A good jewelry caster must be familiar with some of the basic concepts of metallurgy, ceramic science and mechanics, so that he can use these technologies for the benefit of his trade and art. Jewelry casters over the centuries have done fine work, as is attested to by gold castings done by early Egyptians, Africans and the South American Indians (Fig. 1). The lost wax method of metal casting was an early invention, going back three to four thousand years, when beeswax was the only source of wax. Today’s metallurgists look on such past work with considerable respect. For example, the Ashante of Africa cast some very intricate designs, equal in detail to almost anything being cast today. How did they obtain the necessary casting force? They didn’t need to. Casting force beyond that of gravity feed was not a requirement because the casters simply heated the carefully-made clay molds to the same temperature as the poured metal. Then, by a tilting operation, they filled the very hot mold with the molten alloy and simply waited for the mold to cool, with the metal inside solidifying in situ, yielding the casting when the mold was broken open. Apparently the South American Indians learned this technique independently.
Today’s jewelry caster doesn’t have to spend laborious hours building up the elaborate clay mold around the wax pattern, nor does he have to heat his molds to those high temperatures. In addition to gravity feed, he has today’s extra devices for forcing the liquid metal into the intricacies of the pattern cavity at sufficient speed into the relatively cool mold to prevent premature solidification. The modern jewelry caster regularly casts plant leaf designs, all varieties of filigree, and even such challenging patterns as a thistle, with its numerous, sharp, needlelike extensions.
The metallurgist continues to have respect for the jewelry casters’ skills. Yet his thoughts turn to possible further challenges of design intricacy that might be cast using new skills that emerge from today’s evolving technologies.
One interesting aspect of jewelry casting is that its technology requirements are unique-quite different from industrial casting. This means that a jewelry casting technology must be developed on its own.
For example, industrial castings have a solidification behavior that has been well-researched, and the industrial caster can refer to this research as a foundation for judgments for his gate- riser system, the use of vents, chill plates etc. This technology is based on castings that solidify over several minutes in the average casting weighing over a pound.
In jewelry castings, however, the mold cavity is usually filled in about one-tenth of a second (per unit on a casting tree ), and the casting begins to solidify at sharp corners and edges within the first second, taking only 5 to 15 seconds for the entire unit to solidify.
Also the mode of solidification is quite different from that in industrial casting. In industrial castings ( of one pound or more per casting), the wall thickness increases gradually, according to a simple law of wall thickness versus time from moment of complete mold filling. In casting a ring, the individual grains that develop may be thicker than the cross-section of the band of the ring. Experiments indicate that these individual grains “pop” into existence in discrete bursts to full grain size. This happens with little growth of the grain itself before the neighboring grains-which have also popped into existence-begin to crowd around it.
There is also a major difference between the sand mold of the industrial casting and the hard, rigid ceramic mold used in the lost wax method. When the mold is rigid and hard, the casting tends to shrink in the mold cavity, getting hung up by the friction of the mold wall and the resistance of any shoulders that might be present. The result is the problem of possible “hot tears.” In the following discussion of jewelry casting, much of the technology relied on comes from experience and research on jewelry and dental casting rather than from industrial casting. In this article we will cover:
- Casting intricacy and the mode of solidification
- Molten metal and superheat
- Mold material and mold temperature
- Castability charts.
Casting intricacy/ mode of solidification At first glance, it might seem that because each jewelry design is unique, it would be hopeless to try to give a quantitative measure of casting intricacy-but this need not be so. All designs- or at least components of each design-can be classified as three general basic shapes: spheres, rods and discs.
Small beads or nugget-like components can be considered small spheres; wire-like or filigree components are thin rods; and coin shapes, plant leaves and flower petal shapes are thin discs. For each of these basic shapes, the cross-section dimension can be measured or fairly accurately estimated in units of inch fractions or millimeter fractions. On this basis, if we consider a some- what massive class ring, the band portion of the ring might be considered a bent-around disc of 1/16th of an inch thickness, whereas the massive head might be roughly estimated to behave like a 1/4th-inch diameter sphere.
Table I gives the Solidification Rate Index “R ” (last column on the right) for the three basic shapes broken down into several section sizes; “D” the average diameter for each. These “R” numbers are actually proportionalities of the surface-to-volume ratio for each of these. Indeed, Raub and Ott (see reference 3) use the same capital letter “R” as the symbol for the surface/volume ratio for a configuration such as a sphere. In the solidification of a casting, the heat is extracted at a rate proportional to the surface area and the amount of heat to be extracted is proportional to the volume. Thus, the “R” values are a logical index of the solidification rate of the casting, assuming all other variables remain the same: mold temperature, casting temperature, allow etc. The “R” values for the remainder of this article will be treated as an index for comparison purposes as to rate of solidification.
In Table I under the three basic shape designations is a simple equation for calculating the surface-to-volume ratio of the basic shape in question and for different estimated cross-sections.
In the case of the sphere, the equation is:
R = 6/D
This means: six, divided by the average sphere diameter D.
Suppose a component of a de- sign is a bead shape, which can be considered an approximate sphere. The average D would be obtained by estimating the diameter along three different directions through the center of the bead and obtaining their arithmetic average.
In the case of a rod shape, the equation is:
R = 4/D
If the cross-section is oval in shape, the average D would be the average diameter along two perpendicular directions through the cross section of the rod.
In the case of the disc shape, the thickness “t” need be estimated, which represents the average D. Note from the table that for the same average D, the sphere is the most likely to freeze first, having the highest “R” index for a given cross-section D, whereas the disc solidifies the slowest. This can be explained by noting that the sphere has its heat extracted along three perpendicular directions about equal, the rod has two perpendicular directions for heat extraction ( the rod axis direction effectively contributes no heat flow) while the disc has only one significant direction of heat flow-perpendicular to the surface of the disc.
Also note that the smallness of the D dictates the rate of heat extraction; the smaller the casting, the faster the solidification simply because the surface-to-volume ratio increases with casting smallness.
In Table I, the fourth column gives some casting shape examples: poppy seed, dry barley grain etc. To be sure, jewelry generally does not reach the dimensions of a door knob, a standard Coke bottle or a man-hole cover. These are included to add perspective on solidification rates for various sizes of castings. If a door knob takes about 150 seconds to solidify for a given set of casting conditions, the poppy seed should freeze 300 times faster (600:2) or in about one-half second.
Not only is the rate of solidification important, but also the mode of solidification of the casting and its sprue system. The ideal mode to strive for is to have the solid-liquid interface move progressively from the far end of the casting toward the sprue junction, just as the liquid level might rise in a mild bottle filled with liquid.
This scenario of solidification is in part already favored by the fact that the far end of the casting experiences the maximum liquid metal “head pressure,” in both the centrifugal and the gravity feed method, the higher pressure favoring a more intimate contact with the cooler mold surface.
Even in the vacuum-assist method there is a slight favoring of the higher pressure at the lower end of the mold where the quality of the vacuum is a bit better. Furthermore, the thin parts of a pattern to be cast are generally mounted in the mold away from the sprue end, the heavier portion of the casting being located at the sprue. Sharp corners and edges will freeze first, such that the casting solidification starts with a skeleton of the casting framework (Fig. 2A, 2B, 2C, 2D and 2E). The solid builds up on the skeleton with the solid-liquid interface moving toward the entroid of the casting cavity and simultaneously favoring the direction toward the sprue junction.
Being aware of this solidification scenario, the caster can mentally test the placement of the sprue on the wax pattern to see if the mode of solidification will result in the sprue-casting junction being the last to solidify. In general, one finds that in attaching the sprue (or sprues) to the heavier sections of the pattern, the mode of solidification will work out properly. Compromises are sometimes necessary, as one does not place a sprue on fine casting detail such as on the lettering of a class ring. Fig. 2F shows a wax cast ring with two interesting defects. The view is on a cut through the ring in a vertical plane half-way through the band width. The sprue had been attached to the band portion of the ring opposite the head. The major porosity is through the heavy section of the ring. The band portion of the ring was, in effect, a flat strip bent around in a circle, its “R” being high (that of a disc). It solidified well before the head of the ring, the liquid metal feed having stopped before the bulky portion solidified. That bulky portion solidified from the outside inward, causing a liquid pocket that had to shrink in volume by about 5%, hence the microporosity-that is porosity between the dendrites and the grains.
The second defect is found at the sprue junction where fewer but larger voids can be found. In this case, the jet of hot metal flowing along the sprue channel impinged on the mold wall at the entry into the pattern cavity, causing a “hot spot.” This spot also cooled slowly, causing a delay in solidification and, hence, a localized set of shrinkage voids.
3A Figure 3A shows the spruing used for a cast gold alloy mustache comb. In this particular case, the caster would estimate that the “R” for the comb teeth region is approximately 400. For the back portion of the comb, the “R” estimate would be approximately 40. Thus, it is necessary that the sprue be attached to the back opposite the teeth section. Split sprues are shown in Figure 3B to offset the formation of a hot spot on the mold wall. If a heavy class ring were to be cast, there is the problem of spruing opposite the head of the ring, because in the casting the “R” for the band would be much higher than for the bulky head region.
Figure 3C shows a cast ring centrifuged for five seconds after casting to separate the liquid fraction from the portion that had solidified in those five seconds. The incomplete portion of the casting near the oval stone setting indicates an entrapped liquid pocket had existed surrounded by solid to which the liquid feed had been cut off before complete solidification had taken place in this region. Such a liquid pocket would have solidified, but with the liquid metal feed cut off a small shrinkage void would have occurred, perhaps as a depression on the surface or a gas pocket beneath the surface.
Molten metal state and superheat
AJM ran a series of articles on General Metallurgy from August 1985 to February 1986 written by Paul J. Finelt.* In these articles, phase diagrams are discussed as well as the basic process of solidification for the complete liquid state, through the mushy state, and on to the complete solid state. It would be well for the reader not familiar with this aspect of metallurgy to read these articles. However, for convenience, some of the phase diagrams and solidification behavior will be summarized as needed. Figure 4 gives a set of cooling curves from all liquid to all solid for four typical precious metal alloys.
* Aron Finelt, Paul’s father, reintroduced the lost wax method for jewelry in the early 1920s. Dental laboratories had earlier adopted the lost wax method from jewelers; Finelt, a dental technician, saw the possibilities for casting intricate designs, and thus launched the Finelt Jewelry Casting Co. in New York City, operated after his death by his sons Paul, Howard and Jack for many years.
Alloys melt over a temperature range, the lower point of the straight line depicting the mushy state, being the solidus below which the alloy is all solid, while the upper point, the liquidus, is the temperature above which the alloy is completely liquid. All pure substances such as the fine silver, exhibit a cooling curve (see Fig. 4) with the mushy state as a horizontal line. The sub- stance begins to melt on the way up at the horizontal line and begins to freeze on the way down at the same temperature. The so-called mushy state is simply a mixture of solid particles immersed in the liquid, much like the investment slurry poured into a mold with the wax pattern in place.
The mushy state is more clearly understood by referring to the phase diagram for an alloy series. Three phase diagrams are shown in Figure 5 for the alloy systems: gold-copper, gold-silver, and copper-silver. In each of these diagrams, areas are shown in which the alloys are all liquid, all so- lid, or a mixture of solid and liquid-the mushy state. These various areas are delineated by lines which have been plotted from actual experimental data.
The base line of each diagram is a scale giving the gold or silver content for all possible binary alloys of each system. The vertical line at the left of each diagram is the temperature scale for reading off the temperatures to which the various possible alloys are heated. The SS line traced across each diagram is a connected line below which the alloys are all completely solid-it is called the solidus. The LL line, known as the liquidus, correspondingly delineates the region above which the alloys are all liquid.
The area, or areas, lying between these two lines are the mushy state regions. It should be noted that in general, the various possible alloys have a temperature range for the mushy state (with some minor exceptions). The pure components- gold, silver and copper-all melt at one temperature. The mushy state exists, but only at a constant temperature for each. The heat added to an alloy which raises that alloy above the liquidus is called the superheat. Casters are apt to attach importance to the casting temperature, but for basic reasons, the superheat is the important measure for determining the best point along the temperature scale for pouring. It is the superheat that gives the measure of safety-the heat “cushion “-that is needed to assure that the metal enters the mold and fills the cavity before freezing takes place. As will be explained later, in some cases a 10 to 20°F temperature interval above the liquidus will give a good casting, whereas in other cases (intricate castings, for ex- ample ), the temperature interval above the liquidus should be as much as 150°F. One could, under high speed of mold filling and with non-intricate castings such as bulky figurines, cast from the mushy state, as is done in industrial slush casting; but this is not recommended with present day casting procedures for jewelry.
One rule of thumb to follow for jewelry casting is to cast with some superheat, but to favor the low side. This is to save wear and tear on the crucibles and the melting equipment, to avoid loss of volatile elements in the alloy such as zinc, and to prevent harsh surface reactions when metal that is too hot comes in contact with the cooler mold surface, producing unwanted gases and heavy discoloration products on the casting surface. This type of discoloration adds extra man-hours of work in the finishing of the casting.
Jewelry alloys, of course, are more complex than the simple binary alloys referred to above. They are quite often complex alloys of gold, copper and silver -the gold and copper contributing to the yellow color, with the copper also useful for strengthening. The silver is useful in obtaining interesting shades of alloy color, also some hardening and some lowering of the li- quidus and solidus for easier melting and casting. In addition, zinc, nickel, silicon etc. are add- ed for reducing the tendency for gas absorption, to make white gold alloys, and to render the castings “clean” looking-that is, free of discoloration when taken out of the mold. Other elements are sometimes added to reduce the melting range, as in the case for solders.
In jewelry casting, the mold material is generally a ceramic fired gypsum plaster (essentially the same material as in wall plaster). The firing not only eliminates the wax or the plastic part of the pattern, but it also removes most of the chemically-held water in the gypsum, and bonds the silicate powders with a slight amount of melting of the particle surfaces.
After the firing, usually called burnout, the mold is kept hot to add to the cushion of heat already supplied by the superheat of the cast metal to prevent premature solidification. The wax or plastic of the pattern must be completely burned out, as any residual organic material will generate gases when hot metal comes in contact with it. Also the residual material can seal off the interstices (spaces) which develop in the investment, replacing the driven-off chemically-held water; interstices are necessary for air escape as the hot metal plunges into the mold cavity.
The essence of investment casting is that in the casting, there is a race in which, on the one hand, the metal is attempting to flow fast enough to fill the mold cavity completely before freezing and, on the other hand, the cooler mold (relative to the molten metal being cast) is attempting to freeze the metal as soon as the mold is filled. When the race is lost, we have a miscast.
Mention was made above of the high strength imparted to the ceramic mold on firing. This high strength is needed to give the mold the ability to withstand the on-rush of the heavy hot metal plunging its way into the mold cavity, attempting to erode the cavity’s sharp edges and corners. This, unfortunately, has a negative aspect to it. A metal that has been cast continues to shrink relative to the mold after it solidifies, and the strong mold resists this metal contract. Sometimes it resists to the extent that the solid but still-hot metal is pulled apart at the shoulders, causing what are sometimes called “hot tears.” To minimize this effect, casters should avoid quenching the sprue end of a casting on cooling the molds before breaking them up, as this accentuates the metal shrinkage relative to the mold.
In the race referred to above between the metal filling the mold and the solidification of the metal as it is cooled by the cooler mold, the critical factor is the heat transfer rate. A good part of the retardation of the heat transfer rate comes from the heated mold. However, there are also two properties working against this: the high conductivity of the metal and the conductivity of the massive mold (the mold’s conductivity is on the low side, on the order of refractory brick used in lining some furnaces).
The additional property of the combined molten metal and solid mold that affects the heat transfer rate in the “wetting” of the mold by the liquid metal; again, this is on the favorable side, as metals tend not to wet ceramics (mercury on glass is a good example). However the wetting behavior differs for different metal/investment combinations. A quantitative and standard measurement of this property is obtained by determining the contact angle of a liquid metal droplet resting on a flat horizontal mold surface called the substrate (Fig. 6). When the contact angle is small, as in the case of a lead/ tin solder on a clean copper surface, the wetability is considered high and consequently the heat transfer rate, due to the interatomic intimacy between mold and metal, is also high. A small change of approximately 30° in the contact angle can change the heat transfer rate by a factor of two or three.
Mention was made of hot spots on the mold wall produced by impinging hot metal. The metal is generally hotter than the mold by at least 500° F, and metal flow from a sprue might be jet-like. At sudden changes in flow direction caused by a shoulder in the path of the flowing metal, another jet-like action might occur, causing local heating of the mold surface.
Such hot spots are regions of possible entrapped liquid metal as the surroundings solidify. This entrapped metal gets no liquid metal feed as the last portion solidifies. The 5% or so shrinkage of the liquid on solidification then asserts itself as a small void or porosity, sometimes on the surface and occasionally in the interior.
The air in the mold must escape to permit the complete filling of the mold cavity with molten metal. Most of the air manages to escape via the sprue channel as it is in the nature of metal flow in casting first to seek out the furthest reaches into the cavity. Thus the air is pushed back toward the sprue where the molten metal pressure is on the low side to let the air get by. The air that does not escape by this route must be pushed out through the permeable mold.
Therefore, it is rather rare that a mold shows pockets of entrapped air mingled with the metal casting, but on occasion it does occur; as in the race between mold filling and metal solidification, the metal freezes while there is still some air in the mold cavity. In this case, it is not so much that the air cannot escape, but that the air pressure suddenly ceases and, thus, there is no pressure on the air to force its escape. Such defects appear as rivulet-like surface defects. When such defects occur, the remedy is rather simple. Cast at somewhat higher temperature and/or add a bit to the casting pressure (more winds on a spring-wound casting machine) to increase the rate of mold filling. A higher mold temperature will also help.
There is the problem of casting extraction from the mold. If the mold has been partially wetted by the liquid metal, or if the mold temperature and the casting temperature are on the high side, there is an increased tendency for the investment material to adhere tenaciously to the casting, making extraction difficult.
Fortunately, investment manufacturers are aware of this problem, and they have found ingredients that make the mold easily broken when watersoaked; by reducing the wetting tendency by the metal, bonding of the casting to the mold is minimized.
Although there is a wide range of temperature to which molds can be heated for casting, there are limits. Gypsum-bonded investments tend to break down chemically above 1350° F, especially if graphite has been added for the clean cast effect. For the higher temperatures needed for casting the platinum group metals, there are the zirconia, phosphate-bonded and ethyl silicate-bonded investments which can be heated to 2200° F for casting. No doubt if higher mold temperatures are desired, the investment manufacturers can develop such investments.
In the above, frequent references were made to variables that the caster has under his control. Generally he has to begin with some conditions that are prefixed for him-e.g. the particular jewelry design in the form of a wax or plastic pattern. The alloy, or at least the precious metal content, whether it be a karat gold or some other precious metal, is also generally prefixed. He may have available only one or two types of casting facilities in his casting shop, and so the casting method may be prescribed beforehand.
But within this framework, or these fixed conditions, he has some measure of flexibility, and it is here that he can exercise his acquired skill. For example, he may be able to select the alloy, given the precious metal content. He may have a choice of investment material. He certainly can choose the mold and the casting temperatures over fairly wide temperature ranges. There may be a choice of casting method: For the highly-intricate castings, those with the high “R” values, the choice should favor that which uses the higher casting force. Vacuum-assist casting is best suited for casting a large number of pieces per tree.
There may be some control over the casting force, e.g. the speed of rotation of the centrifugal machine, the air pressure on the molten metal, and the degree of vacuum. In some cases, the caster can use a gas cover or flux cover on his melts to favor the cleanliness of the cast metal. It is in the judgments he uses in these variable controls that the caster displays his skills.
As an aid to his acquiring casting skill, it is suggested that castability charts be made available to him for different alloys, different mold materials, different casting intricacy, and for different casting machines.
All metals that can be melted are castable-provided it is possible to fill the casting intricacy in the mold before the metal freezes. Molybdenum and columbium, both having their melting temperatures in the 5000° F region, have been cast. Tungsten and rhenium seem not to have been cast as yet, but, no doubt, if it is necessary, they can be. At the other extreme, mercury, the liquid-at-room-temperature metal, can be cast-but the mold temperature has to be well below-70°F, and, of course, the ambient temperature must remain below the -70° F mark. Such solid mercury castings can be invested in the appropriate investment, like wax patterns, and the solid mercury is melted out to produce the pattern negative in the mold.
If a set of charts (Fig. 7) were available to the caster for different alloys, different investments and for different types of casting methods (centrifugal, vacuum-assist etc.), the caster would first select the appropriate chart. He then must be guided by the intricacy of the patterns to be cast. Castings on the thin section size or with filigree or with other design features indicating a high “R” value should be cast from a region fairly well into the “good casting” zone. Patterns estimated as having low “R” values, such as bulky figurines, may be cast on the line of demarcation between “good casting” and “miscast.” The region gives two temperature conditions: the mold temperature and the metal casting temperature. Here experience can help in making fine distinctions as to the best casting conditions to use. Note that for the gold jewelry alloys, it is not required ever to go over 1300° F in mold temperature, and that little, superheat is necessary if the mold temperature is this high. Also, note that as the mold temperature decreases, the required superheat increases.
One purpose of this article is to urge the jewelry alloy manufacturer to publish charts such as that shown in Figure 7 for all the alloys it markets-charts suitable for different investments and casting methods. At least the liquidus-solidus ranges should be made available, and if a casting temperature is recommended, it should be at some fixed temperature interval above the liquidus in the range of 50° F to 150° F. The caster can then, with experience, learn to adjust his casting conditions above and below this superheat, depending upon the intricacy of the castings and the method of casting. For example, the vacuum-assist method of casting, as against the centrigugal, requires a higher casting temperature because the mold-filling rate is a bit slower.
Should castability charts such as that described herein be made available, it should reduce considerably the learning time for a caster to become fully skilled in his trade.
This article on some jewelry casting technology is hardly the last word on the subject. The references below will give some additional details for those who wish to become more familiar with certain aspects of the technologies discussed. In writing this article, the authors sensed that a book on the subject could be put together from information available in scattered sources, and are considering this project for the future.
HooverandStrong.com | 800-759-9997 | (fax) 800-616-9997