AU1651399A - Dense refractories with improved thermal shock resistance - Google Patents

Description

WO99/32417 PCT/AU98/01049 DENSE REFRACTORIES WITH IMPROVED THERMAL SHOCK RESISTANCE The present invention relates to a refractory material and 5 to a method of manufacturing the refractory material. A simple definition of a refractory material is one which resists the effects of high temperatures. Commonly, the term refractory material is applied to relatively low cost 10 products that are used in many industrial processes, typically operating at high temperatures, to contain corrosive materials, such as molten metal and slags. As such refractories are an important class of materials. 15 The following factors are relevant to the design of refractory materials: chemical compatibility; thermal shock; 20 constraints on start-up; operating conditions; slag penetration; hot strength creep resistance; and 25 cost. Many ceramics materials have properties in common with refractory materials. For example, ceramic materials are characterised by excellent chemical stability, high 30 hardness and a brittle nature. In comparison with refractory materials, typically ceramic materials have poor thermal shock resistance. The combination of poor thermal shock resistance and high cost limits the use of ceramic materials in refractory applications. 35 There are two options to minimise the effects of thermal shock. The first is to avoid the initiation of cracks and WO99/32417 PCT/AU98/01049 -2 the second is to avoid catastrophic crack propagation. Thermal Shock Damage Resistance Parameters (R), a measure of a material's resistance to the above types of failure, were proposed by Hasselman (see Introduction to Ceramics, 5 Kingery 2 md edition 1976 pp 825-30). The physical properties required to compute Thermal Shock Damage Resistance Parameters are thermal conductivity k, thermal expansion coefficient a, Young's Modulus E, effective fracture energy .eff, and strength (MOR) T. Specifically, 10 the Thermal Shock Damage Resistance Parameters R' and R''' can be expressed as: ka R'- Ea(1) 15 and R 2 (2) where R' is the parameter for the resistance to crack 20 initiation and R'''' is the parameter for the resistance to crack propagation. The material characteristics for inhibiting crack formation are high strength with respect to elastic modulus. The 25 requirements for minimising the extent of crack propagation are a high product of work of fracture and elastic modulus with respect to strength. Thus, the design requirements for a material for inhibiting crack formation and crack propagation are different. 30 It is known that resistance to catastrophic failure, which is required in refractory applications, can be improved by the introduction of enough cracks of sufficiently large size so that crack propagation takes place semi-statically.

WO99/32417 PCT/AU98/01049 - 3 It is also known that, alternatively, resistance to catastrophic failure can be achieved by the introduction of microstructural inhomogenieties in any form which serve as stress concentrators in the material. In this way, cracks 5 will form locally, but catastrophic failure is avoided as a result of the small average stress in the material. Conventional refractory materials are designed for chemical stability, thermal shock resistance, and cost. This is 10 achieved through a compromise between reducing the effective surface area for attack and increasing resistance to crack propagation. Typically, a conventional refractory material has an open structure with between 15 and 20% porosity. The open structure allows rapid penetration of 15 slags and gases but inhibits crack propagation. A schematic representation is shown in Fig. 1. The shortcomings of this compromise approach to design were recognised by the late Ronald C. Garvie. He proposed that a 20 dense thermal shock resistant material would offer superior performance to a conventional refractory material. To achieve this goal he introduced micro-cracks into the microstructure. This increased the work of fracture for the material by promoting crack branching. The end result was a 25 dense material with the chemical stability of an advanced ceramic and the thermal shock resistance of a porous refractory. This micro-crack toughened composite material is disclosed in US patents 5,296,420 and 5,334,563 of Garvie. A schematic representation of the composite 30 refractory material is shown in Fig. 2. The essential features of the composite material disclosed in US Patent 5,334,563 are that the material have less than 12% porosity and comprise: 35 a matrix of alumina, with 5 to 90% by volume of the alumina grains having a diameter in the range WO99/32417 PCT/AU98/01049 -4 of 15 to 80 microns; particles of monoclinic zirconia dispersed in the matrix, each dispersed particle comprising an 5 agglomerate of microcrystals which; (a) are strongly bonded together; (b) exhibit a strong thermal expansion 10 anisotropy; and (c) a size such that cracks do not form spontaneously within the agglomerates after cooling from high temperatures in the range 15 of 1600oC; and the alumina and the monoclinic zirconia being chemically inert with respect to each other within the temperatures used in practice. 20 The Garvie US Patent also discloses a number of other combinations, such as: mullite as the matrix and zirconia as the dispersed material; silicon nitride as the matrix and boron nitride as the dispersed material; barium 25 titanate as the matrix and zirconia as the dispersed material; silicon carbide as the matrix and boron nitride as the dispersed material; alumina as the matrix and aluminium titanate as the dispersed material; spinel as the matrix and zirconia as the dispersed material; and 30 fosterite as the matrix and zirconia as the dispersed material. The basis of the Garvie US Patent is the addition of a dispersed second phase in a continuous dense matrix with 35 very particular inter-dependence of the respective thermal expansion coefficients of the phases. Specifically, the use of specific grades of monoclinic zirconia as the WO99/32417 PCT/AU98/01049 - 5 dispersed phase produced an enhanced dilatational/contractional mismatch in a number of matrices such as alumina or zircon. An optimised composition, with respect to thermal shock damage resistance (measured by 5 retained strength) was determined empirically by Garvie to be 8% by weight of zirconia in alumina and 10% by weight in zircon. Extensive chemical attack of ceramic matrix materials, such 10 as zircon and alumina, limits the use of such ceramic composites in many corrosive industrial applications. These include applications where the ceramic composite is in contact with slags used in iron and steel making operations. 15 A further significant problem is the prohibitive cost of production of the composite ceramic materials on an industrially realistic scale. 20 It is known that reaction sintering of zircon mixtures can result in the formation of oxide zirconia dispersions (see for example US Patent 2842447 by Schlotzhauer and Wood and Cambier, Baudin de La Lastra, Pilate and Leriche Brit. Ceram. Soc. Trans. and J. 83 pp 196-200, 1984). As 25 discussed by Cambier et. al. the use of this technique is useful in the manufacture of zirconia in a mullite or alumina mullite matrix. These materials are characterised by high strength with MOR values that can reach 400 to 500MPa. In addition, these materials are typically 30 characterised by pores around the zirconia particles as a consequence of the process. That is, the original zircon particles lose silica to the surrounding matrix. There is a volume decrease reported to be about 20% for the zircon particles converted into zirconia particles. This results 35 in the formation of pores associated with the zirconia grains. Furthermore, Schlotzhauer and Wood (col. 3, lines 11-20 of the US Patent) indicate that the high corrosion WO99/32417 PCT/AU98/01049 - 6 resistance of the final products is a consequence of the lack of cracking associated with the inversion of the zirconia as it is heated or cooled through 1000 0 C. The presence of such pores would accommodate the volume 5 expansion of zirconia associated with the inversion of the zirconia on cooling without the generation of stresses or strains. US Patent 4298385 of Claussen and Steeb discloses a method 10 for producing bodies having high fracture toughness and "substantially equal" mechanical strength. This is achieved by the addition of from 4 to 25 volume % zirconia grains ("embedment material") with a diameter from 0.3 to 1.25 pm in an anisotropic ceramic matrix, such as alumina. The 15 improvement in the properties of the fabricated products resulted, by way of example in the case of alumina with unstabilised zirconia, from the production of extremely fine micro-fissures and a high fissure density in the products. This was reported to significantly increase the 20 toughness, thermal shock resistance and impact strength as compared to products prepared without the zirconia addition. In addition, it was found that it was preferable to disperse the zirconia within agglomerates of zirconia and matrix phase with a size of 2 to 15 pm containing from 25 4 to 25 volume % (preferably 8 to 25 volume %) of the phase (Col 2, lines 9 to 54). For alumina this is equivalent to 5.8 to 32.8 wt% and preferably 11.3 to 32.8wt%. Furthermore, it is also taught that the use of large embedment material is to be avoided, as the strength is 30 considerably reduced (Col 4, lines 43 to 46). From the results presented in Figure 1 of the US Patent it is clearly seen that increasing the particle size embedment materials from 0.3pm to 1.25pm required an increase in the amount of the embedment material from 10 vol% to 15 vol% 35 (14 to 20.6 wt%) and this indicates the benefits of the smaller zirconia grain size. The examples disclose the use of high vol% of the embedment phase. For example in Fig. 6 WO99/32417 PCT/AU98/01049 -7 the vol% ranges from 15 to 25 vol%. It is further reported that such materials are especially suited to high temperature gas turbine elements. 5 US Patent 4,804,644 of Anseau, Lawson and Slasor also discloses a material which includes dispersion of zirconia in a matrix, in this case an O'-sialon matrix. O'-sialon is a solid solution based on silicon oxynitride (Si 2

N

2 0) where there is substitution of Al and 0 for Si and N 10 respectively. The US Patent discloses a number of methods for the preparation of such materials. However, for materials produced according to the methods the zirconia is in the tetragonal form. It is stated that improvements in properties would result from the transformation of meta 15 stable tetragonal zirconia to the monoclinic form in response to a tensile stress typically caused by an advancing crack tip. The transformation results in the formation of compressive stresses that tend to close the cracks. Indeed from example 1E of the US Patent, the 20 zirconia is reported to be in the tetragonal form at room temperature. For the zirconia to be effective the size of the particles must remain small to prevent spontaneous transformation on cooling. There is no report of the physical properties such as strength and thermal shock 25 resistance of such bodies formed. An object of the present invention is to provide a refractory material with enhanced corrosion, erosion and thermal shock resistance which alleviates the disadvantages 30 of the known refractory materials discussed above. According to one aspect of the present invention there is provided a dense refractory material which includes a matrix and a micro-crack initiating single crystal phase formed from 35 fused zirconia dispersed in the matrix. The term "dense" is understood herein to mean that the WO99/32417 PCT/AU98/01049 - 8 refractory material has limited open porosity, typically less than 5% by volume. According to another aspect of the present invention there is 5 provided a dense refractory material which includes a spinel matrix. The spinel group of materials is understood herein to mean materials that are described by the general formula: 10

AB

2 04 where A 2 is typically is either singly or in combination Mg, Fe, Zn and Mn and B 3 is typically either singly or in 15 combination Al, Fe, Cr and Mn. Examples of spinels are magnesium aluminium oxide MgA120 4 , magnetite Fe 3

O

4 , and chromite FeCr 2 0 4 . An example of a "mixed" spinel is Mg(Al,Fe) 2 0 4 . 20 The spinel group of materials have a cubic crystal structure and, therefore, are isotropic. As a consequence, the spinal microstructure is relatively stress-free. 25 Furthermore, the spinel group of materials is relatively stable at high temperatures and while maintained at temperature. The spinel may include one or more additional elements. The 30 additional elements may include Li, Mg, Ca, Ti, Mn, Fe, Co, Ni, Cu, Zn, Sr and Ba, for divalent cations and Al, Cr, Fe, and Mn as the trivalent cations. In addition, spinel phases can exist over a range of compositions with respect to the ratio of the divalent to trivalent cations. 35 The selection of the additional elements may depend on a wide range of factors. By way of example, one factor is WO99/32417 PCT/AU98/01049 -9 the environment in which the refractory material will be used. Specifically, in situations where the refractory material will be in contact with molten slag in metal smelting operations, the additional elements may be 5 selected to optimise the chemical stability of the refractory materials with respect to the slag. By way of further example, another factor is to include additional elements to assist in the manufacture of the refractory material as a dense refractory material. 10 A further advantage of spinels is that they can exist over a range of composition without a change in phase. For example, magnesium aluminate spinel can be magnesium rich, stoichiometric (Mg to Al ratio of 1:2) or aluminium rich. 15 This allows the loss of an element from the crystal lattice without decomposition to form a new phase or compound. Typically, the formation of new phases can result in physical disruption of the refractory body or the formation of less refractory phases. The ability of the spinel to 20 adapt to the environment without a change in phase enhances the stability of the products. It is known that spinels, such as magnesium aluminium oxide MgAl 2 0 4 and chromite FeCr 2 0 4 spinels, have excellent 25 corrosion resistance to slags in metal smelting operations. However, typically, the spinels are coarse and are used as grits or aggregate in refractory bodies for many metal making and cement making operations and not as the matrix of a dense refractory material. Moreover, the refractories 30 that incorporate these spinels are in the form of traditional refractories that are characterised by open porosity and are not dense refractory materials. Furthermore, whilst the Garvie US patents propose the use of 35 spinels in a matrix of a micro-crack toughened refractory material, the disclosure is speculative and not supported by examples.

WO99/32417 PCT/AU98/01049 - 10 It is preferred that the refractory material further comprises a micro-crack initiating phase dispersed in the matrix. 5 It is preferred that the micro-crack initiating phase be no more than 15% by volume of the material. It is preferred particularly that the micro-crack initiating 10 phase be no more than 10% by volume of the material. It is preferred that the spinel matrix be at least 80% by volume of the material. 15 It is preferred particularly that the spinel matrix be at least 90% by volume of the material. It is preferred that the micro-crack initiating phase comprises a dispersion of single crystals. 20 It is preferred that the micro-crack initiating phase be formed from zirconia. It is preferred that the zirconia have a particle size in the 25 range of 5 to 50pm. It is preferred that the zirconia have a particle size in the range of 10 to 20Lm. 30 It is preferred particularly that the zirconia be fused zirconia. The micro-crack initiating phase may be formed from any other suitable material, such as boron nitride and silicon carbide. 35 It is preferred that the spinel be manufactured from low cost precursors.

WO99/32417 PCT/AU98/01049 - 11 According to another aspect of the present invention there is provided a dense refractory material which includes a spinel matrix and a micro-crack initiating phase dispersed 5 in the matrix. According to another aspect of the present invention there is also provided a method of manufacturing a dense refractory material product which includes the steps of: 10 (i) mixing precursor oxides for a spinel material; (ii) calcining the mixture to form the spinel 15 material; (iii)forming the spinel material into a green form of the product; and 20 (iv) firing the green form of the product to produce the final form of the product. It is preferred that the method further includes the step of mixing the spinel material produced in step (ii) with an 25 additive, such as zirconia, selected to form a micro-crack initiating phase dispersed in the fired product. According to the present invention, the spinel material is formed by reaction of the precursor oxides. This is typically 30 carried out in the temperature range of 800 0 C to 1600 0 C and preferably in the range of 1000 0 C to 1400 0 C for dwell times at temperature ranging up to at least 10 hours. Longer times are generally preferred for lower calcination temperatures and shorter times for temperatures in the upper reaches of 35 the range. Dwell times of 1 hour or less are possible for higher temperatures in the range.

WO99/32417 PCT/AU98/01049 - 12 Typically, the spinel material formed is then milled (if necessary) to form a finely divided powder suitable for densification in the secondary heat treatment of step (iv). Typically, the average particle size should be less than 10 5 pm, preferably less than 5pm and more preferably less than 2pm. Typically, the additive which forms the dispersed phase is then added to the spinel powder. 10 Typically, the spinel powder and the additive are then moulded or formed into the desired shape in a green form in step (iii). This can be done with and without the use of additives to increase the plasticity of the powder 15 facilitating forming into the desired "green" shapes. The green shape is then heated to effect densification in the firing step (iv). This is typically carried out in the temperature range of 1000 0 C to 1800oc and preferably in the 20 range of 1400 0 C to 1600 0 C for dwell times at temperature ranging up to at least 10 hours. Longer times are generally preferred for lower secondary heating temperatures and shorter times for temperatures in the upper reaches of the range. Dwell times of 1 hour or less are possible for higher 25 temperatures in the range. Temperatures can also be reduced by use of sintering assists that can be incorporated into the structure of the spinel. However, it is preferable that the firing temperature used in the manufacture be at least as high as the expected operating temperature where the 30 refractory is to be used. Sintering aids may be used to promote densification of the refractory material. These aids can form liquids that result in enhanced diffusion rates thereby increasing the 35 densification rate. Where these additives exist as secondary phases in the final microstructure they can exert a deleterious effect on the performance of products. It is well WO 99/32417 PCT/AU98/01049 - 13 known that the presence of silica-based glasses and calcium containing phases can lead a marked decrease in the high temperature properties of alumina based refractories. 5 Appropriate sintering aids may be used to promote densification at lower temperatures without a loss of performance. The firing cycle of refractory materials can represent a substantial proportion of the cost to manufacture products. Reducing the firing temperature can result in a 10 lower cost to manufacture products. In addition, it is postulated that improved chemical stability is obtained by using a matrix material that contains the main elements of the slag in a solid solution within the crystal structure of the matrix phase or where stable phases are produced as a 15 result of the interaction of elements in the slag with the matrix. The dense refractory material of the present invention contains micro-cracks in the microstructure after 20 fabrication. These micro-cracks are characterised by emanating from the dispersed phase (typically formed by zirconia additions) and extending over several grain diameters in the microstructure. Typically the grain size is the order of greater than 10 pm. 25 As stated above, the spinel group of materials is defined by the general formula AB 2 0 4 where A 2 " is typically is either singly or in combination Mg, Fe, Zn and Mn and B 3 is typically either singly or in combination Al, Fe, Cr and 30 Mn. The spinel may include one or more additional elements. The additional elements may include Li, Mg, Ca, Ti, Mn, Fe, Co, Ni, Cu, Zn, Sr and Ba, for divalent cations and Al, Cr, Fe, and Mn as the trivalent cations. In addition, spinel phases can exist over a range of compositions with respect 35 to the ratio of the divalent to trivalent cations. As indicated above, zirconia is the preferred additive to WO99/32417 PCT/AU98/01049 - 14 for the dispersed phase. It is well known that certain dopants can stabilise the high temperature crystallographic forms of zirconia at room temperature. Typically, dopants include magnesia, ceria, yttria and calcia. For the case 5 of spinels that typically include elements that can stabilise the high temperature forms of zirconia it is surprising that products formed according to the present invention do not contain zirconia particles that are stabilised. Such stabilisation would render the micro 10 cracking mechanism responsible for the improvement in thermal shock inoperative. From thermodynamic considerations it is believed likely that the dopants for zirconia would be partitioned between the zirconia and the matrix. With the exception of certain instances in the 15 finer fractions, this it not observed. Furthermore it is believed that large particle size and hence unreactivity of the zirconia particles is responsible in part or whole for the observed behaviour. This indicates the requirement for large unreactive zirconia particles. Another advantage of 20 such particles is the relative low cost of large unreactive powders and grits as compared to finely divided and reactive powders. Such fine fractions maybe advantageous for other reasons but they do not contribute to the improvement in thermal shock behaviour for refractory 25 bodies as described by the current invention. The present invention overcomes the problems of obtaining low cost refractory materials with high erosion and corrosion stability. 30 In relation to the prior art discussed above, the process disclosed by Schlotzhauer and Wood and Cambier is clearly different to that of the present invention. In the present invention, cracks are deliberately created by the inclusion 35 of the zirconia (or other dispersed micro-crack initiating phase). That is, the stresses and strains associated with the inclusion of zirconia into the matrix can not be WO99/32417 PCT/AU98/01049 - 15 accommodated by the matrix and result in the formation of cracks. A feature of the present invention is the tolerance to 5 impurities and the fact that low cost refractory grade precursors can be used. This allows the use of low cost refractory precursors. It is speculated that the finer fractions of the zirconia materials used are able to react with impurities to produce more refractory phases. 10 In addition, the body disclosed by Claussen and Steeb is substantively different to that in the present invention. The Claussen and Steeb body retains high fracture strength and fracture toughness. This is achieved by the requirement 15 for the use of a large vol% of micron and preferably sub micron zirconia material. The materials produced according to the teachings of the present invention are for refractory type applications. A requirement for this type of material is relatively low cost. This typically means 20 below US$5,000 per tonne for the finished product. Such a final price requires the use of inexpensive raw materials. Sub-micron zirconia powders are expensive. At the current prices for such zirconia powders at the levels required according to the teachings of Claussen and Steeb would 25 equal or in some cases exceed the price of the final product produced according to the present invention. In the present invention the amount of zirconia addition is minimised, thereby allowing bodies to be cost competitive with conventional refractory materials. For materials of 30 the present invention, the strength of the final bodies is sacrificed for improved thermal shock resistance. Materials of the present invention would not be suitable for applications such as turbine blades. However, bodies as described according to the present invention are eminently 35 suitable for applications were high chemical resistance and thermal shock resistances are required but where high strength is not required such as refractory applications.

WO99/32417 PCT/AU98/01049 - 16 It is important to note that in the present invention the presence of <5pm fraction of zirconia does not enhance the thermal properties of the refractories. This is 5 demonstrated when MEL S zirconia particles that contain ~2% particles less than Slim. In the case of the AFM ZCO3 material the percentage of <Slim zirconia particles is increased to >20%. However, this does not lead to an increase in performance in terms of thermal shock. In fact, 10 more ZCO3 material is required as compared to the MEL S zirconia source. This is attributed to the finer fractions being inoperative for the improvement in thermal shock according to the present invention. In addition, it has been observed that the finer fractions of zirconia are 15 preferably attacked in contact with slags typically encountered in operating conditions. The presence of the finer fractions as taught by Claussen and Steeb are clearly inferior for the applications intended for the current materials for the reasons outlined. 20 The following examples are used to describe the invention in a non limiting manner. EXAMPLES 25 Examples 1-9 The object of the Examples 1-9 was to compare the performance of a micro-crack toughened refractory material in accordance 30 with the present invention which includes a low cost single crystal fused zirconia (AFM Grade 3) dispersed phase with a known micro-crack toughened composite material based on agglomerates of monoclinic zirconia (MEL S) proposed by Garvie. 35 The raw materials used were as follows: WO99/32417 PCT/AU98/01049 - 17 Raw Materials Component Supplier A1 2 0 3 Alcoa (A1000) ZrO 2 MEL Grade S or AFM Grade ZCO3 The batch size was a nominal 200g. The batches containing 5 MEL S were designated Examples 1-5 and 6-9 for the AFM containing range. The starting compositions are given in the following table. Starting Composition in Parts 10 Example Batch MEL S ZrO 2 AFM ZrO 2 A1 2 0 3 1 CMA07 0 - 100 2 CMA08 4 - 96 3 CMA09 6 - 94 4 CMA10 8 - 92 5 CMA11 10 - 90 6 CMA13 - 4 96 7 CMA14 - 6 94 8 CMA15 - 8 92 9 CMA16 - 10 90 The alumina powder was combined with zirconia in the proportions given using the milling conditions as outlined in the following table. The objective for all batches was 15 to thoroughly distribute the ZrO 2 rather than reduce the particle size. See the following table for details.

WO99/32417 PCT/AU98/01049 - 18 Ball Milling Conditions Time 0.5 hours Powder 0.2 kg Balls 0.5 kg YTZP Fluid 0.3 1 iso propanol Binder 2.0/2.0g Surcol/Glycerol The resultant slurries were dried at 80 0 C. Segregation of 5 the constituents was avoided by ensuring that slurry viscosity remained high and by use of a shallow drying pan. Bars were pressed from the dried powder with a geometry suitable for strength testing (MOR), Young's Modulus and work of fracture testing (WOF). These were formed by die 10 pressing at a pressure of 55MPa with a bar die of dimensions 5 x 51mm for MOR bars and 7.5 x 102mm for WOF bars. The bars were then bagged and cold isostatically pressed to a pressure of 210MPa. The samples were fired in air on an alumina setter plate. The firing cycle used is 15 given in the following table: Firing Cycle in Air Heat 800C.h -1 Dwell 1600 0 C for 2 h Cool 1000C.h - 1 to room temperature 20 All bars were diamond ground to the dimensions required for compliance to the ASTM E399-83 test for work of fracture testing (WOF). This was a nominal 5 x 10 x 85mm. Bars for strength testing (MOR) were also machined in compliance with the ASTM Standard C1161-94 for flexural testing. 25 Dimensions for these bars were a nominal 3 x 4 x 45mm. For the determination of Young's Modulus, the ends of the WOF bars were ground square. Young's Modulus was determined WO99/32417 PCT/AU98/01049 - 19 by a Transient Vibration Method (ASTM Standard C1259-94) of a right prismatic beam in the flexural mode. Densities were determined by the direct measurement method. The results are presented in the following table. 5 Mechanical Properties Example ZrO 2 Young's Modulus FBD (wt%) (GPa) (g.cm - 3 ) 1 0 375 3.90 2 4 275 3.92 3 6 165 3.93 4 8 142 3.95 5 10 134 3.96 6 4 363 3.94 7 6 244 3.95 8 8 169 3.96 9 10 158 3.98 The results indicate an increase in fired bulk density with 10 increasing zirconia content. This is consistent with zirconia having a significantly higher theoretical density as compared to alumina. Young's modulus was then plotted as a function of zirconia content. 15 From the results in Fig. 3, it can be seen that: (i) there is a decrease in Young's Modulus with increasing zirconia content for both types of zirconia; and 20 (ii) the MEL S material was more effective in forming micro-cracks at lower levels of addition as compared to the AFM ZC03 material. The particle size of the two zirconias was determined (see 25 following table).

WO99/32417 PCT/AU98/01049 - 20 Particle Size of Zirconia Powders Material do (pm) ds 50 (im) d 90 (Im) %<5pm MEL S 10.1 19.1 35.1 2.1 AFM ZCO3 2.4 14.6 36.1 23.7 lot 011709) 5 From the results of the particle size analysis it can be seen that the average size of the MEL S is slightly larger. However, the greatest difference is in the shape of the particle size distribution. The MEL S material has a sharper distribution. This is clearly be seen by the 10 comparing the dlO0 values where the AFM material has a much higher concentration of fines. It is anticipated that isolated zirconia grains in the microstructure with a particle size less than 5 im will contribute little to micro-crack formation. For the AFM material this is almost 15 -25% of the material. The data from Fig. 3 was replotted considering only zirconia particles above 5Lm in size. This plot supports the hypothesis that the sub 5pm grains contribute little to 20 micro-crack toughening. From the results, it can be seen that the smaller fractions of zirconia contribute little to micro-crack toughening. Furthermore, the results indicate that for the particle 25 sizes of zirconia used, a critical zirconia density is required to initiate micro-cracks. Examination of etched polished surfaces of the Examples containing AFM ZrO 2 revealed that this phase was composed 30 of either single crystals or particles with only two or three grains. This indicates that it is possible to use a low cost micro-cracking agent.

WO99/32417 PCT/AU98/01049 - 21 Examples 10-11 The objective of Examples 10/11 was to investigate the performance of a micro-crack toughened refractory material 5 having a spinel matrix in accordance with the present invention produced by the method of the present invention from relatively low cost raw materials. A cup test was used to evaluate the material in contact 10 with both metal and slag. The raw materials used to manufacture the cups are given in the following table. 15 Raw Materials Component Supplier MgC03 Causmag A1 2 0 3 Alcoa (KA13) Fe 2

O

3 Aldrich (31,005,1 purity 99%+) ZrO 2 MEL Grade S Causmag is a refractory grade precursor. This material was crushed in a ring mill to produce a powder with a particle 20 size less than 75pm. KA13 alumina is also a refractory grade precursor. The price of KA13 alumina is roughly an order of magnitude less than A1000 alumina as used in Examples 1 to 9. The starting compositions are given in the following table. 25 Starting Composition in Parts Example MgO A1 2 0 3 10a 27.8 66.7 lla 27.2 61.7 Note: Mg source was added as MgC03 WO99/32417 - PCT/AU98/01049 - 22 The alumina and magnesite were ball milled (see table). Ball Milling Conditions Time 16 hours Powder 3 kg Balls 9 kg Mg PSZ Fluid 4.5 Idistilled water Binder No binder was added at this stage 5 The water was removed by pan drying at 80 0 C. The dried cake of magnesite and alumina were calcined at 1400 0 C to decompose any carbonates or hydroxides present (see table) and to form a spinel. 10 Calcination to Produce Spinel Heat 2000C.h

-

2 Dwell 1400 0 C for 1 hour Cool 2000C.h-1 or natural rate Iron oxide sintering aid was added to the pre-reacted 15 spinel. The overall compositions are given in the following table: Overall Compositions in Parts Example Batch Spinel Fe 2 0 3 Overall Composition 10b SPBOS 94.5 (SP01l) 5.5 Mg(All.Feo.

1

)O

4 llb SPB06 89.2 (SPO2) 10.8 Mg(All.SFe 0

.

2 )0 4 20 The batches were milled using the following conditions: WO99/32417 PCT/AU98/01I049 - 23 Ball Milling Conditions Time 4 hours Powder 3 kg Balls 9 kg Mg PSZ Fluid 4.5 t distilled water Binder 1 wt% PVA, 1 wt% glycerol and lwt% Dispex The zirconia was added to the slurry after milling and 5 immediately prior to spray drying. The zirconia addition to the slurry was at a level of 4 volume % (6.3 weight %). The slurry was continuously stirred prior to spray drying to minimise the effects of settling. 10 The overall starting compositions are given in the following table: Overall Compositions in Parts Example Batch Spinel wt% ZrO 2 wt% 10c SBZ01 93.7 (Mg(Al.

9 Feo.

1 )0 4 ) 6.3 llc SBZ02 93.7 (Mg(Al.aFe 0

.

2 )0 4 ) 6.3 15 Wet bag cold isostatic pressing (CIP) techniques were used to fabricate the cups using tooling consisting of a polyurethane bag and steel mandrel. Lids were fabricated by die pressing followed by bagging and CIP. A pressure of 20 210MPa was used for all runs. The components were fired using the following firing cycle.

WO99/32417 PCT/AU98/01049 - 24 Summary of Firing Cycle in Air Heat 100O0C.h Dwell 130 0 C for 30 minutes Heat 1000C.h -1 Dwell 750 0 C for 60 minutes Heat 100oC.h

-

2 Dwell 900 0 C for 60 minutes Heat 1000C.h " Dwell 1700 0 C for 240 minutes Cool 100C.h-luntil the natural cooling rate takes over. The nominal dimensions of the cups are given in the 5 following table. Nominal Dimensions of Cups Dimension Nominal Measurement Outside Diameter 55 mm Inside Diameter 30 mm Overall Height 55 mm Depth of Bore 35 mm Mass 500 g 10 An iron making slag was used for the test. The lime to silica ratio was in the range of 1 to 1.4 and the FeO content in the range of 0.5 to 10%. The pig iron was machined to be a snug fit into the cups. The slag was loaded into the crucibles and compacted to provide a dense 15 powder bed to protect the metal from oxidation during heat up before melting of the slag. The details of the combined slag-metal-cup test used are given in the following table.

WO99/32417 PCT/AU98/01049 - 25 Cup Test Details Slag Addition 20 g Metal Pig Iron nominal 4% carbon Metal Addition 40 g (close fit in bore slag on top) Temperature 1700oc Dwell at Temperature 4 hours Atmosphere Air (static) Cover Loose fitting lid The performance of these materials was excellent with very 5 low dimensional change observed after the test. The cups were essentially single phase materials with the iron incorporated into the crystal structure of the spinel. There was little slag penetration into the crucibles. Both metal and slag were detected after the test. 10 Examples 12 and 13 The objective of Examples 12/13 was to investigate the effect on performance of variations in composition of the 15 spinel matrix of micro-crack toughened refractory materials in accordance with the present invention. The raw materials used were as follows: 20 Raw Materials Component Supplier MgC03 Causmag A1 2 0 3 Alcoa (KA13) ZrO 2 Magnesium Electron MEL S The Causmag was crushed in a ring mill to produce an agglomerated powder less than 75pm in size. The starting 25 compositions are given in the following table. The final WO99/32417 PCT/AU98/01049 - 26 composition after the total process is given under the "Comments" column. Starting Composition in Parts 5 Example Composition MgC03 A1 2 0 3 Comments 12a SP95 46.7 53.3 Magnesia Rich 70 wt% Alumina 13a SP96 40.6 59.4 Alumina Rich 75 wt% Alumina The alumina and magnesite were mixed in a ball mill (see table). 10 Ball Milling Conditions Time 16 hours Powder 0.7 kg Balls 3 kg Mg PSZ Fluid 1 Idistilled water The water was removed by pan drying. The dried cake of magnesite and alumina were calcined at 1400 0 C to decompose 15 any carbonates or hydroxides present and to form a spinel using the conditions as disclosed in Examples 10 and 11. The spinel was crushed to produce a powder with a ds 50 less than 5pm. The milling conditions used were as follows: WO99/32417 PCT/AU98/01049 - 27 Ball Milling Conditions Batch 12b (SP95) 13b (SP96) Details Time 4 4 hours Powder 0.6 0.6 kg Balls 3 3 kg Mg PSZ Fluid 1 1 Idistilled water Binder Yes Yes 1.5wt% PVA, 1.2wt% glycerol and lwt% Dispex The zirconia was added to the slurry after milling and 5 immediately prior to spray drying. The zirconia addition to the slurry was at a level of 4 volume %. The overall starting compositions are given in the following table: Starting Compositions in Parts 10 Example Composition Spinel ZrO 2 12c SPZ95 93.7 6.3 13c SPZ96 93.6 6.4 The slurry was spray dried. During spray drying, the slurry was continuously stirred to minimise the effects of settling. Wet bag cold isostatic pressing (CIP) techniques 15 were used to fabricate the cups and lids as described in Examples 10 and 11 from the dried powder. The firing cycle used to densify the test cups and lids was the same as described for Examples 10 and 11. 20 After the cup test of Example 12c, examination of a cross section taken from the crucible revealed extensive damage. There was swelling of the cup as the result of the formation of internal porosity in the walls of the cup. The "tearing" of the microstructure is indicative of the 25 formation of the porosity occurring at high temperatures. The performance was assessed as poor.

WO 99/32417 PCT/AU98/01049 - 28 By stark comparison, after the cup test of 13c, examination of a cross section taken from the crucible showed little evidence of slag attack. There was evidence of slag penetration into the body without any sign of major 5 disruption. The performance was assessed as good. The results reveal a dramatic effect of stoichiometry on the chemical performance. The aluminium rich materials were superior to the magnesium rich spinels for the slag tested. 10 Example 14 The objective of the example was to investigate the thermal shock resistance of a micro-crack toughened refractory 15 material having a dispersed single crystal phase in a spinel matrix in accordance with the present invention. The raw materials used were as follows: 20 Raw Materials Component Supplier MgCO 3 Causmag (Milled) A1 2 0 3 Alcoa (KA13) ZrO 2 AFM ZCO3 The Causmag was used as supplied. The particle size of the as-received powder was less than 75pm in size. The starting 25 composition is given in the following table. The composition after calcination is also given in brackets. Composition of Example 14 (Parts) Composition MgC03 (MgO) A1 2 0 3 SP101 (Starting) 41.5 58.5 SP101 (Est.Final) (25) (75) WO 99/32417 PCT/AU98/01049 - 29 The alumina and magnesite were mixed in a ball mill (see table). Ball Milling Conditions 5 Time 16 hours Powder 5 kg Balls 15 kg Mg PSZ Fluid 7.5 1 distilled water The water was removed by pan drying. The dried cake of magnesite and alumina was calcined at 1400 0 C to decompose any carbonates or hydroxides present and to form a spinel. 10 The firing cycle used was the same as disclosed in Examples 10 and 11. The powder was milled using the following conditions: Ball Milling Conditions 15 Time 4 hours Powder 3 kg Balls 9 kg Mg PSZ Fluid 3 distilled water Binder 1 wt% PVA, 1 wt% glycerol and 1wt% Dispex The zirconia was added to the slurry after milling and immediately prior to spray drying. The zirconia addition to the slurry was at a level of 8 weight%. The slurry was 20 continuously stirred prior to spray drying to minimise the effects of settling. Wet bag cold isostatic pressing (CIP) techniques were used to fabricate the bars for thermal shock testing from the 25 dried powder. A pressure of 210 MPa was used for all runs. The components were fired using the firing cycle as outlined for Examples 10 and 11.

WO99/32417 PCT/AU98/01049 - 30 After firing and machining, the bars were subjected to a rapid heating using a gas torch passing across the surface of the samples. The flame was hot enough to cause localised melting on the surface. Although major cracks were formed, 5 the sample was in one piece at the conclusion of the test. By comparison, under this test, an alumina bar fabricated according to Example 1, resulted in the formation of shards roughly a 1 cm in size. This shows the excellent thermal shock resistance of the bodies fabricated from low cost 10 precursors according to the invention. Examples 15-17 The objective of Examples 15-17 was to investigate the 15 chemical stability of refractory materials having a spinel matrix in accordance with the present invention. The precursors used are listed in the following table. 20 Raw Materials Component Supplier MgC03 Ajax (Lab Grade) A1 2 0 3 Alcoa (A16SG) CaC0 3 Ajax (Unilab) SiO 2 5 micron Min-u-Sil The calcium and magnesium carbonates were calcined at 900 0 C to decompose the carbonate and hydroxides present before 25 use. The starting compositions are given in the following table: WO99/32417 PCT/AU98/01049 - 31 Starting Compositions in Parts Example Composition MgO A1 2 0 3 SiO 2 CaO 15 SPO4 28.3 71.7 - 16 SPGO3 28.1 71.0 1.0 17 SPD05 28.1 70.9 - 1.0 The calcined magnesite, alumina and silica precursors were 5 mixed in a ball mill (see table). Ball Milling Conditions Time 16 hours Powder 0.1 kg Milling Media 0.5 kg YTZP Milling Fluid 0.2 t iso propanol Binder None 10 The powder was pan dried at 80oc to remove the fluid. Bars of nominal fired dimensions 20 mm long and with a square cross section of 5 mm were fabricated by uniaxial pressing the powder in a steel die followed by cold isostatic pressing using wet bag techniques at a pressure of 210 MPa. 15 Samples were densified using the firing cycle as outlined in Examples 10 and 11. The fired bulk densities obtained after firing are given in the following table.

WO99/32417 PCT/AU98/01049 - 32 Fired Bulk Density After Firing Example Batch Firing Time FBD AP (%) Temp(oC) (h) (g.cm -3 ) 15 SP04(05)C89 1700 4 3.44 0.5 16 SPGO3(02) 1700 4 3.40 0.4 17 SPD05(01)C107 1700 4 3.42 0.1 Notes FBD Fired bulk density 5 AP Apparent porosity To determine the chemical stability, the materials were heated in contact with a slag. The test consisted of placing a sample in a crucible and surrounding with pre 10 mixed slag. The crucible was removed and the sample extracted from the slag at temperature. Details of the test are summarised in the following table. Pt Crucible Test 15 Crucible 15 ml Pt 5 9 Au Mass of Slag 6 g Mass of Sample 4 g Test Temperature 1550 0 C Duration 2 hours The slag was the same as used in Examples 10 and 11. After the slag test, the degree of slag penetration increased in the order of Example 17 > Example 16 > Example 15. 20 Example 18 The objective of the example was to investigate the firing temperature required to produce a refractory material 25 having a spinel matrix in accordance with the present WO 99/32417 PCT/AU98/01049 - 33 invention. The precursors used are listed in the following table. 5 Raw Materials Component Supplier MgC0 3 Causmag A1 2 0 3 Alcoa (A13) The magnesium carbonate was calcined at 900 0 C to decompose the carbonate and hydroxides present before use. The 10 starting compositions are given in the following table: Starting Composition in Parts Example Composition MgO A1 2 0 3 18 SPF20 28.4 71.7 15 A vibro milling technique was used for mixing and particle size reduction of the alumina. (see table). Vibro Milling Conditions Time 2 hours Powder 0.1 kg Milling Media 0.8 kg YTZP Milling Fluid 0.1 t iso propanol Binder None 20 The powder was pan dried at 80 0 C to remove the fluid. Discs with a nominal green diameter of 25 mm and mass of 10 g were fabricated. Samples were produced using uniaxial pressing in steel dies followed by cold isostatic pressing WO99/32417 PCT/AU98/01049 - 34 using wet bag techniques at a pressure of 210 MPa. The firing cycle as described in Examples 10 and 11 was used for the densification of the samples with the exception of the maximum temperature and dwell times. 5 The fired bulk densities obtained after firing for selected temperatures and times are given in the following table. Effect of Firing Temperature 10 on Densification for MA Spinels Example Batch Firing Time FBD Temp (h) (g.m3) Theoretical (oC) Density 18a SPF20(02)D 1600 1 3.18 89.1 11 18b SPF20(03)C 1650 1 3.35 93.8 110 18c SPF20(04)C 1700 1 3.39 95.0 111 18d SPF20(05)C 1750 1 3.41 95.5 112 18e SPF20(01)D 1700 4 3.42 95.8 10 Notes FBD Fired bulk density 15 The results indicate that the temperatures in excess of 1700 0 C are required to produce high density spinels products. 20 Example 19 The objective of the example was to investigate the density of micro-crack toughened refractory materials having a dispersed single crystal phase in a spinel matrix in 25 accordance with the present invention.

WO99/32417 PCT/AU98/01I049 - 35 The raw materials used were as follows: Raw Materials Component Supplier MgC03 Causmag (Milled) A1 2 0 3 Alcoa (KA13) ZrO 2 AFM ZCO3 5 The Causmag was used as supplied. The particle size of the as-received powder was less than 751m in size. The starting composition is given in the following table. The composition after calcination is also given in brackets. 10 Composition of Example 19 (Parts) Composition MgC03 (MgO) A1 2 0 3 SPF106 (Starting) 41.5 58.5 SPF106(Est. Final) (25) (75) The alumina and magnesite were mixed in a ball mill (see 15 table). Ball Milling Conditions Time 16 hours Powder 5 kg Balls 15 kg Mg PSZ Fluid 7.5 e distilled water 20 The water was removed by pan drying. The dried cake of magnesite and alumina was calcined at 1400 0 C to decompose any carbonates or hydroxides present and to form a spinel. The calcination cycle used was the same as used in Examples 10 and 11. After calcination, the powder was milled using WO99/32417 PCT/AU98/01049 - 36 the following conditions: Ball Milling Conditions Time 16 hours Powder 3 kg Balls 9 kg Mg PSZ Fluid 4.5 distilled water Binder 1 wt% PVA, 1 wt% glycerol and lwt% Dispex 5 The zirconia was added to the slurry after milling and immediately prior to spray drying. The zirconia addition to the slurry was at a level of 5.1 volume % (8.0 weight%). The slurry was continuously stirred prior to spray drying 10 to minimise the effects of settling. Wet bag cold isostatic pressing (CIP) techniques were used to fabricate samples for densification studies. A pressure of 210 MPa was used for all CIP runs. The components were 15 fired using the following firing cycle except for 19d which used the firing cycle as disclosed for Examples 10 and 11. Summary of Firing Cycle in Air Heat 1000C.h' Dwell Max temp for 60 minutes Cool 1000C.h-1 until the natural cooling rate takes over. 20 The densities obtained after firing at selected temperatures is given in the following table.

WO99/32417 PCT/AU98/01049 - 37 Effect of Firing Temperature on Densification Example Batch Firing Time FBD AP Temp (oC) (h) (g.cm 3 ) (%) 19a SPZ106(04)C185 1500 1 2.96 16.5 19b SPZ106(02)FLR 1600 1 3.40 5.5 19c SPZ106(01)C183 1700 4 3.50 2.0 Examples 20 and 21 5 The objective of Examples 20/21 was to investigate the effect of variations in composition of the spinel matrix and firing temperature on the density of micro-crack toughened refractory materials in accordance with the 10 present invention. The raw materials used are given in the following table. Raw Materials 15 Component Supplier MgC03 Causmag A1 2 0 3 Alcoa (KA13) Fe 2 03 Aldrich (31,005,1) ZrO2 MEL Grade S The Causmag was crushed in a ring mill to produce a powder with a particle size less than 75pm. The final composition after the total process is given under the Comments column. 20 Starting Composition in Parts SExample I MgO I A1 2 0 3 Fe 2 0 3 I Comments WO99/32417 PCT/AU98/01049 - 38 20a 27.8 66.7 5.5 Mg(Al.

9 Feo.

1 )O4 21a 27.2 61.7 10.8 Mg(A1 1 .eFeo.

2 )0 4 Note: Mg source was added as MgCO 3 The magnesite, alumina and ferric oxide were ball milled 5 (see table). Ball Milling Conditions Time 16 hours Powder 3 kg Balls 9 kg Mg PSZ Fluid 4.5 1 distilled water 10 The water was removed by pan drying at 80 0 C. The dried cake of magnesite, alumina and ferric oxide were calcined at 1400 0 C to decompose any carbonates or hydroxides present and form a spinel as described in Examples 10 and 11. The batches were milled using the following conditions: 15 Ball Milling Conditions Time 16 hours Powder 3 kg Balls 9 kg Mg PSZ Fluid 4.5 Idistilled water Binder 1 wt% PVA, 1 wt% glycerol and lwt% Dispex The zirconia was added to the slurry after milling and 20 immediately prior to spray drying. The zirconia addition to the slurry was at a level of 4 volume % (6.3 weight%). The slurry was continuously stirred prior to spray drying to minimise the effects of settling. The overall starting compositions are given in the following table: WO 99/32417 PCT/AU98/01049 - 39 Overall Compositions in Weight Percent% Example Batch Spinel wt% ZrO 2 wt% 20b SBZO3 93.7 (Mg(All.

9 Fe 0

.

1 )0 4 ) 6.3 21b SBZO4 93.7 (Mg(Al 1

.

8 Feo.

2 )0 4 ) 6.3 5 Discs were fabricated by die pressing followed by bagging and wet bag cold isostatic pressing at pressure of 210 MPa. The samples were fired using the following firing cycle except for 20f and 21f which used the firing cycle as disclosed for Examples 10 and 11. 10 Summary of Firing Cycle in Air Heat 1000C.h

-

2 Dwell Max temp for 60 minutes Cool 1000C.h - until the natural cooling rate takes over. The densities obtained after firing at selected 15 temperatures are given in the following table.

WO 99/32417 - PCT/AU98/01049 - 40 Effect of Firing Temperature on Densification Example Batch Firing Time FBD AP Temp (oC) (h) (g.cm -3 ) (M) 20c SBZ03(04)C185 1500 1 3.48 2.5 20d SBZ03(03)C184 1550 1 3.54 1.8 20e SBZ03(02)FLR 1600 1 3.56 2.0 20f SBZ03(01)C183 1700 4 3.55 1.8 21c SBZ04(04)C185 1500 1 3.62 1.9 21d SBZO4(03)C184 1550 1 3.60 2.0 21e SBZ04(02)FLR 1600 1 3.64 1.6 21f SBZ04(01)C185 1700 4 3.55 2.8 A comparison of the fired bulk densities after firing for 5 examples 18 to 21 is given in the following table. Effect of Firing Temperature on Bulk Density (g.cm -3 ) Example Composition Temp /Time oC h 1500/1 1550/1 1600/1 1700/4 18 Stoichiometric 3.18 3.39 19 A1 2 0 3 Rich 2.96 3.39 3.50 20 Fe Containing 3.48 3.54 3.56 3.55 21 Fe Containing 3.62 3.60 3.64 3.55 10 From these results, it can be seen that there is a significant advantage in the use of alumina rich spinels over stoichiometric compositions. In addition, the addition of iron results in a significant decrease in the sintering WO 99/32417 - PCT/AU98/01I049 - 41 temperatures. Examples 22-40 5 The objective of Examples 22 to 40 was to investigate the beneficial effect on densification of selected sintering aids for the spinel matrix of micro-crack toughened refractory materials in accordance with the present invention. 10 The raw materials used were as follows: Raw Materials Sintering Aid Component Code Supplier - MgC03 Causmag - A1 2 0 3 Alcoa (KA13) - ZrO 2 AFM ZC03 CaO CaC03 BDH AnalaR TiO 2 TiO 2 BDH MnO 2 MnO 2 BDH NiO NiO BDH CuO CuO BDH Cr 2 03 Cr 2 03 ALDRICH Fe 2 03 Fe 2 03 BDH COO CoO BDH ZnO ZnO AnalaR SrO SrC03 BDH 15 The Causmag was crushed in a ring mill to produce an agglomerated powder less than 75pnm in size. The starting compositions are given in the following table.

WO99/32417 PCT/AU98/01049 - 42 Starting Composition of the Spinel in Parts Composition MgC03 A1 2 0 3 Comments SPFl26 44.7 55.3 Stoichiometric spinel The alumina and magnesite were mixed in a ball mill (see 5 table). Ball Milling Conditions Time 36 hours Powder 5 kg Balls 15 kg Mg PSZ Fluid 7.5 Idistilled water 10 The water was removed by pan drying. The dried cake of magnesite and alumina were calcined at 1400 0 C to decompose any carbonates or hydroxides present and to form a spinel using the conditions as disclosed in Examples 10 and 11. The spinel was crushed to produce a powder with a ds 50 less 15 than 5pm. The milling conditions used were as follows: Ball Milling Conditions SPF126 Batch Details Time 32h hours Powder 5 kg Balls 15 kg Mg PSZ Fluid 5.5 distilled water Binder 1wt% PVA, 1wt% glycerol and lwt% Dispex 20 To examine the effect of the dopant concentration on the sintering behaviour for each dopant, batches were produced at a level of 5 mol% and 10 mol% with respect to the pre reacted spinel (SPF126). The following batches were WO99/32417 PCT/AU98/01049 - 43 produced. The exception was Example 23 that was made reaction sintering of the oxides. Starting Compositions 5 Example Batch 5mo1% 10mo1% 22 SPZ126 - 23 SQEZ02 CuO 24 SQFZ02 CuO 25 SQEZ03 NiO 26 SQFZO3 NiO 27 SQEZ04 Ti02 28 SQFZ04 TiO 2 29 SQGZ05 Fe 2 0 3 30 SQHZ05 Fe 2 03 31 SQGZ06 Cr 2 0 3 32 SQHZ06 Cr 2 03 33 SQEZ07 CaO 34 SQFZ07 CaO 35 SQEZ08 CoO 36 SQFZ08 CoO 37 SQEZ09 ZnO 38 SQFZ09 ZnO 39 SQEZ10 SrO 40 SQFZ10 SrO The zirconia powder was added just 15 minutes before the end of milling to prevent any decrease of zirconia particle size. The concentration of zirconia was 8 wt% of the total 10 batch.

WO99/32417 PCT/AU98/01049 - 44 Milling Conditions for Mixing Powders Type Ball milling Powder see compositions 0.3 kg Media Y-TZP balls 0.9 kg Fluid Isopropanol 750 ml Binder No binder Milling time without ZrO 2 16 h after adding ZrO 2 15 min The slurry including the media was poured from the milling 5 containers into glass containers. The liquid was removed by pan drying in vacuum at 70 oC for 20 h. To remove the milling media from the powder and for a better distribution of the zirconia particles in the powder sieving was applied. The powder was passed through two sieves with a 10 grid size of 4000 pm and 600 pm. Pellets were fabricated from the granulated powder batches. The fired pellets were a nominal 20 mm in diameter with a nominal mass of 10 g for densification studies and bars 15 nominally 20mm long by 5 mm by 5mm. The samples were uniaxially pressed followed by wet bag cold isostatic pressing at 210 MPa. The components were fired using the following firing cycle.

WO99/32417 PCT/AU98/01049 - 45 Summary of Firing Cycle in Air Heat 1000C.h-1 Dwell 130 0 C for 30 minutes Heat 1000C.h

-

1 Dwell 750 0 C for 60 minutes Heat 1000C.h-1 Dwell 900 0 C for 60 minutes Heat 100oC.h - 1 Dwell Maximum for 60 minutes Cool 1000C.h - until the natural cooling rate takes over. The maximum temperatures evaluated for the density studies 5 were 1400 0 C, 1500 0 C, 1600 0 C and 1700 0 C. Samples for the dip test were sintered at 1700 0 C. The results of the densification studies are given in the following tables. 10 Effect of Dopant on Fired Bulk Density (5mol%) Example/ 22 23 25 27 29 31 Firing Temp SPZ126 SQEZ02 SQEZO3 SQEZ04 SQGZ05 SQGZ06 oc C CuO NiO TiO 2 Fe 2 0 3 Cr 2 03 1400 2.24 1.89 2.26 3.48 2.45 2.28 1500 2.59 2.35 2.58 3.52 2.93 2.55 1600 3.43 3.34 3.48 3.50 3.51 3.43 1700 3.54 3.30 3.53 3.40 3.51 3.53 WO99/32417 PCT/AU98/01049 - 46 Effect of Dopant on Fired Bulk Density (5mol%) Continued Example/ 33 35 37 39 Firing Temp oC SQEZO7 SQEZ08 SQEZO9 SQEZ10 CaO CoO ZnO SrO 1400 2.30 2.25 2.25 2.28 1500 2.93 2.63 2.64 2.59 1600 3.40 3.53 3.33 2.62 1700 3.32 3.56 3.58 3.40 5 Effect of Dopant on Fired Bulk Density (10mol%) Example/ 22 24 26 28 30 Firing Temp oC SPZ126 SQFZ02 SQFZ03 SQFZ04 SQHZO5 CuO NiO TiO 2 Fe 2 03 1400 2.24 2.47 2.24 3.36 2.61 1500 2.59 2.93 2.57 3.52 3.12 1600 3.43 3.58 3.53 3.35 3.53 1700 3.54 3.47 3.60 3.26 3.52 Effect of Dopant on Fired Bulk Density (10mol%) Continued Example/ 32 34 36 38 40 Firing Temp oC SQHZ06 SQFZ07 SQFZ08 SQFZ09 SQFZ10 Cr 2 03 CaO CoO ZnO SrO 1400 2.31 2.34 2.24 2.29 2.18 1500 2.55 3.13 2.65 2.60 2.35 1600 3.40 3.17 3.55 3.29 2.66 1700 3.55 - 3.57 3.59 3.32 10 From the results it can be clearly seen the beneficial effects of CuO, TiO 2 , NiO, Fe 2 03, CoO on the densification of MCT spinels. It is important to note that the addition WO99/32417 PCT/AU98/01049 - 47 of Ti and Ca resulted in these phases being detected as discrete second phases by energy dispersive analysis (EDS) in conjunction with the scanning electron microscope (SEM). Typically the Ti had reacted with the finer zirconia 5 fractions inevitably present to form a new secondary phase. The performance of the different doped spinels was investigated using a slag dip test. The conditions were the same as described in Examples 15-17. 10 Dip Test Conditions Crucible Pt5%Au Mass of Slag 6g Mass of Sample 4g Soak Temperature 1500 0 C Soak Time 120 minutes The results after testing are given in the following table.

WO99/32417 PCT/AU98/01049 - 48 Observations after the Dip Test 23 26 28 SQEZO2 SQFZ03 SQFZ04 CuO NiO TiO 2 Good resistance. Good resistance. Complete Slag penetration Slag penetration infiltration of slag with slag limited to 50-200pum limited to 50-200pm composition inslag composition in No ZrO 2 in this No ZrO 2 in this zone centre of the zone. samples the same as original slag with additional Ti detected. 30 32 36 SQHZ05 SQHZ06 SQFZ08 Fe 2 0 3 Cr 2 0 3 CoO Intermediate with Good resistance. Good resistance. evidence of slag Slag penetration Slag penetration wetting all grain limited to 50-200pm limited to 50-2001um boundaries. Thickness of the GB No ZrO 2 in this zone No ZrO 2 in this zone less than in the case of Ti additives 5 From the results it can be seen for the slag tested that the performance of the Cu, Ni, Cr and Co were good but the performance of the Ti doped samples was inferior for the slag tested. As noted, the use of Ti as a sintering assist resulted in a second phase as detected by EDS technique. 10 However it is postulated that in different slags the Ti containing phase could exhibit much greater resistance to slag penetration. This is consistent with the selection of the sintering assist for the intended application. 15 Example 41 to 45 The object of Examples 41 to 45 was to illustrate the WO99/32417 PCT/AU98/01049 - 49 effects of addition of zirconia in the present material and to indicate the mechanism. The raw materials were treated prior to use. The alumina 5 powder was dried at 120 0 C for approximately 16 hours. The magnesite was calcined to remove any carbonates and hydroxides present. Raw Materials 10 Component Supplier MgCO 3 Ajax (Unilab Lab Reagent) -A120 3 Alcoa (A16SG) ZrO 2 MEL Grade S The calcination details are given in the following table. Summary of Calcination Firing Cycle in Air 15 Heat 2000C.h

-

2 Dwell 900 0 C for 3 hours Cool 2000C.h 1 until the natural cooling rate takes over. The dried alumina powder was combined with the magnesium oxide produced from the calcination process in proportions given in the following table.

WO 99/32417 - PCT/AU98/01049 - 50 Starting Compositions Example 41 42 43 44 45 SPZ11 SPZ12 SPZ13 SPZ14 SPZ15 MgO (g) 54.8 53.97 53.1 52.2 51.3 A1 2 0 3 (g) 138.7 136.5 134.3 132.0 129.9 Binder (g) 2.0/2.0 2.0/2.1 2.0/2.0 2.1/2.0 2.0/2.0 Iso propanol 400 400 400 400 400 (ml) Mill media 799.4 801.7 801.1 800.9 799.3 (g) Mill time 14.5 14.5 15 15 15 (hr) At the end of the milling period the mills were removed 5 from the rack, opened and the ZrO 2 added. The mills were then returned to the rack and given an additional 30 minutes of rotation, the objective being to thoroughly distribute the ZrO 2 rather than reduce the particle size. The quantities of the MEL Grade S ZrO 2 added are given in 10 the table. Zirconia Additions Example MEL S (g) 41 SPZ11 6.42 42 SPZ12 9.57 43 SPZ13 12.69 44 SPZ14 15.76 45 SPZ15 18.80 15 The slurries were separated from their respective milling and the slurries dried in a vacuum oven at 80 0 C and 200kPa. Segregation of the constituents was avoided by ensuring that the slurry viscosity remained high and by use of a WO99/32417 - PCT/AU98/01049 - 51 shallow drying pan. Drying time was ~24hours for all batches. For all batches, the dried powder cake was gently crushed 5 by use of a small number of YTZP balls in a coarse granulating sieve then through sieves of decreasing size with the final size being 500pm. The sieved batch weight losses were between 1.0 and 4.0% 10 and the milling media weight loss was zero. Four bars were pressed from all batches. These were formed by die pressing with a bar die of dimensions 7.5 x 101.9 mm and a die face pressure of 35MPa (25 bar hydraulic 15 pressure). The bars were then bagged and cold isostatically pressed to 210MPa. The firing cycle used to densify the samples is given in the following table. Summary of Firing Cycle in Air 20 Heat 1000C.h -1 Dwell 130 0 C for 30 minutes Heat 100oC.h

-

1 Dwell 750 0 C for 60 minutes Heat 100OC.h

-

1 Dwell 900 0 C for 60 minutes Heat 100oC.h Dwell 1700 0 C for 240 minutes Cool 1000C.h -1 until the natural cooling rate takes over All bars were diamond machined to nominal dimensions of 5.0 x 10.0mm cross section and approximately 80mm length. A second firing cycle was carried out on all test bars to 25 remove all traces of the hold down wax used in the machining process.

WO99/32417 - PCT/AU98/01049 - 52 A series of mechanical tests was carried to determine various physical properties. The Modulus of Elasticity in Tension (Young's Modulus, E) was determined by the analysis of a transient flexural vibration of a beam of the test 5 material. The Impulse Excitation of Vibration (ASTM C1259 95) technique used with the Grindo-SonicT

M

. The fracture toughness To, and the fracture energy determinations, y7 (Initiation Fracture Energy) and ywo (Work of Fracture) were determined on the same bars which had been notched to 10 roughly half the depth. These were then tested in 3 point SENB geometry in accord with ASTM E399-83. Strength a, was determined by a four point flexural strength test (MOR) on the tested SENB bar halves. The thermal shock damage resistance parameter R' and R'''' (from Hasselman), were 15 calculated. Ey,

-

2 20 The retained strength after thermal shock was also determined. Bars cut for the flexural test referred to above, were quenched from 700 0 C into boiling water. These were then tested for strength and normalised retained 25 strength, oR/a, computed.

WO 99/32417 PCT/AU98/01049 - 53 Example 41 42 Property SPZ11 SPZ12 Fired Density, p (gcm " 3 ) 3.42±0.01 3.47±0.00 Young's Mod. E (GPa) 238.8±2.0 242.2±0.8 Strength, MOR. y (MPa) 179±9.8 162.2±1.0 Fracture Toughness, K 1 MNm "3/2 2.38±0.32 2.58±0.11 Fracture Energy, Y (Jm "2 ) 31.7±4.1 31.9±2.1 Thermal Cond. K (Wm'K- 1 ) 15.0 15.0 Coeff. Thermal Exp. C (x10- 6

K-

1 ) 7.6 7.6 R' 1.48x10 3 1.32x10 3 R"" 2.36x10

"

' 2.94x10

-

' Retained Strength,*(, (MPa) 11.9±1.0 19.9±4.2 Normalised,** R/a - 7% 127 Example 43 44 45 Property SPZ13 SPZ14 SPZ15 Fired Density, p (gcm " 3 ) 3.49±0.00 3.50±0.02 3.50±0.00 Young's Mod. E (GPa) 239.4±0.1 74.6±1.1 46.4±0.2 Strength, MOR. a (MPa) 126.0±6.3 47.8±1.5 29.9±1.3 Fracture Toughness, MNm -3/2 2.58±0.15 1.82±0.04 1.41±0.03

K

0 Fracture Energy, (Jm- 2 ) 31.3±3.1 62.5±3.4 60.6±1.3 Ywof Thermal Cond. K (Wm'K " 1 ) 15.0 15.0 15.0 Coeff. Thermal Exp. (x10- 6 K ') 7.6 7.6 7.6 a R' 1.04x10 3 1.26x10 3 1.27x10 3 R"" 4.72x10

-

' 2.04x10 - 3 3.15x10 -3 Retained (MPa) 34.0±1.3 34.7±1.42 20.8±1.7 Strength,*aR Normalised,** aR/a - 27% 73% 70% * Denotes the strength measured by MOR after a quench from 700 0 C into water at 100 0 C. 5 ** " the retained strength (*) divided by the pristine strength.

WO99/32417 PCT/AU98/01049 - 54 The flexural strength before and after thermal shock and normalised retained strength after thermal shock testing as a function of zirconia content are shown in Figs. 4 and 5. 5 From these results the addition of zirconia as taught in the present invention increases the thermal shock resistance at the expense of the absolute strength of the body. 10 From the results it can be seen that in stark contrast to Claussen and Steeb (US Patent 4,298,385), products according to the present invention undergo a substantial decrease in the absolute strength and the fracture toughness decreases with increasing zirconia content. This 15 behaviour is the opposite as observed with the Claussen and Steeb materials that showed no significant loss in strength and an increase in fracture toughness. However, the thermal shock behaviour shows that the mechanism that operates in these materials enhances the thermal shock resistance as 20 demonstrated by increased retained strength after the thermal shock test.

Claims (22)

1. A dense refractory material which includes a spinel matrix and a micro-crack initiating phase dispersed 5 in the matrix.

2. The material defined in claim 1 wherein the micro-crack initiating phase is no more than 15% by volume of the material. 10

3. The material defined in claim 2 wherein the micro-crack initiating phase is no more than 10% by volume of the material. 15

4. The material defined in any one of the preceding claims wherein the spinel matrix is at least 80% by volume of the material.

5. The material defined in claim 4 wherein the 20 spinel matrix is at least 90% by volume of the material.

6. The material defined in any one of the preceding claims wherein the micro-crack initiating phase includes a dispersion of single crystals. 25

7. The material defined in any one of the preceding claims wherein the micro-crack initiating phase is formed from zirconia. 30

8. The material defined in claim 7 wherein the zirconia has a particle size in the range of 5 to 50p.m.

9. The material defined in claim 8 wherein the zirconia has a particle size in the range of 10 to 20pm. 35

10. The material defined in claim 7 wherein the zirconia is fused zirconia. WO99/32417 PCT/AU98/01049 - 56

11. A dense polycrystalline spinel refractory material with a dispersed monoclinic zirconia second phase that has an apparent lower thermal expansion coefficient as 5 compared to the matrix on cooling from the fabrication or operating temperature of the refractory to room temperature and results in the formation of a stable network of micro cracks that inhibits catastrophic failure as a result of the effects of thermal shock. 10

12. The material defined in claim 11 wherein the micro-crack network is characterised in that the micro cracks extend over several matrix grain diameters in the microstructure. 15

13. The material defined in claim 11 or claim 12 wherein the monoclinic zirconia particles are in the range of 5 to 50pm. 20

14. The material defined in any one of claims 11 to 13 wherein the monoclinic zirconia particles constitutes 3 to 15 volume % of the total volume of the material.

15. A method of manufacturing a dense refractory 25 material product which includes the steps of: (i) mixing precursor oxides for a spinel material; 30 (ii) calcining the mixture to form the spinel material; (iii)forming the spinel material into a green form of the product; and 35 (iv) firing the green form of the product to produce the final form of the product. WO99/32417 PCT/AU98/01049 - 57

16. The method defined in claim 15 further includes a step of mixing the spinel material produced in step (ii) with an additive, such as zirconia, selected to form a micro-crack 5 initiating phase dispersed in the fired product.

17. The method defined in claim 15 or claim 16 wherein step (ii) is carried out in a temperature range of 800 0 C to 1600 0 C. 10

18. The method defined in claim 17 wherein the temperature range is 1000 0 C to 1400 0 C.

19. The method defined in any one of claims 15 to 18 15 further includes a step of milling the spinel material produced in step (ii) to an average particle size less than 10 pm.

20. The method defined in claim 19, when dependent on 20 claim 16, includes adding the additive to the spinel material after the milling step.

21. The method defined in any one of claims 15 to 20 wherein step (iv) includes firing the green form in a 25 temperature range of 1000 0 C to 1800 0 C.

22. A dense refractory material which includes a matrix and a micro-crack initiating single crystal phase formed from fused zirconia dispersed in the matrix.