Controlling refractory erosion in induction furnaces: the linkage between equipment, alloy system and slag environment affects refractory effectiveness.
The classic conditions for metallurgical corrosion testing involve not only a corrosive media, but also circulation as in the case of a salt spray test. The analogy holds true for a channel furnace inductor or a coreless furnace crucible where the environment is under power from the electromagnetic coil and circulation is intense. Reactive materials are continually brought to the refractory interface, and the byproducts of the ensuing reactions are carried away, making it possible for more corrosion to occur, greatly accelerating the rate of decay.
This is in pointed contrast to the much more static conditions in the upper case of a channel furnace where circulation is sporadic because of convection, charging/tapping and the like. The aggregate of these conditions, and others previously described, provides a sound basis for using refractories of the highest quality in the inductor, throat and crucible areas. If an inductor, for example, is expected to tolerate these conditions for a year or more without maintenance, it must be designed carefully, taking into account these conditions, and be manufactured precisely to design specifications.
Broadly speaking, erosion in induction furnaces can be broken down into three major categories:
* Eutectic fluxing
* Bloating/spalling reactions
Erosion By Fluxing
Flux erosion is a familiar problem often blamed for deterioration of a refractory even when it is not at fault. In treating the subject, the term "eutectic fluxing" is used to cover a range of solid/solid and solid/liquid reactions. Though perhaps not properly named, it is useful here to get the concept across as concisely as possible.
The concept of fluxing can best be explained with a few examples. If a coreless induction furnace were lined with pure silica and nothing else, it would be necessary to raise the temperature of that aggregate to the point of self-bonding, which would be very close to the melting point of the cast iron in the crucible. By adding small amounts of boric acid or boron oxide, the melting point of the silica is lowered in the fine fraction, creating a borosilicate glass that cements the lining together. This is a fluxing reaction used for bonding.
Now, if that same silica lining were to be exposed to cast iron with a high manganese content that resulted in the formation of manganese oxide, that oxide, circulated in suspension over the surface of the silica lining, would convert some of the silica refractory to a manganese silicate glass. The result would be erosion by a fluxing reaction. The presence of other oxides in suspension will often make three-, four- and five-way combinations that become almost impossible to track because of the complex dynamics. Generally, the byproducts of these reactions (including both stable compounds and glasses resulting from melting, alloying and refractory wear) are lumped into the general term "slag."
Slag composition is worthy of attention--particularly in operations involving a range of alloy chemistfies--because it can be more difficult to predict than the weather. Another example for consideration is a channel furnace inductor lined with a magnesia refractory melting cast iron. Magnesia has a very high melting point in the pure state (over 4000F). Practically speaking, considering impurities and bonding agents that flux the melting point down, magnesia is really on the order of 3700F, still well above the normal operating temperature of any industrial channel furnace.
The melting point of the magnesia refractory drops further to less than 3000F in the presence of silica that can easily be circulated through the channel as a derivative of uncleaned returns. If the channel furnace is melting ductile iron pouring at 2850F, the channel, even without clogging, can be expected to run at 2950F or higher, presenting the problem of erosion when silica is introduced. Molding sand is not the only silica source. Silicon in the alloy itself will also oxidize during meltdown by taking oxygen from other sources such as rusty scrap.
A small particle of silica, carried through the channel by the inductor's circulation, touches the magnesia refractory and forms a glass whose melting point is lower than the temperature of the channel. The glass is swept away by metal circulation to rise up into slag or perhaps coat the refractory wall of the upper case and form a shelf. This same general sequence of events can be expected for copper-base alloys, light metals, steel and stainless steel.
Control of Fluxing
Control of fluxing reactions is accomplished by proper refractory selection (i.e., by matching the best acid, basic or neutral refractory with the dominant chemistry of the melt). The word "dominant" is not used lightly here because many alloys change in the course of the melting process. Fluxing control might require a refractory with a chemistry that will offset the most serious corrosion threat.
The other way of controlling fluxing is to control the corrosive oxides that are in the alloy, either by reducing them (if practical), by modifying their chemical reactivity or by adding an alternative reactant. An example would involve the erosion of a silica-bonded alumina refractory by slag resulting from rusty scrap. Adding aluminum-coated or aluminum killed steel can provide an iron oxide reactant that will help to preserve the refractory lining. it is always necessary, however, to consider other ramifications of such a procedure that could easily lead to buildup or clogging from aluminum oxide residuals.
The term "bloating" is more descriptive than scientific. It pertains to chemical reactions between penetrating slag components and the refractory that results in the formation of compounds occupying greater volume. The result is, as the name implies, an internal bloating of the refractory that, being a brittle material, responds by cracking and spalling.
An example would be an aluminum oxide lining in a coreless induction furnace being used to melt low-carbon steel in order to avoid the thermal shock cracking associated with a magnesia lining. The result is excessive amounts of iron and manganese oxide, both chemically basic.
Manganese oxide will react with silica from the refractory, molding sand or the alloy. to form a glass that readily wets and enters the refractory surface. Once in contact with the alumina structure, the manganese silicate glass becomes unstable. The manganese also reacts with the alumina to form a manganese alumina spinel requiring considerably more volume than the alumina itself. The forces generated are more than can be contained within a brittle structure, and cracking occurs.
Cracking opens the structure to the penetration of more manganese silicate glass and causes rapid erosion. A section through such a lining will graphically show the progression of the glassy phase, development of the spinel and the resulting internal cracking. This is an example of the coexistence of fluxing of the silica bond in the alumina lining and the spalling reaction that results from the spinel formation. Together, they produce rapid deterioration of the refractory.
Many other bloating-type reactions are found in both ferrous and nonferrous melting (e.g., magnesia/alumina spinel, zinc/alumina spinel, mullite). Since the thickness of induction furnace refractories is kept to a minimum in powered areas and because this type of reaction will readily damage very high-purity refractories, a serious concern is posed for induction furnace users. The speed with which cracking, spalling and finning can develop, given the right circumstances of reactive materials and circulation, should not be underestimated.
One of the chemical anomalies in metallurgy is that, as temperature increases, the bond strength of most oxides decreases because carbon's affinity for oxygen becomes stronger. Certainly, from the point of view of protecting alloys from oxidation, this is fortunate. However, when temperatures become extreme and, in some cases even at normal processing temperatures, the ability of carbon to break down refractory oxide compounds becomes a major cause of erosion.
The two principal oxides that fall into this category are silica (SiO[.sub.2]) and chrome oxide (Cr[.sub.2]O[.sub.3]). In coreless induction melting of cast iron and some nonferrous applications where silica is also used, carbon, either present in the alloy as an element in solution or as a cover, will break down silica in the lining.
The best example is one in which carbon's affinity for oxygen results in an equilibrium at the low end of the normal pouring temperature. This means that, above the equilibrium, carbon's affinity for oxygen will result in the breakdown of the silica lining once the more available oxygen from weak oxides in suspension (iron oxide) is used up.
Conversely, when the temperature falls below this equilibrium, oxygen is made available in the melt, and internal oxidation of silicon in solution occurs. To preserve the life of a silica lining, care must be taken to minimize time at high temperature, particularly after the bath has had an opportunity to refine out the other oxides in suspension or in solution. At that point, the silica lining is open to major deterioration by reduction. This is also a major cause of saturation in alumina-silicate-lined inductors by creating space for penetration.
Similarly, a chrome oxide-bearing refractory must also be protected from reduction. This can easily occur in a channel or coreless furnace left on hold at high temperature because the equilibrium point for carbon chrome is even lower than for carbon silicon. Once the available oxygen in the melt has been consumed by the carbon, the melt will take it from the chrome oxide. Dissociation of 6-12% of a refractory's bond structure, or up to 40% dissociation of the mortar, leads to rapid erosion and subsequent serious consequences.
Oxidation/reduction and equilibrium temperatures should not be viewed as pertaining only to one or two oxides. Only a few oxides are affected in the normal operating temperature ranges, but all refractory oxides are subject to reduction at high temperature and in the presence of carbon. Induction furnaces have that high temperature capability.
When a refractory with a highly developed saturation network is bridged and either has carbon in its chemistry or contains carbon deposits, temperatures can exceed 3000F and reduce silica, alumina and (under some circumstances) magnesia.
The reduction mechanism is not confined to carbon. Many of the light metals, particularly aluminum and magnesium, will take oxygen from other more weakly bonded refractory oxides such as silica. They will literally burn in a thermite reaction similar to an exothermic hot top or a burning bar. This is of major concern in melting aluminum and magnesium in high-powered induction furnaces.
This same mechanism can also contribute to erosion in iron or copper-base alloy melting. For example, iron melted using aluminum-coated steel scrap will attack the silica in the lining by both carbon and aluminum reduction. The excess aluminum in solution will become an oxide at the expense of the silica in the lining. The same general mechanism holds true for aluminum bronze, contributing not only to erosion, but also to the formation of buildup on the surface as the aluminum oxide deposits, creating superheated channels.
Saturation Network Failure
Metal network development takes place in virtually every powered area of an induction furnace, except where there is effective glassy phase sealing as in a silica-lined coreless furnace. When this network becomes too concentrated for the power applied, remelting and superheating within the refractory structure results. This leads to the formation of "worm holes" and/or "rat holes" that generally follow isotherms.
In channel furnaces, this can result in the development of a secondary channel parallel to the normal operating one. This can lead to serious problems if the secondary channel remains isolated and superheats. Often parallel channels tunnel through and become connected to the operating channel again. As a result, they receive sufficient cooling and pose no lining threat, but this is not a desirable condition because secondary channels contribute significantly to inefficient inductor operation. Controlling the saturation network density will correct this.
It is common for an inductor to become partially saturated during the early stages and also to develop sufficient plugging to superheat the network. This results in early washout of the penetrated refractory, and the furnace is forced to run most of its life as an "old" inductor. Controlling the initial clogging and saturation would have the inductor operating on the original channel surface and yielding much longer refractory life with more efficient performance. Occasionally the erosion pattern is in the form of a spiral that reflects the torsional circulation pattern in the channel.
Refractory melting in an induction furnace is abnormal but, unfortunately, not uncommon. It is not a normal erosion mechanism but, rather, is a consequence of buildup, clogging or bridging. Entrapped metal in a high-energy induction field can easily reach the failure temperature (melting point) of the refractory material. A clogged inductor can easily melt some of its own inductor refractory, which then circulates to the obstructed channel exit where it encounters cooler metal and solidifies. This seals the inductor and/or lower throat from the upper case. When this happens, a "melt through" is inevitable unless corrective action is taken immediately.
Bridging of the lower circulation pattern of a coreless furnace by midcoil buildup accomplishes the same thing. As the temperature builds, the coil may go into vapor block, initiating a runout. A pocket or fin of molten metal internal to the refractory wall can cause the same failure. This latter situation, called "pocket erosion," is normally encountered in high-frequency, self-tuning coreless furnaces. A fin or saturation zone receives concentrated electromagnetic flux, creating in the wall a bridged condition that is enlarged by refractory melting.
Personnel should be alert for hot spots or metal running from holes/cracks in the walls of coreless induction furnaces under power. In aluminum and other light metal melts, these hot spots may become incandescent as the aluminum and/or magnesium in the alloy dissociates the refractory through a thermite reaction.
Summary and Conclusions
It is evident that the deterioration of refractories in induction furnaces incorporates mechanisms that are unique to this type of equipment. In deciding how to handle a particular situation, it is important to keep in mind the interrelationships that exist between the equipment, the refractory, the alloy system and the slag environment. Often, a problem with one can be most economically and effectively solved by changes in another.
For 30 years, induction furnaces have never ceased to present a material engineering challenge to me. Nevertheless, I would also like to emphasize that the problems described in this series should not be taken as an indictment of the induction furnace as a melting tool. I am convinced that it is by far the best melting system available. It is sophisticated, complex and presents serious material design challenges. Each induction furnace style must be fully understood by today's foundrymen so that they can derive the full benefits of these versatile melting devices.
Ronald A. Stark Refractory Consultant Hubbardston, Massachusetts
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||part 3|
|Author:||Stark, Ronald A.|
|Date:||Aug 1, 1991|
|Previous Article:||Sand reclamation cost-savings worksheet.|
|Next Article:||Spotting marketing opportunities.|