CA2290708C - Cleaning and finishing a ceramic mold - Google Patents

Abstract

A technique for removing loose powder from the interior surfaces of a ceramic mold made using layer manufacturing processes, such as three-dimensional printing processes. The interior of a mold can be filled with water which is boiled and the particles are entrained in the flow caused by the boiling liquid. Alternatively, fine particles are introduced into the mold and agitated in the mold to dislodge the particles and the powder so that they can be poured out of the mold. Such technique tends also to remove the surface finish of the mold. Any of the particles which remain can be dissolved in a liquid and removed in the liquid from the mold. Further, the surface finish of the mold can be improved by casting a slip of fine particles on to the surface to form a generally level, and preferably non-conformal, coating on the surfaces.

Description

CLEANING AND FINISHING A CERAMIC MOLD
Introduction The invention is useful in connection with the fabrication of molds and other bodies that are made by layer manufacturing techniques.
This invention relates generally to ceramic molds for metal castings and, more particularly, relates to ceramic molds made by three dimensional printing techniques using a layered process and to techniques for the removal of loose l0 powder from within the ceramic mold and the improvement of the surface condition and especially the surface finish of the interior of the mold.
Background of the Invention Ceramic molds for metal casting can be created directly from a computer model using layer techniques, i.e., three dimensional printing processes which can be defined as processes that construct objects in layers using a computer model of the objects. Exemplary processes of this type are described, for example, in European Patent Application 20 Publication No. 0,431,924 A2 published on June 12th, 1991 (see also Canadian Patent 2,031,562). As described therein_ a ceramic mold for metal casting can be created directly from a computer model using such process wherein the mold is created on a layer by layer basis. As shown specifically in FIGS.
1(A)-1(F) herein, the mold may be created by spreading powder 1 using roller 3 within a confined region as defined by a piston 5 and cylinder 7 arrangement. A further material 9, 1a e.g., a binder material, is then deposited at specific regions 11 of a layer as determined by a computer model of the mold.
The further material acts to bind the powder within the layer and between layers. This process is repeated layer after layer until all layers needed to define the mold have been printed. The process results in a bed of powder 13 which contains within it a ceramic mold 15. When the further -2_ material which a used to b=nd the powder cor_tains a ceramic, r i re r~ v rori n 1 p., ~ ~,j er t he e__t__ _ be.. can be f ~_ _.. at a_. e_ _ : a _.. te~~:, ature . Next , the powder on the exterior suraaces e-_' the mold is removed to provide a mold which is still filled w=th loose powder on the interior surface 17 thereof. The loose interior powder must then be removes to yield a hollow mold 19. As t'_~_e passageways with'__~_ the mold can be long and complex, the geomet=-y often precludes the use of a tool such as a brush to aid in interior powder removal . One techr_ic~ue for removing the powder which is disclosed i~ the above applicaticr_ is to wash or flush the powder from t'.~_e inter=or surface. In some cases, and par titular 1y when the fur =he= material used to bind the powder contains a pelvmer, the mold can be removed =rom the powder bed be=ore fir=ng.
It will be understood that other methods might also be used to create ceramic molds directly from a computer model.
For example, selective laser sinterinc might also be used to create such a mold from ceramic powder. Thus, the current invention can be applied to molds made d=rectly from a computer model, regardless of the process used.
A major problem with such techni~:e is that the powder is not always thoroughly removed when a mere flushing operation is used. Accordingly, it is desirable that other techniques be devised to provide more complete and efficient removal of the powder from the interior of the mold.
Moreover, another major problem that arises is that, since the mold is fabricated from a plurality of layers, a non-smooth surface, e.g. a "stair-stepped" surface on the interior of the mold is usually produced during the layering process. while the magnitude of this surface effect can be reduced, by reducing the layer thickness, an undesirable 13 increase in the fabrication time results. Acceptable surface finish is, therefore, not always achievable when using a practical fabrication time cycle. It is desirable, therefore, to devise techniques to achieve the desired surface finish utilizing further operations after the mold has been fabricated and the powder is removed from the internal passages.
Brief Summary of the Invention In accordance with one embodiment, the invention provides a process for removing powder that is lightly bonded to an interior passage of a body to be used as a mold for casting a part of a metal alloy, the body made by a layer manufacturing method, the process comprising the steps of: a. pouring metal particles into the interior passage wherein the sizes of the metal particles are chosen to be smaller than the smallest interior passageway in the body;
b. agitating the body to dislodge the lightly bonded powder;
and c. removing the particles and the powder from said body, wherein any of the metal particles that are not removed from the body will be incorporated into the cast part.
Tn accordance with another embodiment, the invention provides a process for removing powder that is lightly bonded to an interior passage of a body to be used as a mold for casting a part of a metal alloy, the body made by a layer manufacturing method, the process of removing powder comprising the steps of: a. pouring metal particles of the same alloy as will be used for casting the part, into the interior passage wherein the sizes of the metal particles are chosen to be smaller than the smallest interior passageway in the body; b. agitating the body to dislodge the lightly bonded powder; and c. removing the -3a-particles and the powder from said body, wherein any of the particles that are not removed from the body will be incorporated into the cast part.

Detailed Description of the Invention The invention can be described in more detail with the help of the accompanying drawings wherein:
FIGS. 1(A)-1(F) depict a process sequence for making a ceramic mold by three dimensional printing techniques showing a mold during printing, after removal of exterior powder, and after removal of -interior powder;
FIGS. 2(A)-2(D) depict a sequence of operations performed to remove loose powder from within a ceramic mold using microwave energy;
FIGS. 3(A)-3(C) depict a sequence of operations to remove loose powder from within a ceramic mold using vacuum suction applied externally to the mold:
FIGS. 4(A)-4(E) depict a sequence of operations performed for the removal of lightly adhered powder from within the mold by using the agitation action of small metal powder particles;
FIGS. 5(A) and 5(B) depict a sequence of operations performed for providing an improved surface finish on a mold using a non-conformal coating thereon FIG. 6 depicts the use of a filter layer to enable a non-conformal coating of fine particles to be deposited on a substrate made of coarser particles.
As seen in FIGS. 1, 1(A)-1(F), and as described above, a ceramic mold 15 is formed on a layer-to-layer basis by applying successive layers of ceramic powder to a confined region, e.g. as depicted by the arrangement of piston 5 and -S-cylinder 7, and then applying binder material to selective regions of each layer (FIGS. 1(A)-1(D)), as discussed in the above identified Sachs et a1 application. The exterior loose powder and piston/cylinder arrangement are removed to leave a mold having loose powder in the interior cavities thereof (FIG. 1 (E) ) . The interior loose powder is removed, by washing or flushing, as discussed in the above application to produce the desired ceramic mold shown in FIG. 1(F).
As the shape of the ceramic mold can be suite complex, the ceramic powder must travel a complex route in order to come completely out of the casting mold. One technicrue helpful in a washing operaticn is to place the mold in an ultrasonic tank and apply ultrasonic energy to the mold to help loosen the powder on the interior surfaces and thereby facilitate its removal.
Another effective powder removal technique in accordance with the invent=on is described with reference to FIGS. 2(A)-Z (D) . In such method, a mold 17 is immersed in a suitable liquid such as water 21 contained by a vessel 23 as shown in FIG. 2(A). A small amount of surfactant may be added to the water to improve the wetting characteristics thereof. The water relatively rapidly penetrates through the porous ceramic mold and fills the void spaces between the loose ceramic powder within the interior cavities of the mold.
Alternatively, for molds of low porosity, the water may be poured into and contained by the mold itself with no vessel required. The ceramic mold, once placed in the bath of water in container 23, is then placed in an apparatus for applying microwave energy thereto which apparatus consists of a housing 25 and a microwave energy generator 27. A suitable apparatus is a well-k.~owa commercial microwave oven such as those used to cook food. As shown in FIG. 2(B), the microwave generator is tu=nod or. and microwave ene_cy 29 is applied to the mold and therei n to cause tha water to boil. The boil ing action of the water within the loose powder on t'~e interior surfaces of the mold leads to the formation of bubbles 29 of steam and the agitation from the boiling causes the powder 31 to be ejected.
A c=itical aspect of the use of the m'_crowave enercv is that the bell=ng action takes place pre=erentialiy within the loose powder on the i_~.side of the mold. This action is cortrast~d with a boili:g action that weu'_d result when the mold is merely place in a tarac of hot water and subjected to boiling by the applicat_on of heat t: the exterior of the tar~c. In the iatte= case, the prima=,r boil=ng action takes place at the interface between the ware= an3 the tank in which it is contained. Since the mold, and in particular the interior of the mold, is at a somewhat lower temperature, relatively little boiling action takes place within the mold.
As a result, there would be relatively lithe ejection of particulate matte. therefrom. In the case where microwave energy is used, however, the water is heated uniformly throughout the interior cavities by the microwave energy. The loose powder particles, on the interior of the cavities now act in effect as °boiling chips", as that term is understood in the practice of certain chemical processing technia_ues, wherein boiling chips act as nucleation sites where bubbles of steam form. Thus, in the case of boiling by microwave energy, the boiling takes place preferentially at precisely the locations desired, i.e., within the loose powder which is inside the ceramic mold cavities. It will be understood that liquids other than water can be uses for the purpose of causing boiling and ejection of the powder and that the freauency of the microwa~re excitation energy should be appropriately chosen to match the properties of the liquid which is used.

-7_ FIG. 2 (C) depicts the eject=c n of powder at a later stace wherein a substantial amount oz the ejected powder 33 has accumulated at the bottom of the vessel 23. FIG. 2(D) shows the completely emptied mold cavities wherein all the ejected powder 35 has accumulated at the bottom of vessel 23 and the microwave power has now been tur_~.ed of f .
The boiling actior_ described above is sufficient to excel loose powder through passages of relatively complex shape.
However, it is preferable to orient the mold with its pourir_g cup as s::own in FIGS . 2 (A) -2 (D) . Since the steam bubbles tend to rise, it is generally preferred that the main exit passage for the powder be orie__~.ted upwarcly.
An alternate means to create boiling within the mold is to immerse the mold in a ligu_d such as water within a pressure vessel such as an autoclave. While at an elevated pressure, the liauid is raised to, or near to, the boiling point of the liauid. Rapid decompression (reduction of pressure) will induce boiling of the liquid within the mold and the boiling action will aid in the ejection of the powder from within the mold much as in the case of boiline with microwave energy as described above.
An alternative method for the evacuation of powder from the inside of a ceramic mold is described with reference to FIGS. 3(A)-3(C). As seen therein, the mold is fitted with a cap 37 and suction is applied to the pouring cup of the mold at tubE 39, the suction drawing air in through the wall of the mold, thereby I~oseninc the powder therein and causing the loose powder Z4 from the interior of the mold to flow out from tile pouring cup as shown in FIG. 3 (C) . Ceramic molds are typically fairly porous, a requirement generally imposed by the casting process itself, and the air will flow relatively freely throuc:~ the porous mold wall . After the suction is applied for seine time period, the cap 3'7 c,an be removed yielding the empty mold shown in FIG. 3(C).
While the powder removal methods descr'_bed above are high? y effective, it is possible that a small a..~.,ount cf powder might be left behind on the interior surfacss of the mold, such powder being especially lightly aahered to the interior walls of the mold. A further process that is useful for removing such ligttly adhered materials is dep=cted i_~. FIGS.
4 (A) -4 (E) where=r. a s~na_1 ruantity cf a mater'_a 1 whi ch wil l act to remove the loose powder by a combinat'_cn of abrasion and impact actions is poured into t~:e mold (F .G. 4 (A) ) . In the preferred embodiment depicted very s;,ta_1, generally spherical metal particles 51 are poured into a mold 47 which has a small amount of lightly adhered powder 39 on its inter for surface . The size of the metal par =icles must be smaller than the size of the smallest passageway inside the mold. Thus, for example, if a thin-walled turbine blade is being made with a wall thickness of 570 microns, the particle sizes should not exceed 500 microns and preferably should be in the 100-200 micron range. After being poured into the mold, the mold is covered with a cap 55, as shown in FIG.
4(B). The mold car, then be turned over as shown in FIG. 3(C) allowing the metal particles 57 to tumble with_n the mold and thereby impact and loosen the lightly adhered ceramic powder 59 so that it is free to move within the mold. It is understood that the tumbling action can be effected in any manner, for example, by alternately rocking to mold back and forth or by tumbling it continuously in one direction. It is further understood that the shape of the partic? es poured into the mold can vary. However, it is found that spherical particles have an advantage of being highly flowable and, therefore, easy to pour into and out of the mold. It is also _g_ urde_stocd that fort he. more aggressive agitation of the mold Wit h the ts:zbling particles inside can be util'_zed, e.g. as by vibe acing t_te mold or by other means , as long as the impact of the partic'_es is not so violent as to cause damage to the mold.
The material of the particles that are poured into the mo~.d can also vary. For example, it is possible to use ceramic par titles rather than ~retallic particles . However , it appears pry=enable to use metal lit particles as they will have mor=_ mass aid momentum and, therefore, do a superior job of knocking loose the lightly adhered ceramic powder. After su_tabie tumbling, cap 55 is removed and the metal particles 61 together with the loosened ceramic particles 63 are poured out, as shown in FIG. 4 (D) . The resulting finished mold 19 is shown in FIG. 4 (E) .
A further aspect of the process is to use metallic particles of the same alloy that will be case within the mold. Such use provides an advantage that, should a few, e.g., one or two, metal particles be left behind in the mold, they will simply melt and be incorporated into the casting when the alloy is poured. One problem that may arise in achieving such incorporation is that the ceramic mold often goes through a firing step to preheat the mold before the metal is poured. As a result, there is a risk of oxidation of one or more metal particles that are left behind during the firing step. Accordingly, it is a further aspect of this invention to use metal particles of the alloy that is to be cast and to plate such particles with a very thin plating of a noble metal material, such as platinum. In this manr_e~, the metal particles will resist oxidation during the fir_ng step and will then become readily incororated into the alloy during casting.

-IO-T_n anot'_her embodiment of this invention, t he powder which is peur~d into the mold can be a soluble material. For example, a metal salt, such as sodium chloride, can be used as the powder for tumbling. Such powder will act to loosen unprinted, but lightly adhered, powder on the interior of the mold in a manner similar to that described above. Any salt powder that remains after the tumbliac operaticr_ can be removed by immersing the mold in a solvent, such as water, and dissolvinc the salt out.
A fur ther advantage in using the t=c.~_~.icrue of FIGS . 4 (A) -4 (E) is that the tumbling action of the par:' cles (w::ether they are ceramic or metallic) improves the sur=ace finish of the interior surfaces of the mold. In processes which use powder particles to form a part, such as in a three-dimensional printing process, for example, some powder particles on a surface of the part being fabricated may be only lightly bonded with the majority of the particles protruding from the surface. The tumbling action of the metal or ceramic particles in the mold can remove such barely bonded particles from the interior surfaces thereof. Further, the tumbling action also tends to smooth out the stair-stepping configuration that occurs between the layers of the printing process. Such stair-stepping configuration is the result of the sequential building of the par in layers of finite thickness.
Another approach to improving the surface finish of a porous part made directly from a computer model by methods, such as a three-dimensional printing process, is to coat the inter for of the mold in such a manr_er t:~at the under l ying surface roughness decreases as the coating grows. In order to create the greatest improvement in the surface finish of the molds, the coating should preferably be non-conformal. FIG.

5A shows a surface, as created by a three-dimensional printing process, for example, which sur_ace is constructed of three layers, resulting in steps 67. Steps 67 are a primary source of surface roughness for the mold. In addition, defects 69 may arise in an individi:al layer, which defects can also lead to surface rougrness . Figure 5H shows a non-confo "al coating 71 which is th'_cker i_~_ the deoressior_s of the stair-step configuration and which, there=ore, tends to smooth the sur::ace .
While coatings have been used in the mold casting industry, they have bee_~. most prominently applied to sand molds. Their primary purpose in such application is to provide a barrier to the molten metal used in the casting process so that it does not penetrate into the sand mold and, in the process, the surface finish of the casting tends to be improved. In standard practice, these coatings are applied as paints, using either brtah_ng, spraying or dipping technigues .
The coatings are generally appl ied in relati~rely thick layers, e.g., 150-2000 microns thick, and often are obtained by multiple applications of the coating material. The coating materials generally are prepared with a very high solids content, e.g., as high as 40°s by volume in order to build up such large thick:-~esses . 8ecaus.e of the methods of preparation and application, the thickness of the coatings can not be well controlled and, hence, the geometric control of the casting may suffer. bioreover, the ability of the coatings to coat the mold in a non-conformal fashion is limited. Accordingly, existing sand mold coating technology is poorly suited to the needs of molds made by layer manufacturing methods, such as three-dimensional printing processes, for example.
In accordance with the invention, however, slips of fine particles can be used to create a casting, and preferably a non-c:,nfor;~a1 coating of the mold surface. In c,e embodiment, for example, a slip, or dispersion oL partic__s in a liauid vehicle, is poured into the mold and the lic,:id vehicle is made to flow into the porous mold by the actic~ of capillary forces which draw the liquid into the mold. T.e mold acts as a filter with respect to the particles, with the result that these particles are deposited on its surface . 'I .e steps which have a greater pcre~ volume per unit area at the surface of the porch's body tend to draw more liquid in and, he::ce, causes the depcs=tion of great=r amounts of par _iculate ;;.utter from the s1 -is . Thus, the zon-ccnfornal ar_d a general _eTrel ing nature o~ to ccating is achieved.
Slip casting is a method that has been k:~wn in the art for making cerar.:ic bodies . Slip casting is the depositicn of particles on the surface of a porous mold thrcu=h the flow of a lir-uid vehicle which disperses those partic_es. The term, "slip", usually refers to the particulate dis~ersior_. Slip casting is commonly used for the fabrication of complex shaped components by casting them on the inside of porous forms.
These forns are usually made of two halves and can be split to remove the part after drying. The cast thickness is found to be rather uniform and makes complex shapes accessible to the ceramic designer.
In accordance with the invention, the cast layer thickness will be much thinner than that cbtained using conventional slip cast parts and will be made to adhere to the surface of the part by the physical and chemical composition of the layer so that it will not spall or separate during use.
Thus, in the present invention, the slip cast layer is intended to become a part of the finishes mold and to cons=tute the inner layer or face coat of th a mold. It is this layer which is exposed to the molten meal during the casting operation. The casing c~ tie inns= layer thus presents another advantage since the com~ositic: of the inner layer can be different than the compcsition of t .e bulk of the mold and, thus, a material may be chosen which either minimizes reaction with the molts-~_ metal or promotes the nucleation of grains, depending or- the parti~~slar casting application.
The cast layer thickness will have to be accounted for in the design of the component . Fortunately, the layer thick:-iess can be precisely controlled s_nce t:ze deposit=on rate is a function of certain process parar"eters. T!.= cast layer thickness, h, varies with time, t, according tc .he following approximate eauation, = d= e' 2 ~
t 18K~ (y-1)(1-E)= r where d is the particle size in t::e slip, "E" is the void fraction of the cast layer, "u" is the viscosity of the liquid vehicle, y is a function of the solids content of the slip, "y" is the surface tension of the licruid vehicle, K is a constant, and r is the pore size in the mold shell. The radius of curvature of contoured shapes will also change because of the deposition of material on the curved surfaces of the component. In a typical application, sub-micron pa rticles (e. g., having sizes of 0.1 - 1.0 microns) can be used and the process typically builds up a layer havi= g a thickness of up to 60 microns thick, for example. As i= is only the radii of dimensions comparable to the layer thickness which will be effectec', most geometries w_11 not be significantly altered. The use of thin cast layers is indicated when it is desired to minimize the alteration of the geometry of the mold _1G__ during the casting.
Sli ps of rel a=ive_y lc~r sc._ids cc :tents are pre=erably used in accordance wit h the . ve-tior_. Such slips can help to create the thin cast layers whit:. are the desired aim of this invention as the thick.~ess of the cast layer depends on the quantity of liquid vehicle abscrbed into the porous body and on the solids conte~t of the sl_o wric'_~. is filtered out as the liquid vehicle is absorbed. Thus, by a combination of control of the viscosity and surface te.~.s:c:. of the liauid vehicl a of the slip, the solids conte::t ad t he casting time, the slip thickness may be co.~.trolle~?. Typically, slips with a solids content of between 1% ar_d 10~ by vclume might be used.
Fine surface finish and gced mechanical adhesion to the mold shell surface requires the use cf very fine powders.
Slips of such very fire powders ca_~. be obtained by a variety of methods known to persons skilled in the art. Fine powders of ceramics or metals are available commercially.
Alternatively, such powders may be classified by sedimentation' or centrifugation. Particle size distributions as narrow as 0.2 to 0.3 microns can be made by such methods. Such narrow size distribution slurries dr~~ to fore films with surface smoothness that is of optical quality. These slurries can be made stable with respect to flocculation by proper selection of the pH or by the presence of a dispersant. It is well within the skill of those in the a_-~ of fire particle dispersion to determine the exact conditions by which to disperse fine particles of a giver material. Alumina particles are, for example, dispersed in water when the pH is below 4.
When the mold produced directly from a computer model is made with small particles, for example, particles that are sub-micron in size, the methods described above may be practiced with litale or no special p=eparat_on of the mold.
In such a case, the fine porosity of t..e mold :vill filter out the particles ir_ the slip and these particles will deposit entirely on the surface of the mold.
Often the mold is produced using larger particles and, as a result, the pores of components prepared by rapid prototyping methods are freauently larger than those needed to filter out the fine particles of a slip. Two approaches are possible in such a case. In one approac, the slip is designed to penetrate a small distance into t=a porous mold, but to aaglomerac_ and stop the pee=ratio- after it has reached an approximate, but determin~le a-controllable depth. Such controlled penetration can, z~= example, be achieved by utilizing particles which have a t_adency to lock against each other as they come into contact. For example, plate-like particles will behave in this man.:.=__-. A possible benefit associated with the partial penetration of particles into the mold is that the slip-cast layer will be more adherent to the mold.
Another approach, in the case where the pores of the mold prepared by rapid prototyping methods are larger than that needed to filter out the desired fine particles of a slip, is to suitably prepare the porous parts before receiving the coating of fine particle suspensions. In one case the pores on the surface of the part are modified so that fine particles of the slip are filtered by the surface. The pore structure can be very effectively modified by dipping the part in a gel - forming solution and drying it prior to slip casting. The gel forming material creates a filter within the large pores so that the fine particles are appropriately filtered during slip casting. The gel can then be removed after coating by firing in the case of an organic gel. Alternatively, in another case, inorganic gels can be used, which gels remain in the compcnent duri~c use. Colloidal silica, of the type used for investment casting mcl,-~.s, for example, is an appropriate gel forming material whit= can be used. The colloidal silica is drawn to the sur=ace during drying to fill the large pores with nanemeter-scale silica particles. This procedure produces cr. effective filter which pre~rents passage cf the alumina par titles uper_ subsequent slip casting. Such p~_coat procedure need not be carried out with colloidal silica. More refractory solids Iik~ that ef zirconic, yttria, o. alumina may be just as effective. rigure 6 s:~cws thre= dimensional printed pare made cf three layers of fairly la==a particles 73. The filter lave= 75 has been formed withi: the pa=t and at the sur=ace ef to part. Fine particulate= 77 have been slip cast onto the fil ter layer with the result t .at the sharp steps in the pa=t have been smoothed into gentle ccntours 79.
A variation on the above procedure is to use a partially saturated preforn in a reactive casting process. The Iictuid used, however, is selected to flocculate the slurry and plug the pcres before much slurry has beer. sucked into the body.
An example of this reactive casting approach is to cast alumina slurries or. parts slightly saturated with ammonium hydroxide solution. The acidic alumina dispersion flocs within the interior of the part when in comes in contact with the basic solution.
One problem which can arise is the cracking of cast films, i.e., cracks can form during drying as a result of capillary stress. The formation of cracks has beer the subject of recent investigations, e.g., as discussed in the article of R.C. Chiu and M.J. Cima, "Drying of Granular Ceramic Films: II. Drying Stress and Saturation Uniformity";
submitted in 1992 to the Jou~a1 or t'~e American Ceramic Society. As described therein, cracking can be effectively pre~re_nte3 by several methods . For example, t a cracks will not for-z in films with thickness less than a critical value.
The critical cracking thickness for granular a_umina films is roug'.~_l y 60 microns . Thin fil ms, a .g. , having t hiclsnesses less than _60 microns, do not present a cracking problem since they do not have enough stored elastic energy to fore and extend cracks. An alternative approach is to strengthen the film by the addition of a binder. Both organic and inorganic binders can be used. Excellent examples fcr investment casting toolig are zirconium acetate, yttrium sols, o. boehmite sols whic . are comp atible w=th ac=dic alumna slips. These binders will nct only strengthen the green films but a_so improve the bondi ng of the film to the thr ee dimensional pr inted component after firing.
While the above description of various embodiments of the invention discuss a variety of preferred embodiments thereof, modifications thereto may occur to those in the art within the spirit and scope of the invention. Hence, the invention is not be constrsed as limited thereto, except as defined by the appended claims.

Claims (3)

CLAIMS:

1. A process for removing powder that is lightly bonded to an interior passage of a body to be used as a mold for casting a part of a metal alloy, the body made by a layer manufacturing method, the process comprising the steps of:
a. pouring metal particles into the interior passage wherein the sizes of the metal particles are chosen to be smaller than the smallest interior passageway in the body;
b. agitating the body to dislodge the lightly bonded powder; and c. removing the particles and the powder from said body, wherein any of the metal particles that are not removed from the body will be incorporated into the cast part.

2. A process for removing powder that is lightly bonded to an interior passage of a body to be used as a mold for casting a part of a metal alloy, the body made by a layer manufacturing method, the process of removing powder comprising the steps of:
a. pouring metal particles of the same alloy as will be used for casting the part, into the interior passage wherein the sizes of the metal particles are chosen to be smaller than the smallest interior passageway in the body;
b. agitating the body to dislodge the lightly bonded powder; and c. removing the particles and the powder from said body, wherein any of the particles that are not removed from the body will be incorporated into the cast part.

3. A process in accordance with claim 2, wherein the metal particles are coated with an oxidation resistant metal, so that the metal particles will not subsequently oxidize before any part is cast in the mold.