From the beginning of its industrial use, in the middle of the XXth century, titanium alloys have been finding application in areas in which a good relationship mechanical strength/density and the ability to keep high mechanical properties at high temperature are the most significant selection factors, and the relevance of the parameter cost is tiny, or even non considered-in the military, aerospace and aeronautic industries. However, after the end of the Cold War, the demand of these sectors of activity has abruptly decreased, and there are references to production decreases over 40% by the beginning of the nineties.

Nowadays, an effort of widening the domain of application of titanium alloys is on course, either through the introduction of new alloys, or by using new processing technologies, namely foundry, although the decrease of production costs has been revealed itself as a crucial factor. In this context, the use of traditional foundry techniques may become an efficient method for cost reduction, so contributing in a significant way for widening the present application domains.

The use of traditional foundry techniques presents, however, a lot of difficulties, in consequence of the characteristics of the metal itself, namely its high melting temperature, low fluidity at the pouring temperature and high reactivity with almost all the elements. In face of these imperatives, the resources needed for the production of titanium castings are different, and present more sophistication, than those needed for casting parts made of traditional alloys, what strongly contributes to the high cost of those castings. On the other hand, the technical difficulties and the equally high cost of post-casting machining operations, when necessary, limit the number of moulding processes liable to be used, so becoming necessary to use processes leading to"near net shape products", that means, precision casting processes Melting operation Nowadays, the production of titanium casting alloys includes multiple remelting before the final melting charge is reached, with the purpose of obtaining the desired chemical composition and its homogenisation. The final melt is performed in arc or induction skull melting furnaces (using cold wall copper crucibles), under controlled atmosphere.

Ceramic crucible induction melting might be a valuable alternative to the actual melting routes, not only for titanium alloys but for reactive alloys in general. Although its use is actually limited to research purposes at lab scale, there are some references to a few industrial applications, where the castings quality is not critical [1, 2].

The main limitation to the development of a ceramic crucible induction melting technique suitable for reactive alloys concerns mainly the refractory material to use on the crucibles production. Nevertheless, there is some research work on this field [1-4], and it is believed that once such problem is solved the production of titanium castings will grow significantly. The use of ceramic crucible furnaces presents several advantages over the traditional cold wall melting equipment: Lower equipment cost; . Reduced area of the melting plant; . Easier metal temperature control; Easier overheating procedure ; Suitable to be used with any titanium alloy; Lower cost of the melting stock; Improved safety of the melting operation.

Results obtained so far with the purpose to develop a ceramic crucible suitable for titanium alloys are not very encouraging. Reports usually refer poor thermodynamic stability, with negative consequences on the final castings, such as: Metal contamination with different elements, as a consequence of metal- crucible interaction; . Presence of gas porosities and non-metallic inclusions; Crucible destruction or deterioration during the melting operation.

On the first case, contaminants usually appear forming interstitial solid solutions with titanium, leading to poor mechanical and metallurgical properties, namely hardness increase and structure heterogeneity. On the second case, inclusions are a consequence of metal-crucible reaction or erosion of the crucible wall, and porosities are probably due to low solubility of gaseous elements/compounds formed during the same reaction.

On the third case, the causes of crucible destruction are usually related with poor thermal-shock resistance of the crucible material, but erosion phenomena, rnetal- crucible reaction or even dissolution of the crucible material may also be present.

The US Patent 4710481 refers the melting of commercially pure titanium in crucibles of calcium oxide (CaO) that, in spite of the high thermodynamic stability that they present, drove to oxygen incorporation in the cast metal. The use of calcium oxide crucibles is also referred by Sato [1,2], that refers an incorporation of 0.4% oxygen in weight for commercially pure Ti, besides porosity formation inside the castings, after solidification. The use of CaO is, however, delicate, given the high hygroscopicity it presents.

The US 5102450 and EP 0526159B1 Patents also refer the use of CaO crucibles in the melting of intermetallic alloys based on the y-TiAl, to which 2 to 8% of Nb and 2% of Cr was added, with the purpose of reducing the tendency for oxygen absorption by the metallic bath. The metallic charges were introduced in the CaO crucibles in a predefined sequence (in first place Al, Cr and Nb, were melted and only after the melting of the alloy so obtained the titanium was introduced, with the purpose of shortening the time of contact of the titanium with the crucible). The alloys so produced revealed oxygen absorption between 0.1 and 0.25%, depending on the chemical composition of the alloy.

However, the use of this process for the obtaining y-TiAl is not viable, since the decrease of the contamination of the alloy with oxygen is reached by significantly altering the chemical composition of the alloy itself.

Some researchers have been developing work in this domain, using other refractory types, namely Y203. The US Patent 4040845 respects the melting of the alloys of the Ti-Be system in crucibles constituted by a refractory material composed by about 60% of Y203, the remaining being a mixture of rare earth elements, namely Dy203, Yb203, Er203, Gd203, Ce02, La203, Sm203, Nd203 e Pr203, and still CaO, A1203 and TiO2. The use of such crucibles leads to an absorption of 0,52 wt% O by the molten metal. The process still reveals other drawbacks, like the high production cost of the crucible, due to the high cost of the refractory materials used on its production.

Y203, MgO and CaO stabilized ZrO2 were also evaluated as crucible materials for titanium alloys. It's use lead to high incorporation of oxygen (1,55 to 6%) [5,6] and zirconium (1,3%) [6] in the molten metal.

Besides those referred, other oxides have been tested in the manufacturing of crucibles for Ti alloys, namely MgO, Th02 and Dy203, for which oxygen absorptions between 0.5 and 3.4% in weight were registered [7] Its use in crucible production presents, however, problems of different nature: the radioactive nature of the Th02, the low thermal shock resistance of MgO and the high cost of Dy203.

The use of other materials, besides oxides, in crucibles manufacture was also object of study. The results obtained with several carbides (TiC, ZrC, WC), borides (TiB2, ZrB2, CrB2) and graphite generally lead to strong contamination of the metallic alloy in elements resulting from the metal-crucible interaction phenomena [8,9, 10].

The time of contact between metal and crucible is one of the main factors responsible for the contamination of the metallic alloys. The US5299619 Patent refers the development of a method to decrease that time of contact and, so, to reduce metal contamination. This invention proposes the use of two melting charges, molten in different equipments that react exothermically after entering in contact, so decreasing the total time of melting, and consequently the time of metal-crucible contact. When applied to the production of TiAl, the process consists on the melting of an aluminium charge in a conventional furnace, the transport of the liquid metal to the furnace where the mixture Ti-Al will be done, and the pouring, under air conditions, of the liquid aluminium in the resident crucible of this furnace. This process doesn't avoid the surface oxidation of the molten aluminium during the melting, transport and pouring operations.

A different approach for obtaining TiAl castings is presented in the W09832557A1 Patent: a graphite crucible is united to a ceramic mould obtained by successive coatings of an expanded polystyrene pattern, to obtain a sealed set- up, with a direct connection between the crucible's mouth and the mould's ingate. The mould-crucible set-up is then heated at high temperature, so producing the vaporization of the polystyrene and the complete drying of the mould. Later, the melting charge (a billet with the geometry of the interior of the crucible and with the desired final chemical composition) is introduced in the crucible through its bottom, and the set-up is placed with the crucible inverted in the upper position in the melting furnace. The melting of the charge is very fast, since the crucible was already pre-warmed, so reducing the incorporation of carbon in the metal. This process demands, however, the use of pre-melting, to obtain the billet with the desired final chemical composition.

Moulding operation Immediately after pouring, an interaction between the liquid metal and the mould material starts, which dissolves to a certain extent. The high solubility in liquid titanium that moulding materials usually present, and the turbulent flow of the liquid metal, lead to an uniform distribution of contaminants in the liquid metal, which concentration will dictate the properties of the produced alloy. When solidification starts, metal-mould interaction continues, but diffusion of elements in the solid metal is much slower and and limits to the outer layer of the part, so promoting a gradient of impurities between its inner and outer regions. In practice, titanium alloys castings usually present a superficial layer of oxides, besides a concentration gradient of residual elements decreasing from the surface to the interior, with particular emphasis for oxygen. The presence of that gradient results in a gradient of hardness, the two being almost coincident. This superficial layer, significantly harder than the interior of the casting is known as"alpha-case", and requests the"grinding"of the surface of the casting by "chemical grinding"processes, before its use.

Requests of moulding and crucible materials are quite similar, namely its thermodynamic stability facing the metallic alloy. Concerning the mould, the metal contacts not only the refractory grains, but also the binder used to produce the ceramic.

The US4787439 Patent describes a process for obtaining moulds by the lost wax moulding process, using contact and filling layers of different nature. The contact layers are produced starting from slurries consisting in Y203, Zr02, mixtures of Y203 and ZrO2 and mixtures of ZrO2 and Si02. For all the cases, the contact layers are made with a slurry of colloidal silica and alumina silicate, using a coating of higher granulometry alumina silicate. In the pouring of Ti6A14V, the moulds with contact layer of Y203 result in castings with"alpha- case"extension between 2.5 and 25 ; j. m, while the other moulds produce"alpha- case"extension values between 250 and 500 jum.

A similar process is described in the US5944088 Patent. Ti6A14V parts were obtained starting from moulds with an interior layer fully produced in Y203 (the slurry consisting in colloidal Y203 and granular Y203, and the coatings made with larger granulometry Y203), the outer layers to be applied starting from slurries of colloidal silica and alumina silicate, with coatings of alumina silicate of larger granulometry. With this process it is possible to obtain Ti6A14V castings with"alpha-case"extensions between 25 arid 50 ßm.

The US5407001 Patent also refers the use of moulds of variable composition to obtain Ti6A14V castings. However, according to this patent, the contact layers are produced starting from slurries consisting in mixtures of Y203 and Zr02, in proportions varying between 100 and 75% of Y203, colloidal Si02 being used as binder. After the immersion of the pattern in the slurry, the grapes are covered with granular alumina. The outer layers are obtained starting from a slurry consisting of alumina using ethyl silicate as binder. The coatings after immersion in the slurry are made with alumina of larger granulometry. With this process it is possible to obtain Ti6A14V castings with"alpha-case"extensions between 25 and 200 um, according to the thickness of the part.

There are references to moulds integrally produced using Zr02 (either as binder, under the colloidal form, or as refractory) [11], in the production of commercially pure titanium castings, for which"alpha-case" extensions between 25 and 75 pm were found.

Summary of the invention The main subject of this invention is the development of a casting technique for yTiAl alloys, suitable to obtain contamination free castings, with characteristics that allow its industrial use.

In that context, the invention suggests the melting of yTiAl in low cost ceramic crucibles, suitable to achieve the rnain task of the invention. The developed ceramic crucibles reveal high thermal-shock resistance and are thermodynamically stable facing yTiAl alloys. When compared with CaO crucibles used up to now at laboratory scale, the developed crucibles are not hygroscopic, which makes easier their manipulation.

In order to reduce significantly the present production cost of yTiAl, this invention suggests the use of melting stocks consisting in commercially pure titanium and aluminium, in order to avoid pre-melting operations needed to obtain sub-products (billets) with the desired final chemical composition.

Another aspect of the invention concerns the moulding process and materials used to obtain the desired moulds. The invention anticipates the production of multi-layer moulds using the investment casting process, using refractories chemically stable facing yTiAl, without significant changes on the methodology, procedures and equipment of that moulding process. The moulding technique allows the production of very low surface contaminated castings with good surface finishing, similar to that of ferrous castings obtained by the traditional investment casting process.

Presently, the production of y-TiAl parts by foundry is not usual.

Such parts are usually produced by mechanical forming processes, from billets with a composition previously adjusted through a complex sequence of remeltings. Titanium and traditional titanium alloys castings are produced by the traditional investment casting process, but they usually present many casting problems, like porosities, chemical composition heterogeneity, and hardness variation. Moreover, the actual production costs are very high, as consequence of high investment costs in equipment and the high production costs themselves (melting operation, finishing costs, raw rnaterials).

The developed technique allows the elimination, or significant decrease, of those problems, namely the absence of casting porosities and very low contamination of the base metal in elements known to be harmful to the performance of any part (oxygen for example). The"alpha-case"extension is lower than 25 um, nevertheless the hardness value is still in agreement with the existent standards for y-TiAl, and therefore being not needed any later operation of removal of the surface metal. The developed technique relies upon low cost equipments, so decreasing the global production costs, and allows its use by any foundry that can use the traditional lost wax moulding process, without need of great modifications of its production processes.

Brief description of the drawings The following description is based on the enclosed drawings, which, without any limiting character represent: Figure 1-Schematic representation of the centrifugation structure of the melting furnace used on this work ; Figure 2-Acceleration curves used in pouring ; Figure 3-Thermal sinterisation cycle of the ceramic shells; Figure 4-Variation of the micro hardness values in the constituents a2 + y e from the surface to the interior of the samples; Figure 5-Graphic representation of the oxygen concentration variation, versus the distance to the surface of the samples; Figure 6-Portion of the Surface roughness profile of a casting sample.

Detailed description of the invention This invention concerns the development of a casting process to obtain y-TiAl parts.

Comparatively to the existent inventions in this domain, this is the only invention applied to y-TiAl, since this material is in development phase for industrial application, and, to the present da-te, its processing has just been done with the purpose of producing billets, for later transformation by plastic formation. The known approaches to the production of this material by melting in ceramic crucible respects to the US5102450 and EP 0526159B1 patents, but the material produced is not y-TiAl, but a quaternary Ti-Al-Nb-Cr alloy, which purpose is to decrease the affinity of the alloy to oxygen.

On the other hand, melting is performed starting from a charge of pure aluminium and titanium, technique until now not used to produce y-TiAl.

Using this process, a casting is obtained by only one single melting operation, in the same way of the traditional casting of metallic alloys, so avoiding the multiple previous remeltings needed to correct the chemical composition.

As it can be observed in figure 1, the furnace is constituted by a hermetic chamber (1), used both for melting and pouring, having a glass window (2) for introduction of a thermocouple for temperature reading and visualization of the course of the melting operation. In the lower part of the chamber, where melting is performed, there is a quartz crucible (7) that, during melting, is involved by an induction coil (10), that guarantees the heating of the melting charge. Inside the chamber, and inside the quartz crucible, a set consisting of an inner crucible (fusion) (6) and an outer crucible with a spout, for protection of the quartz and pouring crucibles (5), is placed. Between the outer crucible and the quartz crucible a ceramic support is placed (9), to avoid the fracture of the quartz crucible because of thermal shock. The mould (3) is placed in horizontal position, inside the same chamber, hold by an appropriate support (4), linked to the spout of the pouring crucible by a ceramic positioner (8). The chamber is tied up to a metallic shaft structure, which is fixed to the top of a vertical axis by a screwing device (12), around which the whole apparatus rotates. In the same metallic structure a counterbalance exists (13), at the opposite side of the chamber, relatively to the rotation axis, to gllarantee the balance of the apparatus.

Pouring is done centrifugally, by the rotation of the whole set described, around the referred axis.

Crucibles used in the melting operations are commercially available crucibles, made of ZrO2 stabilized with 9% of Y203, produced by the company Marketech International, U. S. A. , with porosity from 10 to 30%, to which interior a coating of Y203 is applied. The Y203 coating is applied as an emulsion with the characteristics presented in the table 1, produced by the company ZYP Coatings, U. S. A.

Table 1-Characteristics of the Yttrium oxi de (Y203) suspension used to inside coating the multi-layered crucibles.

Solid component Y203 (99. 9%) Liquid component Ethyl alcohol wt % solid component 54 wt % liquid component 46 Viscosity, R 120 cps pH 7 Composition after drying Y203-99.7 wt% C-0.3 wt% The coating operation of the crucibles is done by filling them with the Y203 suspension, followed by 30 seconds of permanence inside it, followed by dripping out for 5 minutes. With this procedure a uniform thickness coating is obtained, about 100 pm thick. After this operation, crucibles are dried at 200°C for 24 hours, and the mentioned operations repeated once again, except for the time and drying temperature that will be 24 hours a-t 500°C. With this procedure an interior coating of the crucible is obtained with a film of Y203, with a uniform thickness of about 200 Fm.

The mould is obtained by the lost wax moulding process, using successive coatings, of variable composition. The purpose of the use of different coating compositions is to decrease the production costs, since the material used in the outer layers is of significantly lower cost tha. n the one used in the inside layer. The moulding process begins with the construction of one, or several, wax patterns and their filling runners and ingates (grape), using commercially available wax for lost wax moulding. The different elements of the grape are obtained by gravity pouring the wax in metallic counter-moulds built for that purpose. The pouring temperature of the wax is 11 5°C, and the temperature of the counter-moulds is room temperature. The stabilization of the metallic counter-moulds temperature is attained by their immersion in cold water, immediately after each pouring. As releasing agent, silicon oil is used, applied by aspersion. The wax grape is later immersed in a solution of ethyl alcohol (50%) and acetone (50%), with the purpose of degreasing the surface and improving the adhesion characteristics of the first ceramic coating.

The grape so produced is then covered with a ceramic slurry with the characteristics shown in table 2. In tables 3 and 4 are shown the characteristics of the binder and of the refractory used in the slurry preparation.

Table 2-Composition of the slurry used in the first coating (type B1) Constituent Type Nature Amount (wt %) Binder Colloidal Y203 Y263 60 Refractory Y203 Y2O3 20 Urea--10 Acetic acid 99%--10 Anti-foam RSD-10 burst (*)-0. 25% on the binder Wetability promoter Victawet (*)-0. 25% on the binder (*) Produced by the company REMET Corporation, U. S. A.

Table 3-Composition of the binder used in the preparation of the slurry used in the first coating Water (wt %) 77 Y2O3 (wt %) 14 Acetic acid (wt %) 9 pH 7 Viscosity, µ (at 25°C) (cps) 10 Volumic Mass, p (kg dom-3) 1.17 Granulometry of the solid comp. (nm) 5-10 (*) Produced by the company NYACOL, U. S.A.

Table 4-Characteristics of the refractory used in the preparation of the slurry used in the first coating (*).

Composition wt % Si02 0. 1 Y203 99. 9 Granulometric index 325 mesh (< 45 urn) Volumic Mass, p (kg dm-3) 5. 0 (*) Produced by the company Micronmetals Incorporated, U. S. A.

The slurry is prepared using traditional mixing techniques, using a stem mixer of variable rotation speed. In the actual case of this slurry, preparation begins with the addition of urea to the binder (with the wetability promoter and the anti-foam already included), and later addition of the refractory.

This addition is done along 120 minutes, at a rate of 25% every 30 minutes.

Finally the acetic acid is added, in bulk, but very slowly (in about 5 minutes).

The goal of using urea is to decrease the tendency for-the slurry to jellify, so guaranteeing that that phenomenon is not verified for values of pH lower than 8.5. The use of acetic acid is needed to significantly lower the initial pH of the slurry to values that guarantee its acid nature, since the basic nature of the Y203 refractory would cause its jellification. The pH of the slurry is going to be adjusted along its time life, with small additions of acetic acid.

After coated with the slurry, the grape is dried for 24 hours under controlled atmosphere (temperature and moisture).

After this first coating, the grape is successively covered by immersion in another slurry, immediately followed by coating with a granular refractory material, being followed by new drying under controlled atmosphere.

The grain size of the refractory used is increased from coating to coating, in order to avoid sudden variations of the morphology of the coatings used in each shell, to avoid possible detachments between successive coating layers, during the drying or during the thermal processing of the mould. This series of operations is repeated as many times as considered necessary to give the shell the mechanical strength needed to its handling.

The nature of the outer coatings is different from that of the inner coating (Y203), in order to reduce costs. In table 5 the composition of the slurry used is shown, and in the tables 6 and 7 are shown the characteristics of the binder and the refractory used in its preparation.

Table 5-Composition of the slurry used in the outer coatings (type B2) Constituent Type Nature Amount (wt %) Binder Colloidal ZrO2 ZrO2 33 Refractory Zur02 ZrO2 66 Acetic acid 99%--1% on the total charge Anti-foam RSD burst-10 (*)-0. 25% on the binder Wetability promoter Victawet (*) -0. 25% on the binder (*) Produced by the company REMET Corporation, U. S. A.

Table 6-Composition of the binder used in the preparation of the slurry used in the first coating (*).

Water (wt %) 77 Zr02 (wt %) 20 Nitric acid (wt %) 3 pH 6 Viscosity, u (at 25°C) (cps) 20 Volumic mass, p (kg dm-3) 1.28 Granulometry of solid comp. (nm) 50 (*) Produced by the company NYACOL, U. S. A.

Table 7-Characteristics of the refractory used in the preparation of the slurry used in the first coating (*).

Composition wt % Zr02 > 94.7 Si02 < 0. 5 CaO 4.0 A1203 < 0. 4 Fe203 < 0. 1 Ti02 < 0.3 Y203 Granulometric Index 325 mesh (< 45 um) Volumic Mass, p (kg dns3) 5.8 (*) Produced by the company REMET Corporation, U. S. A.

The preparation of this slurry is done using the following procedure : Addition to the binder of the anti-foam agent followed by the wetability promoter agent; Slow addition of the refractory to the mixture binder + addictives, that is: > Addition of 50% of the refractory material in-the first 15 minutes of mixing ; > Addition of 25% more of the refractory material along 5 hours, at the rate of 5% per hour; Addition of the remaining 25% of the refractory material along the next 40 hours, at the rate of 5% each 8 hours.

Table 8 presents the characteristics of the granular refractory material used in the outer coating operations after coating with the slurry.

Table 8-Characteristics of the refractories used in the outer coatings of the ceramic moulds Granulometry Type Nature Nominal Dispersion (wt %) (mesh) Aperture (mm) > 0. 6 0.3-0. 6 0.15-0. 3 0.075-0. 15 < 0.075 RI Zr02 30-50 < 1 94 < 6-- R2 Zr02 50-100-3 93 4- R3 Zur02 100-325--< 1 70 <30 After manufacturing, the ceramic shell is submitted to a thermal processing at high temperature, for elimination of the wax and coalescence of the refractory grains, so driving to an increase in its mechanical strength.

After this operation and the cooling of the shell, the grape is introduced in a metallic box, the remaining space inside the box being filled with sand agglomerated by the sodium silicate/CO2 process. This way, the mould for the pouring operation is built.

The melting operation is done using an electric induction furnace, and the pouring is accomplished by centrifugation. The melting crucible is introduced inside an alumina crucible with a spout, for protection of the induction coil in case of fracture of the melting crucible. The spout of the crucible has a direct connection with an existent hole in the mould for its filling.

The whole process occurs inside a chamber under controlled atmosphere (pressure and nature) (figure 1). However, before proceeding to the melting and pouring to the mould, the set crucible + mould + melting charge is subjected to a heating stage at 300°C for 2 hours with the purpose of eliminating any moisture that may be present.

The described process allows the obtaining of y-TiAl castings of thickness up to 20mm, in which the level of contaminant elements doesn't surpass 0.11 at% of Y and 0.66 at% of O (0. 29 wt%), free from inside porosities, without need of removal of the"alpha-case", without formation of surface oxides films and with surface finishing identical to the usually presented by the ferrous alloys castings obtained by the lost wax moulding process.

In sequence, an example of sample production using the described process is presented.

Example Cylindrical samples with 20 mm diameter and 85 mm length were produced in y-TiAl. Melting crucibles were multi-layered ZrO2 + Y203 ones, prepared according to the methodology described, and with the dimensions: /4ext = 54 mm, i,, t = 40 mm, length = 80 mm. Melting charges were made of commercially pure Ti and Al, with the composition indicated in table 9.100 g charges were used, with 52% Ti and 48% Al.

Table 9-Composition of the commercially pure titanium and aluminium used in this example.

Chemical composition (wt %) O H N C Fe Ti Si Cu Al Ti CP 0.25 0.012 0.03 0. 08 0.2 > 99. 5 Al CP 0. 1 0. 05 0.01 > 99.8 After pre-heated up to 300°C, the set crucibles-mould-melting charge was placed inside the melting furnace chamber, in the position schematically shown in figure 2. After closing, the chamber was evacuated up to the pressure of 5 x 10-6 bar, and after, a washing operation was done with high purity argon (N60 Argon-02 < 10 ppm, N2 < 5 ppm, H20 < 1 ppm), with a 10-1 m3/minute caudal under the pressure of 2.5 bar. This"washing"operation was performed three times, and the argon pressure was then stabilized at the level of 10-3 bar, and then the effective melting process was initiated. The beginning of the melting operation started very slowly, to avoid the occurrence of thermal shock on the crucibles. Thus, the heating began using 1 kW power, kept for 10 minutes, time enough to allow the aluminium to change to the mushy state. After this phase, heating was performed under a 2 kW power, for 5 minutes, followed by a heating stage under 3 kW power, for 4 minutes, the beginning of the melting of the alloy being verified by the end of it. After this phase, the molten alloy was maintained under the same power for 60 seconds, the pouring being executed afterwards, around 1550°C, and the samples were allowed to cool inside the chamber until room temperature, maintaining the argon atmosphere. Pouring was performed by centrifugation, at a rotation speed of 418 rpm, with an acceleration in agreement with the curve shown in figure 2.

Moulds were produced according to the described techniques, the first coating was made with an Y203 slurry, without any refractory coating, and the following coatings used a Zur02 slurry and the ZrO2 granulated refractories of the types R1, R2 and R3. The procedure used is described bellow in detail: 1. Manufacture of the wax grape, according to the described methodology; 2. Immersion of the wax grape in the Y203 slurry (type B1) for 5 seconds; 3. Dripping out of the slurry for 10 seconds; 4. Drying of the mould for 24 hours, at a temperature of 30 °C zt2 and a relative moisture of 40-50%; 5. Immersion of the wax grape in the Zr02 slurry (type B2) for 5 seconds; 6. Dripping out of the slurry for 10 seconds; 7. Coating of the slurry to saturation with refractory of the type R3. This operation was done allowing the refractory to drop by gravity over the shell, through a 50 mesh sieve; 8. Drying of the mould for 24 hours, at a temperature of 30 °C 2 and a relative moisture of 40-50%; 9. Repetition of the steps 5 to 8; 10.2 repetitions of the steps 5 to 8, using a refractory of the type R2 and a 30 mesh sieve in the step 7; 11. 2 repetitions of the steps 5 to 8, using a refractory of the type R1 and a 10 mesh sieve in the step 7.

Using this procedure, ceramic shells with 7 coatings were obtained, the first one made exclusively with Y203 and the following ones with Zr02, with growing granulometry from the inner to the outer coating.

In the case of the Y203 slurry (type Bl) the pH value immediately after mixing was 5.2 but it increased gradually up to 7.8 by the end of 3 days. A new addition of acetic acid was made, lowering the pH value to 6.5 that raised again up to 7. 5during the 2 following days. During the lifetime of the slurry (10 days at the temperature of 25°C 2), its viscosity value varied significantly, accompanying the variations of the pH value. In the case of the Zr02 slurry (type B2), the pH was adjusted for 3.1. A tendency for continuous pH increase was verified with a viscosity decrease, which has been controlled with small daily additions of acetic acid. In table 10 the evolution of the slurry characteristics along the days is presented.

Table 10-Variation in time of the slurry characteristics Parameter After mixing During the life time of the slurry B1 B2 B1 B2 PH 5.2 3.1 5.2-7. 8 3.0-3. 3 Viscosity (s) 12 12 12-16 11-13 Solid content (%) 28 29 28-31 29-30 The viscosity values were ascertained through the measure of the flow time of the slurries through the hole of a Zahn #4 cup.

After the production of the shells a thermal shock operation to eliminate the wax was done. The shells were placed in an adequate support on the floor of a dedicated furnace, which was later properly lifted to the interior of that furnace, previously heated to the temperature of 900°C. After elimination of the wax, the shells were left to cool until room temperature. The final thermal processing of the shells was done according to the thermal cycle represented in figure 3.

Samples obtained by this process present a microstructure consisting in a lamellar a2+y microconstituent and a y phase. In table 11, chemical composition of the microconstituents versus its distance to the surface of the sample is shown. The alloy revealed a small contamination with Yttrium (0,11 at. %), being lightly superior at the metal-mould interface (0, 15 at. %).

Microhardness values are uniform along the whole sample, presenting a slightly higher value (about 3% higher) at the metal-mould interface, in the ot2+y constituent (figure 4). However, the registered increase is a characteristic of all casting parts, as a consequence of different cooling speeds on the periphery and the interior of the castings. In this example, it cannot be considered that an"alpha-case"equivalent to the thickness of that harder zone exists, since the increase of hardness doesn't force any type of surface finishing and is located between the values suggested by the bibliography for the y-TiAl alloys with the composition of the alloy produced in this work.

The oxygen concentration, determined by SIMS, is also uniform, a slight increase being registered for distances to the surface of the sample less than 20 um (Figure 5). This increase, is however non significant, since its influence in the microhardness value in the same zone is negligible.

Table 11-Variation of the chemical composition of the microconstituents versus the distance to the surface of the sample, in a sample of Ti-48Al, obtained by melting it in a ZrO2 crucible inside coated with Y203 and pouring in multi- layered Zr02 + Y203 ceramic moulds.

Distance to sample Inside surface average value (urn) (5000) 5 10 25 50 100 200 300 400 500 Chemical composition (% Ti 54.23 53.89 54.77 54.25 55.28 54.04 55.49 53.57 54.16 53.36 α2 + γ Al 45.62 45.98 45.15 45.66 44.62 45.06 44.40 46.37 45.79 46.53 Y 0. 15 0.13 <0. 10 <0. 10 0.10 0.10 0. 11 <0. 10 <0. 10 0. 11 Zr Under the equipment detection limit (0, 1 %) Ti 45.51 43.96 45.13 46.13 44.48 44.67 44.57 fly Al 54. 45 55.95 54.77 53. 90 55.45 55.26 55. 33 <0. 10 <0. 10 0.0 <0. 10 <0. 10 <0. 10 0.10 Zr Under the equipment detection limit (0,1 %) Average composition of Ti-52.66 A1-47. 23 Y-0.11 the sample The surface roughness value (Ra) obtained for the samples was 1. 6 µm (figure 6) and the surface finishing is in agreement with the 3/OS1 pattern of the 359-01 Technical Recommendation of BNIF. The samples haven't developed any superficial film of oxides, being only visible by electronic microscopy small Y203 portions adherent to the surface, with a thickness non greater than 2 um, easily removable using the cleaning processes conventionally in foundry.

The characteristics of the castings obtained by this process allow its direct industrial use, except in the cases in which the geometrical complexity demands the use of local machining operations. The main markets for this type of castings are the aeronautics and automobile industries, in which the relationship mechanical strength/density is important, and the properties at high temperature are a prioritary requirement. However, the developed process allows a significant reduction of the production costs, since it demands neither the use of complex equipment nor the use of high cost handwork with specific"know how", and doesn't force the accomplishment of surface finishing operations as a consequence of the absence of the"alpha-case." Thus, it may be anticipated that new markets will appear, whenever the application of this material is justified for technical reasons, that until now have been made unfeasible because of the high costs of the option. The present markets also should suffer a significant increase, based in new applications until now avoided for the reasons.

An important consequence of the development of this casting process is the possibility of its use by traditional lost wax foundry companies, since the modifications of the present techniques and methodologies are insignificant. This fact allows the opening of new markets, decreasing the risk situations in case of recession of some of the present markets.

References [1] SATO, T.; YONEDA, Y.; MATSUMOTO, N.-"A Technique for Casting Titanium Alloys with Lime Refractory: Communication presented to the 58th World Foundry Congress". Cracow, September 1991.

[2] SATO, T.; YONEDA, Y.; MATSUMOTO, N. -A New Process of Producing Titanium Alloy Castings. Transactions of the Japan Foundrvmen's Society. Vol.

11, (October 1992), p. 27-33.

[3] REETZ, T. -Keramische Werkstoffe fur Aluminium und Titanschmelzen.

Giesserei. Vol. 83, n° 16 (August 1996), p. 53-56.

[4] SCHADLICH-STUBENRAUCH, J.; SAHM, P. R. -Entwicklung einer Schleuderfeingieptechnik fur die Herstellung kleiner, dunnwandiger und filigraner Gupteile aus Titan und Titanlegierungen. Giessereiforschung. Vol. 43, n°4 (1991), p. 141-161.

[5] CHAPIN, E. J.; FRISKE, W. -A Metallurgical Evaluation of Refractory Compounds for Containing Molten Titanium: Part-Oxides. Washington, DC: Naval Research Laboratory, 1954. NRL 4447 Report.

[6] SAHA et al-On the Evaluation of Stability of Rare Earth Oxides as Face Coats for Investment Casting of Titanium. Metallurgical Transactions B. Vol. 21B, (June 1990), p. 559-566.

[7] REETZ, T. -Keramische Werkstoffe fur Aluminium und Titanschmelzen.

Giesserei. Vol. 83, n° 16 (August 1996), p. 53-56.

[8] CHAPIN, E. J.; FRISKE, W. -A Metallurgical Evaluation of Refractory Compounds for Containing Molten Titanium: Part I-Carbon, Graphite and Carbides. Washington, DC: Naval Research Laboratory, 1954. NRL 4467 Report.

[9] CHAPIN, E. J.; FRISKE, W. -A Metallurgical Evaluation of Refractory Compounds for Containing Molten Titanium: Part II-Borides and Sulphides. Washington, DC: Naval Research Laboratory, 1955. NRL 4478 Report.

[10] GARFINKLE, M. ; DAVIS, H. M. -Reaction of Liquid Titanium with Some Refractory Compounds. ASM Transactions. Vol. 58, (1965), p. 521-531.

[11] FRUEH, C.; POIRIER, D. R.; MAGUIRE, M. C.; HARDING, R. A.-Attempts to develop a ceramic mould for titanium casting-a review. International Journal of Cast Metals Research. Vol. 9, n° 4 (1996), p. 233-240.