrO2) ceramics are summarized in Table I. The numbers 30 and 60 in the sample codes indicate the weight percent of zirconia in the composite. The remaining sample codes stand for the consolidation process: FG - die pressing of freeze-dried granules; SC - slip casting; GC - gelcasting; and HAS - hydrolysis-assisted solidification. Five weight percent Al2O3 was replaced by an equivalent amount of AlN in the HAS suspensions. All the suspensions contained 50 vol% solids. They were prepared by wet ball milling for 24 h in the polypropylene bottles using alumina balls (the charge to balls ratio was 1:3). For the SC, HAS, and FG processes, an aqueous solution of a polycarboxylic acid without alkalis, Dolapix CE 64 (Zschimmer and Schwarz, Chemnitztalstrasse, Germany), at the ratio of 0.4 wt% on the dry powder weight basis, was used as the dispersing medium. The as-obtained suspensions were filtered off and degassed for 5 min by vacuum pumping. Crucibles with 100 mL volume and 5 mm wall thickness, and cylindrical rods with 20 mm diameter and about 20 mm in height were fabricated by SC in plaster molds. For consolidation via HAS, the required amounts of AlN were added to the degassed suspensions and the mixing was continued for a further 2 h under the same ballmilling conditions. The final suspensions were then consolidated into split-type aluminum molds (60 mm 30 mm 30 mm). For the freeze granulation process, the degassed suspensions were diluted to 35 vol% by adding the required amount of distilled water and the 3 wt% on a powder weight basis of an emulsion binder, Duramax D1000 (Rohm and Haas, Lauterbourg, France). This procedure is justified because deagglomerating powder particles in a suspension is easier at moderate-high solids loading in comparison with dilute systems. On the other hand, concentrated suspensions tend to exhibit shear thickening characteristics when subjected to high shear rates as those prevailing upon flowing through a narrow (0.7 mm diameter) spraying nozzle. The 35 vol% solids suspensions were sprayed into liquid nitrogen (1961C) to obtain granules by freeze granulation (Power Pro freeze granulator LS-2, Gothenburg, Sweden). The granules were then dried at 491C under a pressure of 1 103 torr in a freeze-drying system (Labconco, LYPH Lock 4.5, Kansas City, MO) for several days. The dried granules were uniaxially pressed (200 MPa) in a metal die to obtain pellets with 30 mm diameter and B8 mm height. The suspensions for GC were prepared by dispersing the ZTA powder precursor mixtures in an aqueous-organic premix solu-
tion13,14 obtained by dissolving 20 wt% MAM (methacrylamide), MBAM (mithylenebisacrylamide), and NVP (nvinylpyrrolidinone) in a 3:1:3 weight ratio in deionized water in the presence of 0.4 wt% Dolapix CE 64 (on the dry powder weight basis) as a dispersing agent. Deagglomeration was performed for 24 h in polypropylene bottles containing alumina balls (12 mm diameter) using a roller mill. The weight ratio between the powder and the balls was 1:3. The as-obtained suspensions were filtered off and degassed for 5 min by vacuum pumping. 1-octanol at the ratio of 1 mL/g of suspension as a deairing agent, a polymerization initiator (10 wt% aqueous solution of ammonium per-sulfate, APS), and a catalyst (tetramethylethylenediammine, TEMED) at the ratio of 4 and 2 mL/g of suspension, respectively,13,14 were added to the GC suspensions. After being degassed again for 2 min, these suspensions were poured into nonporous white petroleum jellycoated split-type aluminum molds (60 mm 30 mm 30 mm). The gelation was carried out under ambient conditions for about 1 h after adding the polymerization initiator and the catalyst. The consolidated parts were dried under controlled humidity conditions to avoid cracking and nonuniform shrinkage. The ZTA samples consolidated by the different processes were heat treated at a rate of 11C/min up to 5001C and held at this temperature for 2 h. Then the sintering was conducted at a heating rate of 51C/min up to 16001C with a 1-h holding time at this temperature.
(2) Characterization Techniques Particle size analysis of the powders was performed by light scattering (Coulter LS 230, Miami, FL, Fraunhofer optical model). The viscosity of the suspensions was measured using a rotational Rheometer (Bohlin C-VOR Instruments, Worcestershire, U.K.). The measuring configuration adopted was a cone and plate (41, 40 mm, and a gap of 150 mm), and flow measurements were conducted between 0.1 and 800 s1. The same configuration was used to obtain information about the gelation behavior and the evolution of gel stiffness under dynamic measurements of G0 and G00 in the linear viscoelastic region. A time sweep was conducted for 30 min at a constant frequency of 1 Hz. A frequency sweep from 1 to 100 Hz under a constant stress of 100 Pa was performed immediately after 30 min of gel time, i.e., the time corresponding to the crossover point of the G0 and G00 curves. The BD, AP, and WA capacity of various sintered ZTA ceramics were measured in aqueous media according to the Archimedes principle (ASTM C372, ASTM International, West Conshohocken, PA) using a Mettler balance (AG 245, Mettler Toledo, Zurich, Switzerland). On average, three density measurements were performed for each sample (70.01 error).28,29 XRD patterns were recorded on a Rigaku (Tokyo, Japan) XRD equipment using a diffracted beam monochromated CuKa (1.5406 A˚) radiation source.30 Crystalline phases were identified by comparing with PDF-4 reference data from the International Centre for Diffraction Data (ICDD).31 The
Table I. Precursors Compositions Used for the Preparation of ZTA-30 (70 wt% Al2O3130 wt% ZrO2) and ZTA-60 (40 wt% Al2O3160 wt% ZrO2) Ceramics Following Various Aqueous Colloidal Processing Routesw Samplez
30FG 30SC 30GC 30HAS 60FG 60SC 60GC 60HAS w
Processing route
Slurry medium
Al2O3 (wt%)
ZrO2 (wt%)
AlN (wt%)
Die-pressing of freeze dried granules (FG) Slip casting (SC) Gelcasting (GC) Hydrolysis assisted solidification (HAS) Die-pressing of freeze dried granules (FG) Slip casting (SC) Gelcasting (GC) Hydrolysis assisted solidification (HAS)
— Water Premix solution Water — Water Premix solution Water
70 70 70 65.98 40 40 40 38.03
30 30 30 29.99 60 60 60 60.35
— — — 4.02 — — — 1.62
See experimental section for the details. The numbers 30 and 60 in the sample codes indicate the zirconia concentration of the composite; in the case of HAS suspensions 5 wt% Al2O3 was replaced by an equivalent amount of AlN. All the suspensions used in the casting processes contained 50 vol% solids.
z
January 2009
microstructures of dense ZTA ceramics were examined by scanning electron microscopy (SEM) (Hitachi S-4100, Tokyo, Japan) on thermally etched (15001C for 10 min) surfaces coated with carbon. The mechanical properties evaluated were hardness (H), fracture toughness (KIc), and flexural strength. Hardness was determined by applying a Vickers indenter and calculated as H 5 P/2d2, d being the half-diagonal indentation impression and P the indentation load (10 kg). The fracture toughness (KIc) was calculated on the basis of the indentation method [KIc 5 H a1/2 0.203 (C/a)3/2].32 Here, 2a represents the Vickers indent diagonal length, 2C the resulting crack length, and H is the Vickers hardness (HV 5 Kg/mm2 5 10 MPa). Five to six samples were tested for fracture toughness, with five measurements in each sample. The results presented are therefore the average of at least 25 measurements in order to assure good reproducibility.13,14 The data of hardness, crack, and diagonal length were collected using a microhardness tester (Leitz Wetzler, Wetzlar, Germany) by holding the indenter tip (with 1371) under a 10-kg load for 20 s on the surface of the sample polished to a mirror finish. The flexural strength of the green and sintered samples was measured using a three-point bending test (JIS-R1601). About 15–20 samples were tested per case and the results are presented as mean values (70.01 error).
III. Results and Discussion (1) Suspension Characteristics The viscosity versus shear rate curves of the ZTA-30 and ZTA60 suspensions prepared for SC, GC, and HAS are shown in Fig. 1. Irrespective of the processing method and the precursor composition, all suspensions exhibit smooth shear thinning behaviors for shear rates g_ < 300s1 , and some flow disturbances for higher shear rates. Furthermore, the organic additives in premix solutions tended to slightly increase the viscosity of the ZTA-30 precursor mixtures relative to the simple dispersant solutions. No noticeable differences in viscosity could be seen among the ZTA-60 suspensions, which are generally more viscous than the ZTA-30 ones. The viscosity differences between the two precursor mixture suspensions could be due to differences in the isoelectric points,33,34 Hamaker constants,35 particle sizes,36 etc., of the powders. The smaller particle size of the zirconia powder in comparison with alumina will enhance the viscosity of the suspensions that are more zirconia rich.36 Furthermore, the Hamaker constant of zirconia is greater than that of alumina, leading to stronger attractions between the particles.35 Moreover, Dolapix CE 64, being an anionic dispersant, will show a higher tendency to be adsorbed onto the surfaces of alumina particles with an isoelectric point of about 100
30SC 30HAS 30GC 60SC 60HAS 60GC
8–9,33,34 than onto the surfaces of ZrO2 particles with a lower isoelectric point of about 4–6.33 This means that a better dispersing efficiency should be expected for the precursor mixtures that are more alumina rich, ZTA-30, as observed. The addition of AlN did not show any remarkable effect on the viscosity of ZTA-30 and ZTA-60 suspensions. The evolution of storage modulus, G0 , and the loss modulus, G00 , of 30GC and 60GC suspensions upon addition of the polymerization initiator (APS) and the catalyst (TEMED) is presented in Fig. 2 as a function of the reaction time. The time at which the curves of G0 and G00 intersect each other is taken as the gelation time, i.e., the time at which the elastic component, G0 , starts predominating over the viscous component. Both the G0 and the G00 parameters first increase with time up to the crossover point (G0 5 G00 ), with a predominance of G00 . After the crossover point, the situation is reversed; G00 tends to decrease while G0 continually increases, reaching values that are 1–2 orders of magnitude higher than the G00 . The crossover points occurred at the end of 5 and 8 min in the 30GC and 60GC systems, respectively, indicating a faster gelation process of the 30GC slurry. In order to compare the strength of the network structure of 30GC and 60GC suspensions upon gelation, mechanical spectra (Fig. 3) were carried out at ambient temperature after 30 min of the crossover point of G0 and G00 (gel time). Because of the huge differences in magnitude between G0 and G00 , the dynamic analysis was restricted to the elastic component only. In fact, under these conditions, the recording of G00 is no longer reliable because it depends on the instrumental resolution of the phase lag between sinusoidal stress and strain.37 Figure 3 shows that both curves run almost parallel to the x-axis, a characteristic behavior of stiff gels, although a stronger gel network has been formed in the 30GC system. It has been reported that Al2O3 favors the polymerization of MAM and MBAM monomers by the free radical initiation of APS.9,13 This might explain why the ZTA-30 system that is more alumina rich forms a stiffer gel.
(2) Characteristics of Green Consolidates Table II lists the values of the green density, percentage linear shrinkage upon drying, and green (three-point bending) strength of differently consolidated ZTA ceramics, together with the values of viscosity of suspensions at g_ ¼ 140s1 , the predicted amounts of NH3 released upon complete hydrolysis of AlN, and the setting times. Although the gel time defined as the crossover 60GC
G' G''
10000
1000
G'/G'' (Pa)
10 Viscosity (Pa.s)
11
Aqueous Colloidal Processing of ZTA Composites
1
100 30GC 10000
0.1 1000 0.01 –100
100
300 500 700 Shear rate (1/s)
900
1100
Fig. 1. The flow behavior of aqueous (30SC, 30HAS, 60SC, and 60HAS) and aqueous-organic premix (30GC and 60GC) suspensions containing 50 vol% solids loading. In the case of 30HAS and 60HAS suspensions, 5 wt% Al2O3 was replaced by an equivalent amount of AlN. HAS, hydrolysis-assisted solidification; SC, slip casting; GC, gelcasting.
100 0
2
4
6 8 10 Time (min)
12
14
Fig. 2. Setting and visco-elastic behavior of suspensions containing 50 vol% ZTA-30 and ZTA-60 precursor mixtures dispersed in the aqueousorganic premix solution.
12
Vol. 92, No. 1
Journal of the American Ceramic Society—Olhero et al. 100000 30GC 60GC G' (Pa)
70000
40000
10000 1
10 Frequency (Hz)
100
Fig. 3. Elastic modulus of the gels formed by the free radical polymerization of organic monomers in the 30GC and 60GC suspensions measured immediately 30 min after the gel time. GC, gelcasting.
point (G0 5 G00 ) has varied in the range of 5–8 min (Fig. 2), the consolidated parts only acquired sufficient handling strength after about 30 min. Therefore, Table II also presents the shortest time after which the gelcast samples could be unmolded. The time reported for the 30 and 60SC systems was the time required to slip cast unmolded 100 mL crucibles with 5 mm wall thickness. The 30 and 60HAS systems required almost 24 h for setting and for safe handling of the consolidated parts. Table II shows that green ZTA-60 (richer in ZrO2) bodies are denser than ZTA-30 ones due to the higher density of ZrO2 (5.890 g/cm3) in comparison with 3.98 g/cm3 for Al2O3. Concerning the different consolidation techniques, SC (involving liquid removal) gives higher green densities in comparison with direct consolidation techniques (without liquid removal), as expected. The presence of about 4 wt% AlN (less dense, 3.260 g/cm3) and the possible entrapment of some NH3 gas released during consolidation due to hydrolysis of AlN might also account for the lowest green density measured for the 30HAS green bodies. Micrographs of fracture surfaces of green 30HAS and 30GC are presented in Figs. 4(a) and (b), respectively. The 30HAS sample broke more easily and its fracture surface was smoother in comparison with the stronger 30GC sample (Table II), showing a kind of grain pullout. The higher relative density of the 30GC sample in comparison with the 30HAS sample (Table II) is also apparent from these micrographs. The 3D network of fine capillaries formed in the former sample upon polymerization of MAM, MBAM, and NVP could lead to higher drying shrinkage and to packing of denser particles (Table II). Furthermore, the presence of relatively fine aluminum hydroxide par-
ticles derived from AlN hydrolysis can be seen in the 30HAS micrograph (Fig. 4(a)). Green bodies consolidated by SC and GC exhibit higher shrinkage values upon drying than those consolidated by HAS. This was expected from the higher green density of SC bodies and the closeness of the particles forming finer pores with a higher suction pressure, the driving force for shrinkage upon drying. Although the proximity of the particles is less in the GC, the polymerization creates a 3D network of fine capillaries that exert a high suction pressure, equivalent to an external applied pressure that promotes shrinkage. This is not the case in the HAS process in which consolidation takes place with a pH shift toward the isoelectric point of alumina, leading to an in situ flocculation, less prone to form fine capillaries. Nevertheless, all green bodies exhibit linear drying shrinkage o3%, due to the high solids loading of the starting suspensions. It was reported that linear drying shrinkage of complex-shaped ceramic components consolidated by near-net shape-forming processes should be kept below 3% to assure successful fabrication.15,24 Table II also reveals that samples consolidated by GC exhibit higher green strength values than those prepared by SC or HAS processes. This improved strength of GC green bodies is due to the network formed by the polymerization of MAM, MBAM, and NVP monomers.10 In the case of the HAS process, aluminum hydroxides formed upon AlN hydrolysis exert a cementing action, conferring stiffness to the body.11,38,39 The cementing effect tends to increase with increasing concentration of AlN in the suspension. However, the amount of AlN should be such that whatever the NH3 gas released upon AlN hydrolysis (AlN13H2O-Al(OH)31NH3) should be soluble in the aqueous medium (NH31H2O-NH4OH). Otherwise, the excess gas will not be accommodated in the solution, forming bubbles that may become entrapped in the concentrated suspensions, which may lead to the formation of pores in green and sintered parts, which act as strength degrading flaws. 30 and 60GC exhibited green strengths of 17.2970.84 and 17.3771.325 MPa, respectively, being the highest values reported up to now for greens of ZTA ceramics. Green strength does not seem to be influenced by the precursor mixture, because 30 and 60GC ceramics exhibit similar values. Such a level of green strength enables green machining; thereby, expensive post sintering machining operations can be minimized. These results indicate GC as a better process for fabricating high green strength ZTA bodies in comparison with HAS or SC.
(3) Characteristics of Sintered Bodies The values of BD, shrinkage upon sintering, theoretical density (TD), AP, WA capacity, hardness, fracture toughness, and
Table II. Slurry Characteristics and Properties of Green ZTA-30 (70 wt% Al2O3130 wt% ZrO2) and ZTA-60 (40 wt% Al2O3160 wt% ZrO2) Ceramics Consolidated by Various Aqueous Colloidal Processing Routesw Green density Samplez
Slurry viscosity (mPa s at g_ ¼ 140s1 )
NH3 to be released by AlN hydrolysis (wt%)
Slurry setting time (min)y
(g/cm3)
%TDz
Linear shrinkage upon drying (%)
Green strength (MPa)
30FG 30SC 30HAS 30GC 60FG 60SC 60HAS 60GC
— 64.2 78.3 151 — 265 346 264
— — 1.67 — — — 0.67 —
— B20 o1440 B5 (30) — B20 o1440 B8 (30)
2.6570.18 2.8070.23 2.4070.06 2.7670.23 2.9570.31 3.2070.29 2.8470.42 2.9970.02
58.20 61.49 52.71 60.62 57.55 62.42 55.40 58.33
— 2.5170.37 1.5870.20 2.3670.27 — 2.4370.37 1.6270.31 2.7170.21
o0.25 4.0470.29 7.7672.34 17.2970.84 o0.25 2.7670.44 5.3570.76 17.3771.32
w
See experimental section for the details. The numbers 30 and 60 in the sample codes indicate the zirconia concentration of the composite; in the case of HAS suspensions 5 wt% Al2O3 was replaced by an equivalent amount of AlN. All the suspensions used in the casting processes contained 50 vol% solids. y The gel time, i.e., the time corresponding to the crossover point (G0 5 G00 ) was 5 and 8 min for the 30GC and 60GC systems, respectively. The values given in brackets represent the shortest time at which samples could be unmolded. In the case of 30SC and 60SC, the values represent the time required for removing a 100 mL volume and 5mm-thick crucible from the plaster mold. z The percentage of theoretical density values of ZTA-30 (4.553 g/cm3) and ZTA-60 (5.126 g/cm3) were calculated based on the rule of mixture of alumina (3.98 g/cm3) and zirconia (5.89 g/cm3) theoretical densities. HAS, hydrolysis-assisted solidification; GC, gelcasting. z
January 2009
13
Aqueous Colloidal Processing of ZTA Composites
(a) 60HAS
Intensity (a.u.)
60FG
10 μm
30HAS
(b) 30FG
((ZrO2)0.91(Y2O3)0.09)0.917 m-ZrO2 α-Al2O3 10
10 μm Fig. 4. SEM micrographs of fracture surfaces of green 30HAS (a) and 30GC (b) samples. Some pores, probably derived from the entrapment of ammonia gas due to the hydrolysis of AlN particles are circled in (a). GC, gelcasting; HAS, hydrolysis-assisted solidification; SEM, scanning electron microscopy.
flexural strength of ZTA-30 and ZTA-60 ceramics sintered for 1 h at 16001C are presented in Table III. The TDs of ZTA-30 (4.553 g/cm3) and ZTA-60 (5.126 g/cm3) were calculated based on the TDs of alumina (3.98 g/cm3) and zirconia (5.89 g/cm3) according to the rule of mixture and were used to determine the %TD. For the ceramics consolidated by FG, SC, and GC, ZTA-60 samples are denser (496%TD) than ZTA-30 (93.3%– 95.7%TD) ones. Differences can be attributed to the fineness of zirconia powder and its better sintering ability. For both compositions, the highest sintered densities were obtained for slip-cast bodies, certainly due to the higher starting green
20
30
40 50 2θ (degrees)
60
70
Fig. 5. XRD patterns of various 30FG, 30HAS, 60FG, and 60HAS ceramics sintered for 1 h at 16001C. HAS, hydrolysis-assisted solidification; GC, gelcasting; XRD, X-ray diffraction.
densities. Conversely, HAS was the less performing consolidation method (TDo91% for ZTA-60 samples and TDo88.5% ZTA-30 samples) probably due to the entrapment of released ammonia gas (AlN13H2O-Al(OH)31NH3) bubbles. Irrespective of the chemical composition and consolidation technique used, all the sintered ceramics exhibit low AP (o0.9%) and WA capacity (o0.8%) values. The results presented in Tables II and III show a close relation between the green and the sintered densities, confirming that porosity elimination upon sintering strongly depends on the homogeneity and particle packing achieved in the green bodies, justifying the less dense materials obtained by HAS. In this case, the decomposition of aluminum hydroxides formed upon hydrolysis of AlN and the oxidation of
Table III. Properties (Bulk Density, Apparent Porosity, Water Absorption Capacity, Linear Shrinkage, Hardness, Fracture Toughness, and Flexural Strength) of ZTA-30 and ZTA-60 Ceramics Consolidated by Various Processing Routes and Sintered for 1 h at 16001Cw
Samplez
30FG 30SC 30HAS 30GC 60FG 60SC 60HAS 60GC w
Bulk density (g/ cm3)
Percent theoretical density (%)y
Apparent porosity (%)
Water absorption capacity (%)
Linear shrinkage (%)z
Hardness (Hv or kg/ mm2)
Fracture toughness (MPa m1/2)
Flexural strength (MPa)
4.23 4.34 4.01 4.30 4.94 4.95 4.65 4.94
93.315 95.742 88.462 94.859 96.371 96.566 90.714 96.371
0.06 0.32 0.90 0.00 0.00 0.40 0.01 0.00
0.01 0.07 0.79 0.00 0.00 0.08 0.02 0.00
14.8070.09 21.5170.77 20.7770.60 20.1270.89 17.0570.08 13.2370.85 18.7270.41 12.3870.93
1493.0750 1435.4724 1315.5727 1440.0729 1246.1761 1263.077.9 1293.6739 1366.5722
4.7670.3 4.6470.19 4.8670.48 4.1270.24 4.7770.19 4.9270.48 5.6570.91 5.9970.81
431.8741.2 409.979.4 463.5716.6 603.9749.7 476.4739.2 509.0755.9 483.0737.4 655.3714.3
See experimental section for the details. The numbers 30 and 60 in the sample codes indicate the zirconia concentration of the composite; in the case of HAS suspensions 5 wt% Al2O3 was replaced by an equivalent amount of AlN. All the suspensions used in the casting processes contained 50 vol% solids. yThe percentage theoretical density values of ZTA-30 (4.553 g/cm3) and ZTA-60 (5.126 g/cm3) were calculated based on the rule of mixture of alumina (3.98 g/cm3) and zirconia (5.89 g/cm3) theoretical densities. z Shrinkage underwent upon sintering. HAS, hydrolysis-assisted solidification; GC, gelcasting. z
14
Vol. 92, No. 1
Journal of the American Ceramic Society—Olhero et al.
(a)
(b)
10 µm
(c)
10 µm
(d)
10 µm
10 µm
(e)
(f)
10 µm
(g)
10 µm
(h)
10 µm
10 µm
Fig. 6. SEM micrographs of 30FG (a), 60FG (b), 30SC (c), 60SC (d), 30HAS (e), 60HAS (f), 30GC (g) and 60GC (h) ceramics sintered for 1 h at 16001C. HAS, hydrolysis-assisted solidification; GC, gelcasting; SEM, scanning electron microscopy.
any nonhydrolyzed AlN will hinder the densification process, further contributing to less dense ceramics.13 However, the measured sintered densities are well comparable to those reported in the literature for similar materials.1–7 A commercial ZTA ceramic (Astro Met, Inc., Springfield Pike Cincinnati, OH) consisting of 15% YPSZ (3% Y2O3 incorporated)185% Al2O3 possesses a BD of 4.1 g/cm3 (TD 5 96.10%), a WA capacity of 0%, hardness of 1750 kg/ mm2, flexural strength of 760 MPa, and a fracture toughness of
6 MPa m1/2. Another commercial ceramic (Dynamic Ceramic Ltd., Crewe, Cheshire) of the sample chemical composition exhibits similar properties but a slightly higher fracture toughness (6–8 MPa m1/2). The percentages of TD achieved in the present work for the 60FG, 60SC, and 60GC samples compare well with those of commercial samples. Figure 5 compares the XRD patterns of 30HAS, 60HAS, 30FG, and 60FG sintered for 1 h at 16001C with the standard XRD data reported in ICDD files for corundum (a-Al2O3, No.
January 2009
15
Aqueous Colloidal Processing of ZTA Composites
(a)
100 μm
(b)
85.7 μm
values than ZTA-30 samples. This can be attributed to the toughening mechanisms identified in zirconia-reinforced alumina (ZTA) ceramics, namely the volume expansion ( 4%) and the shear strain ( 6%) associated with the t-m transformation of zirconia that reduces the propagation of the cracks.1–8,40 Microcracks (either residual or stress induced) extending in the stress field of a propagating crack can absorb the fracture energy, increasing the material toughness by the microcrack toughening mechanism. Cracks can also be deflected by localized stress fields developed as a result of phase transformation, thermal expansion mismatch, or by fracture of second phase particles, thus contributing to the increase of fracture toughness. The indentation micrographs of 30GC and 60GC samples are presented in Figs. 7(a) and (b), respectively. Different magnifications were used to show the crack lengths clearly. Anisotropy in the fracture seems to be absent as cracks extending from the four corners are similar in all the tested samples. It can be seen that crack lengths are shorter in 60GC ceramics than in 30GC ones. Moreover, GC samples showed shorter crack lengths in comparison with those consolidated by HAS, SC, and FG, indirectly suggesting that enhanced homogeneity has been achieved by GC. The mechanical properties reported in Table III reveal some expected trends: higher hardness and lower toughness and flexural strength values for the richer in alumina samples, and the opposite situation for the samples richer in zirconia. A detailed analysis of the mechanical data for each composition shows that ceramics consolidated by SC and GC routes tend to exhibit better properties and similar distributions of the alumina and zirconia grains in the sintered microstructures (Fig. 6). This is consistent with the higher degrees of homogeneity achieved by these two processing routes.41,42 Among the various shaping techniques, the best green and sintered mechanical properties were obtained for ceramics consolidated by aqueous GC.
Fig. 7. Indentation micrographs of (a) 30GC and (b) 60GC samples. GC, gelcasting.
00-046-1212), monoclinic zirconia (m-ZrO2, No. 00-037-1484), and tetragonal yttria partially stabilized zirconia {[(ZrO2)0.91(Y2O3)0.09]0.917, No. 01-083-0113} (YPSZ). The main XRD lines of all sintered materials belong to corundum or to YPSZ phases. This last phase undergoes stress-induced transformation, improving the fracture toughness and the strength of the material. No lines belonging to AlN or other compounds that could be formed by reacting AlN with corundum (AlON) or ZrO2 are seen, indicating that AlN has been totally hydrolyzed/oxidized upon sintering. As expected, zirconia peaks of ZTA-60 samples are stronger than those of corundum. The SEM micrographs of 30FG, 60FG, 30SC, 60SC, 30HAS, 60HAS, 30GC, and 60GC, sintered for 1 h at 16001C, are presented in Figs. 6(a)–(h), respectively. The white portions correspond to zirconia, and the gray/dark ones to alumina. These micrographs reveal four distinct features. (1) The fractions of white grains are higher in the ZTA-60 ceramics, as expected. (2) No other intergranular phases besides alumina and zirconia can be seen. (3) ZTA-30 ceramics exhibit a higher number of flaws and defects in comparison with the higher density ZTA-60 ceramics. (4) 30FG and 60FG ceramics contain fine alumina grains, less open porosity, and rather irregularly distributed large segregated ZrO2 grains in comparison with ceramics consolidated by casting routes. The ceramics consolidated by HAS show some pores and some coarse alumina grains that can be attributed to local enrichment of alumina due to the hydrolysis of AlN. Ceramics consolidated by SC and GC routes exhibit more uniform distributions of alumina and zirconia grains in comparison with those consolidated by other techniques. Table III shows that the ZTA-60 samples (except those consolidated by HAS) exhibit higher fracture toughness and flexural strength values and, as a consequence, lower hardness
IV. Conclusions Different types of dense ZTA ceramics with potential for orthopedic applications could be consolidated by colloidal processing techniques such as SC, HAS, and GC processes. ZTA ceramics consolidated by GC and sintered for 1 h at 16001C exhibit uniformly distributed alumina and zirconia grains and improved green and sintered mechanical properties in comparison with those consolidated by SC, HAS, and conventional die pressing of freeze-dried granules. The replacement of 5 wt% Al2O3 by an equivalent amount of AlN to promote consolidation via HAS seems to be excessive for the systems richer in alumina(ZTA-30), tending to degrade the green and sintered properties of the consolidated bodies.
Acknowledgments IG thanks SERC-DST (Government of India) for the BOYSCAST fellowship awarded(SR/BY/E-04/06). The first author wishes to thanks FCT (Foundation for Science and Technology) for the Finantial support under the grant BPD/SFRH/ 27013/2006.
References 1
N. Claussen, ‘‘Fracture Toughness of Al2O3 with an Un-Stabilized ZrO2 Dispersed Phase,’’ J. Am. Ceram. Soc., 59, 49–51 (1976). 2 J. Wang and R. Stevens, ‘‘Review—Zirconia-Toughened Alumina (ZTA) Ceramics,’’ J. Mater. Sci., 24 [10] 3421–40 (1989). 3 K. Tsukuma, K. Ueda, and M. Shimada, ‘‘Strength and Fracture Toughness of Iso-Statically Hot-Pressed Composites of Al2O3 and Y2O3Partially-Stabilized ZrO2,’’ J. Am. Ceram. Soc., 68 [1] C4–5 (1985). 4 K. Tsukuma, T. Takahata, and M. Shiomi, ‘‘Strength and Fracture Toughness of Y-TZP, Ce-TZP, Y-TZP/Al2O3, and Ce-TZP/Al2O3’’; pp. 721–728 in Advanced in Ceramics, Vol 24: Science and Technology of Zirconia III, Edited by S. Somiya, N. Yamamoto, and H. Yamagida. The American Ceramic Society, Columbus, OH, (1988). 5 W. H. Tuan, R. Z. Chen, T. C. Wang, C. H. Cheng, and P. S. Kuo, ‘‘Mechanical Properties of Al2O3/ZrO2 Composites,’’ J. Eur. Ceram. Soc., 22 [16] 2827–33 (2002).
16 6
Journal of the American Ceramic Society—Olhero et al.
Y. Shin, Y. Rhee, and S. Kang, ‘‘Experimental Evaluation of Toughening Mechanism in Alumina–Zirconia Composites,’’ J. Am. Ceram. Soc., 82 [5] 1229– 32 (1999). 7 R. H. J. Hannink, P. M. Kelly, and B. C. Muddle, ‘‘Transformation Toughening in Zirconia-Containing Ceramics,’’ J. Am. Ceram. Soc., 83 [3] 461–87 (2000). 8 D. Casellas, M. M. Nagl, L. Llanes, and M. Anglada, ‘‘Fracture Toughness of Alumina and ZTA Ceramics: Micro-Structural Coarsening Effects,’’ J. Mater. Proc. Tech., 143–144, 148–52 (2003). 9 M. A. Janney, S. D. Nunn, C. A. Walls, O. O. Omatete, R. B. Ogle, G. H. Kirby, and A. D. McMillan, ‘‘Gelcasting’’; pp. 1–15 in The Handbook of Ceramic Engineering, Edited by M. N. Rahman. Marcel Dekker, New York, 1998. 10 G. Tari, ‘‘Gelcasting Ceramics: A Review,’’ Am. Ceram. Soc. Bull., 82 [4] 42–6 (2003), and references therein. 11 T. Kosmac, S. Novak, and M. Sajko, ‘‘Hydrolysis-Assisted Solidification (HAS): A New Setting Concept for Ceramic Net-Shaping,’’ J. Eur. Ceram. Soc., 17, 427–32 (1997). 12 I. Ganesh, N. Thiyagarajan, D. C. Jana, P. Barik, G. Sundararajan, and J. M. F. Ferreira, ‘‘Dense b-SiAlONs Consolidated by a Modified Hydrolysis Assisted Solidification Route,’’ J. Eur. Ceram. Soc., 28 [4] 879–85 (2008). 13 I. Ganesh, N. Thiyagarajan, D. C. Jana, P. Barik, and G. Sundararajan, ‘‘An Aqueous Gelcasting Route to Dense b-Si4Al2O2N6-0.5SiO2 Ceramics,’’ J. Am. Ceram. Soc., 91 [5] 1566–71 (2008). 14 I. Ganesh, D. C. Jana, S. Shaik, and N. Thiyagarajan, ‘‘An Aqueous Gelcasting Process for Sintered Silicon Carbide Ceramics,’’ J. Am. Ceram. Soc., 89 [10] 3056–64 (2006). 15 K. W. Kirby, A. T. Jankiewicz, R. F. Lowell, and R. L. Hallse, ‘‘Near Net Shape Fabrication of Ceramic Radomes’’; U.S. Patent No: 6,083,452, July 4 2000. 16 J. S. Reed, Introduction to the Principles of Ceramic Processing. John Wiley and Sons, New York, 1988. 17 J. M. F. Ferreira, ‘‘Role of the Clogging Effect in the Slip Casting Process,’’ J. Eur. Ceram. Soc., 18, 1161–9 (1998). 18 R. Moreno, A. Salomoni, and I. Stamenkovic, ‘‘Influence of Slip Rheology on Pressure Casting of Alumina,’’ J. Eur. Ceram. Soc., 17, 327–31 (1998). 19 J. M. F. Ferreira and H. M. M. Diz, ‘‘Effect of Driving Force on Pressure Slip Casting of Silicon Carbide Bodies,’’ J. Eur. Ceram. Soc., 18, 1171–5 (1998). 20 J. R. G. Evans, ‘‘Injection Moulding’’; in Materials Science and Technology the Processing of Ceramics, Vol. 17A, Edited by R. J. Brook. VCH, Weinheim, 1996 (Chapter 8). 21 T. J. Graule, F. H. Baaber, and J. L. Gauckler, ‘‘Shaping of Ceramic Green Compact Direct from Suspensions by Enzyme Catalyzed Reactions,’’ Cfi/Ber. DKG, 71, 317–23 (1994). 22 O. Lyckfeldt and J. M. F. Ferreira, ‘‘Processing of Porous Ceramics by ‘‘Starch Consolidation’’,’’ J. Eur. Ceram. Soc., 18, 131–40 (1998). 23 H. M. Alves, G. Tari, A. T. Fonseca, and J. M. F. Ferreira, ‘‘Processing of Porous Cordierite Bodies by ‘‘Starch Consolidation’’,’’ Mat. Res. Bull., 33, 1439– 48 (1998). 24 B. E. Novich, C. A. Sundback, and R. W. Adams, ‘‘Quicksett Injection Moulding of High Performance Ceramics,’’ Ceram. Trans., 26, 157 (1992).
25
Vol. 92, No. 1
R. D. Rives, ‘‘Method of Injection Moulding Powder Metal Parts’’; U.S. patent 4113480 1976. 26 A. J. Fanelli, R. D. Silvers, W. S. Frei, J. V. Burlew, and G. B. Marsh, ‘‘New Aqueous Injection Moulding Process for Ceramic Powders,’’ J. Am. Ceram. Soc., 72, 1833–6 (1989). 27 P. J. Bendeich and P. A. Walls, ‘‘Aqueous Milling and Near Net Shape Forming of SiAlON Ceramics’’; pp. 627–32 in Proceedings of the 2nd International Meeting of Pacific RIM Ceramic Societies, Vol. 2, Edited by P. Walls, C. Sorrel, and A. Ruys. Australian Ceramic Society, Queensland, Australia, 1996. 28 I. Ganesh, S. M. Olhero, A. H. Rebelo, and J. M. F. Ferreira, ‘‘Formation and Densification Behavior of MgAl2O4 Spinel: The Influence of Processing Parameters,’’ J. Am. Ceram. Soc., 91 [6] 1905–11 (2008). 29 I. Ganesh, K. A. Teja, N. Thiyagarajan, R. Johnson, and B. M. Reddy, ‘‘Formation and Densification Behavior of Magnesium Aluminate Spinel: The Influence of CaO and Moisture in the Precursors,’’ J. Am. Ceram. Soc., 88 [10] 2752–61 (2005). 30 B. D. Cullity, Elements of XRD, 2nd edition, Addison-Wesley, Reading, MA, 1978. 31 H. P. Klug and L. E. Alexander, ‘‘X-Ray Diffraction Procedures for Polycrystalline and Amorphous Materials,’’ J. Appl. Cryst., 8, 573 (1975). 32 G. R. Anstis, P. Chantikul, B. R. Lawn, and D. B. Marshall, ‘‘A Critical Evaluation of Indentation Techniques for Measuring Fracture Toughness: I Direct Crack Measurements,’’ J. Am. Ceram. Soc., 64, 533–8 (1981). 33 G. A. Parks, ‘‘The Iso-Electric Points of Solids Oxides, Solid Hydroxides, and Aqueous Hydroxy Complex Systems,’’ Chem. Rev., 65 [2] 177–98 (1965). 34 Y. K. Leong, P. J. Scales, T. W. Healy, and D. V. Boger, ‘‘Effect of ParticleSize on Colloidal Zirconia Rheology at the Isoelectric Point,’’ J. Am. Ceram. Soc., 78, 2209–13 (1995). 35 L. Bergstrom, ‘‘Hamaker Constants of Inorganic Materials,’’ Adv. Colloid Interface Sci., 70, 125–69 (1997). 36 G. Tarı´ , J. M. F. Ferreira, and O. Lyckfeldt, ‘‘Influence of the Stabilising Mechanism and Solid Loading on Slip Casting of Alumina,’’ J. Eur. Ceram. Soc., 18, 479–86 (1998). 37 S. M. Olhero, G. Tari, M. A. Coimbra, and J. M. F. Ferreira, ‘‘Synergy of Polysaccharide Mixtures in Gelcasting of Alumina,’’ J. Eur. Ceram. Soc., 20, 423– 9 (2000). 38 S. Novak, T. Kosmac, K. Krnel, and G. Draz, ‘‘Principles of the Hydrolysis Assisted Solidification (HAS) Process for Forming Ceramic Bodies from Aqueous Suspension,’’ J. Eur. Ceram. Soc., 22, 289–95 (2002). 39 K. Krnel and T. Kosmac, ‘‘Protection of AlN Powder Against Hydrolysis Using Aluminum Di-Hydrogen Phosphate,’’ J. Eur. Ceram. Soc., 21, 2075–9 (2001). 40 M. A. Janney, C. A. Walls, D. M. Kupp, and K. W. Kirby, ‘‘Gelcasting SiAlON Radomes,’’ Am. Ceram. Soc. Bull., 83, 9201–6 (2004). 41 S. Deville, J. Chevalier, G. Fantozzi, J. F. Bartolome, J. Requena, J. S. Moya, R. Torrecillas, and L. A. Diaz, ‘‘Low-Temperature Ageing of Zirconia-Toughened Alumina Ceramics and its Implication in Biomedical Implants,’’ J. Eur. Ceram. Soc., 23, 2975–82 (2003). 42 F. F. Lange, B. I. Davis, and E. Write, ‘‘Processing-Related Fracture Origins: IV, Elimination of Voids Produced by Organic Inclusions,’’ J. Am. Ceram. Soc., 69 [1] 66–9 (1986). &
