You are on page 1of 40

Inclusion Formation and Removal

The term "clean steel" is commonly used to describe steels that have low levels of the solute elements sulfur, phosphorus, nitrogen, oxygen and hydrogen; controlled levels of the residual elements copper, lead, zinc, nickel, chromium, bismuth, tin, antimony and magnesium; and, a low frequency of product defects that can be related to the presence of oxides created during the act of steelmaking, ladle metallurgy, casting and rolling. This last definition causes extreme problems to the steel manufacturer as the definition of "clean" is not absolute, is based upon the product formed from the casting and the in-service use or life of the product. In addition, the definition "clean" is comparative as each customer of a steel product has the ability to buy steel from around the world and compare the performance level of a given product based upon the supplier. In this system, as steel is a commodity, the best steel producer defines the level of quality that is expected by a customer and, as steel producers are continually striving to produce "cleaner" steels, the cleanliness standard desired by the customer is continuously changing as a function of time and technological improvements. The term "clean steel" is therefore continually variable depending upon the application and the competition between steel suppliers. Due to the variable nature of the term "clean steel" it is more accurate to talk about "high purity steels" as steels with low levels of solutes and "low residual steels" as steels with low levels of impurities that originate from scrap remelting and "clean steels" as those steels with a low frequency of product defects that can be related to the presence of oxides. For example, there are "high purity, low residual clean steels" such as ultra deep drawing steel sheets for automobiles which require ultra-low carbon contents (< 30 ppm), low nitrogen contents (< 30 ppm) and the absence of oxide inclusions with diameters greater than 100 microns; and "low residual clean steels", such as those used for drawn and ironed cans, which are a standard low carbon steel (1006), without particular high purity requirements, but are ultra clean with the requirement that oxide diameters must be less than 20 microns. In addition, in forging and bearing grades, there are "clean steels" that require strictly controlled inclusion size distributions. As measurement is a key in the determination of permissible inclusion content thin products which undergo significant drawing are very susceptible to variations in inclusion size distribution, as cracks or wire breaks can be easily counted. For example, in drawn and ironed cans, it is very common to count the number of cracked flanges per million formed cans and generally a performance of less than 10 ppm would be looked upon as an excellent performance. Another application which is amenable to measurement is bearing life. It is well known that total inclusion content (as measured by total oxygen content) has traditionally correlated with bearing life and decreased total oxygen contents (below 10 ppm) have lead to significant increases in bearing life and thus the drive to very low inclusion contents in bearing steels. In addition to total oxygen content, the total length of stringer inclusions after forging correlates well with bearing life and, at low total oxygen levels, efforts to reduce inclusion clustering leads to very long fatigue life for bearings. There is one constant in the world of high purity, low residual and clean steels and that is the continual drive to reduce solute and residual contents in all steels and to control the frequency and size distribution of the inclusions that are found in all steels. Thus this chapter will focus

upon an understanding of the fundamentals of the production of high purity, low residual clean steels, emphasizing the limits in current technology and the potential for the production of high purity, low residual steels with a low frequency of inclusions with average diameters less than 5 microns. Clean steels are steels with a low frequency of inclusions of average diameter less than 5 microns. The major problems in clean steel manufacture are incomplete separation of clustered solid inclusions (> 5 microns in diameter), the presence of sporadic larger liquid inclusions due to emulsification of covering slags and the presence of solid materials that originate from the refractories used to contain steels. The equipment used to produce clean steel varies greatly between different steel plants; however, current clean steelmaking and casting practices are based upon the following principles: Oxygen, which is dissolved in liquid steel at the steelmaking and melting stage, must be transformed into a solid or a gas and removed before casting.

The external oxygen sources which are responsible for the reoxidation of liquid steel, must be eliminated at every step in the process.

The physical entrapment of the liquid fluxes used during steel refining and casting must be eliminated.

Refractories in contact with liquid steel must be chemically stable and resistant to corrosion and erosion.

These practical principles of clean steel manufacture are based upon an understanding of the importance of maintaining chemical equilibrium between the elements dissolved in liquid steel and the slag and refractory systems which are in contact with the liquid steel, and, of controlling fluid flow to avoid conditions at liquid slag-steel interfaces which could result in the physical entrapment of the covering slag. Clean steel manufacture is dependent upon an understanding of the fundamental steps necessary to produce a clean steel:

generation of the inclusion; transport of the inclusion to an interface; separation of the inclusion to the interface; and, removal of the inclusion from the interface.

Success or failure in the production of clean steel is dependent upon a complete understanding and execution of these four issues. 1. Inclusion Generation The formation of inclusions during steelmaking is inevitable as oxygen is more soluble in liquid iron than in solid iron. In addition, at 1600 C and gas phase oxygen partial pressures

greater than 6 x 10-9 atmospheres, liquid iron will spontaneously oxidize to liquid iron oxide. The saturation limit for oxygen in liquid iron in contact with iron oxide at 1600 C is 0.23 wt%, a value which decreases with temperature according to the equilibrium between iron, oxygen and iron oxide[1-2] The soluble oxygen content of pure liquid iron in contact with a gas phase can be reduced if the gas phase oxygen partial pressure is less than 6 x 10-9 atm: a partial pressure which can be achieved with carbon monoxide/carbon dioxide mixtures with less than 1% carbon dioxide, for example. Practically, it is necessary to shield liquid iron at all times to prevent the formation of iron oxide and generally this is accomplished by use of a liquid slag, with a low diffusivity of oxygen, as a physical barrier between the ambient atmosphere and the liquid steel. The solubility limit for oxygen in solid iron is low and the partial pressure of oxygen in equilibrium with solid iron and iron oxide at 1000 C is 1 x10-15 atm.[3]. Thus, if a pure liquid iron sample in equilibrium with FeO is solidified and cooled, iron oxide would precipitate interdendritically as spherical liquid iron oxide inclusions. The key to making a clean steel is to determine the mechanism by which an oxide inclusion can be removed from liquid steel. For example, if it possible to transport the inclusions from the bulk of the liquid steel to a free surface, where the inclusions can separate from the liquid steel, the steel will become cleaner. Of course, the only inclusions which can be removed are those which have formed while the liquid iron is molten. Pure iron solidifies at 1534 C where the maximum solubility of oxygen is approximately 0.17 wt/o. Thus, between 1600 C and 1534 C, 0.06 w/o soluble oxygen will transform to iron oxide. This is maximum possible amount of oxygen which could removed under these conditions and, regardless of one's processing abilities, it would be impossible to produce pure solid iron with less than 1700 ppm oxygen in the solid state, if liquid iron was allowed to become saturated with oxygen before solidification. To produce liquid iron of lower than saturation oxygen content it is necessary to precipitate the oxygen, as an oxide, at liquid steel processing temperatures. Again, if liquid iron is used as an example, the equilibrium between liquid iron, oxygen dissolved in liquid iron and liquid iron oxide can be written as follows: [Fe] + [O] = (FeO) ... [1]

where [O] is oxygen dissolved in liquid iron and (FeO) is pure liquid iron oxide. The equilibrium constant K1 = 1/ao for pure liquid iron and iron oxide and the oxygen level at saturation is a function of temperature only. The equilibrium oxygen content at a given temperature in pure liquid iron can be reduced by modification of the activity of FeO by dissolution into another chemical species. For example, liquid calcium aluminate could be added to the liquid iron surface as a sink for the liquid iron oxide. As the iron oxide dissolved in calcium aluminate, the mole fraction of iron oxide present in the newly formed slag would be reduced and the equilibrium oxygen activity of the steel would also be reduced as the liquid iron attained equilibrium with the iron oxide activity in the slag. Thus, modification of the activity of the deoxidation products the first method that can be used to reduce dissolved oxygen levels during processing.

A second method to reduce dissolved oxygen levels is to modify chemical equilibrium by changing the equilibrium reaction. In steelmaking, carbon is always present, and the equation [2] often sets dissolved oxygen activities. [C] + [O] = CO ...[2] In the BOF, for example, carbon content sets the lower limit for the oxygen activity at a given temperature and as the CO pressure approximates to 1 atmosphere during refining, lower carbon contents lead to higher oxygen activities and higher soluble oxygen levels. Oxygen content in carbon deoxidized steels can be further modified by vacuum or inert gas treatment which will reduce the equilibrium carbon monoxide pressure and lead to the production of steels with lower oxygen contents at a given carbon level. The equilibrium constant for reaction [2] increases with decreasing temperature and large quantities of CO are evolved during solidification of steels containing only carbon as a deoxidant. This "rimming" phenomena was utilized in ingot casting to produce a cleaner ingot surface; however, the porosity associated with even mild rimming is undesirable in many castings and the rimming action can be controlled by addition of deoxidants to "kill" the action in the steel during solidification. Elements such as silicon, manganese and aluminum are routinely added to liquid steel to reduce oxygen activities to below that which will cause CO evolution in a mold. The practical limit of steel cleanliness in normal steels is set by chemical equilibrium and, to reach this value, all inclusions formed during the deoxidation reaction must be separated before solidification[4] . In general, in steels deoxidized with silicon and manganese, the equilibrium is set by the formation of a liquid manganese silicate; in aluminum killed steels, by the formation of solid alumina; and, in calcium treated, aluminum killed steels by the formation of a liquid calcium aluminate. To further reduce soluble oxygen content at a given temperature, manganese silicate or alumina can be dissolved into a slag. For example, dissolution of alumina into a slag can result in the soluble oxygen content of an aluminum killed steel being reduced to less than 2 ppm. Dissolution of silica into a slag of calcium aluminate, for example, can significantly reduce the activity of silica and lead to soluble oxygen contents of silicon killed steels in the range of aluminum killed steels. The use of chemical equilibrium to reach low soluble oxygen contents is widespread; however, to be effective, all phases in contact with the liquid steel must be at the same equilibrium condition. If aluminum killed steels are at equilibrium, reaction [1] must also be at equilibrium; therefore, FeO activity in slags must be very low or aluminum will oxidize by reaction with the slag. Similarly equilibrium MnO and SiO2 slag activities in the slag are uniquely set by the addition of aluminum and any increases above the equilibrium slag levels will result in reaction with the slag and the formation of alumina. Therefore, in practice, the slag systems must be designed to be very low in FeO and MnO after the steel has been killed. Currently, furnace slags tend to be high in FeO and MnO and this has led to the development of slag-free tapping techniques and, also, slag killing using calcium carbide or aluminum. Equilibrium between the refractories and the steel must also be considered. For example, high alumina refractories are not 100% alumina and numerous binders are added which can contain oxides which are less stable than alumina. Under these conditions the container refractories will react with the liquid steel, if not adequately designed. Future clean steel

developments will be aimed at developing steels which are in equilibrium with both their container refractories and the slag present on top of the liquid steel. Recently, residual magnesium from ferro-alloys, scrap or recycled aluminum or from reduction of refractories has caused significant processing problems due to low residual levels (< 5ppm) causing the primary deoxidation inclusion to be transformed into a high magnesia containing inclusion which often is a magnesium aluminate spinel, a highly refractory inclusion. Often, this causes nozzle clogging during steel pouring. Recent studies by Itoh et al.[5] have fully documented the problem and magnesium residuals must be reduced to less than 1 ppm. It is interesting that originally the term slag denoted "any non-metallic substance, formed together with a metal during a metallurgical process" [6]; however, this very general definition has been refined with time to refer only to the covering liquid oxides that are present on top of the liquid steel, i. e., ladle slags, tundish slags and mold slags. The primary and secondary deoxidation or reoxidation product are not referred to as slag inclusions but as simply non-metallic inclusions or inclusions. A second source of cleanliness problems is incomplete separation of emulsified slags[7]. This occurs during processing events where there is a high dissipation of energy at the slag-metal interface. Events such as vessel filling, vessel drainage and level fluctuations at the slag metal interface are generally responsible for the generation of such inclusions. The size range of entrapped slag inclusions varies from less than 20 to greater than 200 microns; however, their frequency can be low. Generally, these type of inclusions are practice specific and lead to cleanliness problems which are apparently random, until their source is recognized[8-13]. The last source of inclusions is the erosion and corrosion of refractories and carry-over of the nozzle well sand during processing. These problems are solved through better process design. The final level of cleanliness is a balance between the rate of inclusion removal and the rate of inclusion formation. Clearly to refine liquid steel the rate of removal must be greater than the rate of creation; therefore, reductions in the rate of reaction between the steel and the slag and the steel and the refractory system will enable the lowest possible inclusion contents in the shortest processing time. Reoxidation of liquid steel by reaction with air causes very high rates of inclusion formation and, often, causes the rate of formation of inclusions to be greater than the rate of inclusion removal. Thus, a natural limit to any process which enables inclusion removal is the point at which air reoxidation begins.

2. Transport to an Interface In the above section on generation of inclusions it must be recognized that although thermodynamics can allow the prediction of the chemistry of the most stable inclusion, it does not allow a prediction of the size of an inclusion and, when inclusion removal is a goal, the inclusion size distribution is a key to understanding the rate at which an inclusion can be transported to a surface where it has the opportunity to separate from liquid steel. Deoxidation and reoxidation inclusions must first be nucleated within the liquid steel and once nucleated these inclusions will grow until the rate of growth is limited by diffusion in the liquid. The smallest measured deoxidation inclusions are of the order of 15

nanometers[13,14] however, these quickly grow to between 1 and 5 microns in diameter in practical steelmaking conditions. Liquid turbulence can aid in agglomeration of inclusions and solid inclusions can easily cluster and form large three dimensional rafts of sintered small inclusions. Thus it is not unusual to find a high frequency of small inclusions less than 5 microns in diameter and a much smaller frequency of inclusion rafts which are from 5 to 200 microns in diameter in liquid steels. The goal of clean steel manufacture is to minimize the total numbers of inclusions with diameters less than 5 microns and to eliminate all clustered inclusions before casting. Initial studies of inclusion removal were focused upon the removal of inclusions due to buoyancy driven flow where Stokes' Law applies[15-16]:

where VT is the sphere velocity, (( is the differential density between the particle and the liquid, ( is the liquid dynamic viscosity, g is acceleration due to gravity and r is the particle diameter. Stokes law describes the velocity of a solid sphere under the influence of buoyancy forces, due to density differences, in a static bath. Stokes Law only holds for rigid, spherical particles within the viscous flow regime, where the Reynolds number is less than 0.1. For spherical alumina particles in liquid steel the minimum particle radius for adherence to Stokes Law is 33.2 microns. For flows where the Reynolds number is greater than 0.1, it is common to view the problem as a balance between the gravity force and the combination of a buoyancy and a frictional force. Details of friction factors for different flow conditions and particle types are given by Schwerdtfeger[15] and corrections for slip, interfacial tension, particle rigidity and wall effects are given by Iyengar et al.[16]. Unfortunately, actual metallurgical vessels are not quiescent and are subject to natural convection, as has been pointed out by Szekely[17] and Iyengar[16] and inclusion removal rates in industrial vessels cannot be explained solely by Stokes' Law and its modifications. For example, Iyengar calculated that the critical temperature gradient to avoid natural convection, for steel held in a small crucible, was 0.08 C/cm and that fluid velocities of 0.015 cm/s were sufficient to ensure that the drag force due to the fluid flow was an order of magnitude larger than the buoyancy force. In actual steelmaking practice, temperature gradients are high against the refractory container walls and natural convection occurs leading to natural stirring patterns within the ladle or the tundish. Liquid steel velocity calculations, by Joo et al.[18] indicate fluid velocities which are greater than 1 cm/sec due to natural convection in ladles and tundishes. The effect of fluid flow in practical vessels is to provide a transport path to an interface and the problem can be viewed as a standard boundary layer problem where fluid flow leads to a momentum boundary layer which has a thickness which can be related to the flow field and the physical properties of the fluid. Inclusions are carried by fluid motion to a boundary layer at an interface where, due to buoyancy forces, they may traverse this boundary layer. In order to calculate, from first principles, inclusion removal in a stirred system, it is necessary to calculate the exact form of the boundary layer and then calculate the fraction of inclusions

which can traverse the boundary layer and separate during their contact time with the boundary layer. The Navier Stokes equations must be solved for the coupled heat and mass transfer conditions to take into account natural convection, to determine the boundary layer thickness against the container walls and the slag-metal surface (taking into account the fact that this is a liquid-liquid boundary). The local flow conditions for each particle at the boundary layer and the frequency of separation must then be determined as a function of particle size. Calculations of this type are quite numerically complex; however, results have been given by Joo et al.[18] where large inclusions with a large rising velocity separate more completely within the tundish and the efficiency with which inclusions are removed decreases with decreases in inclusion size. Joo's work indicated that even large inclusions can pass through a tundish and not be removed due to the effect of fluid flow and that particles with diameters less than 40 microns ( float out velocities of 0.5 mm/s) have removal efficiencies of less than 50%. Inclusion removal in ladles has not, as yet, been computer modeled for inclusion removal in the same manner as tundishes; however, the basic phenomenon of liquid phase mass transfer is well recognized. Turkdogan[19,20] suggested an equation of the following type is an adequate representation of practical results: Ct = Co exp ( - kt ) .... [3]

where Co is the initial inclusion content, Ct is the concentration of inclusions at a time t and k is the apparent flotation rate constant for a given type and intensity of stirring. Turkdogan models inclusion removal is a first order process where the removal rate is proportional to concentration. Schwerdtfeger[21] has also analyzed plant results of inclusion removal from aluminum-killed steels and found that a similar first order equation could be used to describe the data from total oxygen analysis: [O] = < [o]i - [o]e > exp [ - A keff t / V ] + [O]e .... [4]

where [O]i and [O]e are the initial and equilibrium oxygen concentrations, Keff is an effective mass transport coefficient related to surface area, A, and V is steel volume. Schwerdtfeger has shown that keff is a function of stirring intensity and that effective mass transfer coefficients from inclusion removal studies are similar to those seen in ladle desulfurization and aluminum oxidation. Stirring of any kind in a ladle will increase fluid velocities, decrease boundary layer thickness and promote improved efficiency of inclusion removal (as long as no other phenomena occur which promote inclusion generation). Mass transfer coefficients are often plotted as a function of energy input via stirring and correlations of inclusion removal trends can be developed in this manner; however, such an approach is only useful when small, closely sized particles (such as alumina) are being removed. Fortunately, this approach is appropriate for ladle metallurgy. Mixing in the liquid also has a strong effect on inclusion size and turbulence can result in more frequent contact between particles and inclusion agglomeration. The level of turbulence necessary to optimize inclusion agglomeration is currently an active field of research; however, as yet, no clear method of optimizing turbulence has been developed22,23].

Inclusion agglomeration also occurs on bubble surfaces and on refractory surfaces and small bubbles with float out velocities lower than that of the liquid steel recirculating velocities are excellent sites for inclusion agglomeration, as is seen in continuous cast product[22] and there have been many attempts to use fine bubble generation as a mechanism of inclusion removal[24]. Another important method of inclusion agglomeration is on refractory surfaces, the most obvious example being nozzle clogging. Agglomeration on refractories can be particularly troublesome as the clog is easily removed from the nozzle surface, freeing quite large agglomerated inclusions into the casting. Filtration of steel, an method of increasing inclusional separation to refractories, has been studied in detail for liquid steels over the last 10 years[25-28]; however, as yet, no long life filter or method of in-situ filter replacement has been developed. 3. Separation or Stabilization at an Interface The next step in clean steel manufacture is the separation of the inclusion from liquid steel to an interface. It can be shown from thermodynamics[29-32] that all inclusions have a lower energy when separated from the liquid steel to either a liquid steel-slag, liquid steel-gas or liquid steel refractory surface. Thus if the inclusion can separate to a surface it will be stabilized on the surface. In order that an inclusion separates to an interface the liquid between the inclusion and the interface must drain and then, a hole must spontaneously form and grow between the two interfaces in order for the particle to complete separation at the interface. The energy of hole formation is related to the interfacial energy between the interfaces and the liquid steel and the distance between the particle and the interface. Clearly energy must be supplied to create the hole which when created will spontaneously increase in size as the total interfacial area decreases with adsorption of the particle or droplet. This last step in inclusion transport can result in inclusions, which arrive at the interface but do not have sufficient energy to overcome the interfacial energy separating the two liquids, exhibiting a rest time phenomenon where the particles or droplets are stabilized for significant times before separation. In flowing systems, this can lead to droplets or particles which reach the surface moving across the interface due to the velocity gradients across the boundary layer and eventually being re-entrained. Thus in particle or droplets which exhibit rest time phenomena, the separation efficiency at an interface is less than unity and droplets and particles under this condition can be extremely difficult to remove completely from liquid steel. Thus it is important to increase particle or droplet size in liquid steels to the point that the buoyancy plus inertial force of the particle can overcome the interfacial forces to ensure complete separation at the interface. Thus fluid flow at the interface is also important in inclusion removal. Separation of inclusions to a refractory interface can lead to inclusion stabilization at an interface; however, it is possible for inclusions stabilized at refractories to be released back into the fluid during times of fluid turbulence. In filtration or during nozzle clogging the refractory interface acts as a accumulator and agglomerator of inclusions and the presence of large clusters of inclusions in cast product can often be related to disintegration of inclusion build-ups stabilized against refractories. Separation of inclusions to a bubble interface can be either helpful or deleterious to steel quality. The bubbles can cause inclusion stabilization and agglomeration on the bubble

surface. If the bubble buoyancy causes the bubble and its associated bubble raft to separate to a slag-metal interface then the inclusions can be completely removed from the steel; however, if the bubble size is too small top overcome the natural convection currents during process the bubble itself becomes an inclusion agglomeration site within the liquid and can also be responsible for quite large inclusion rafts that appear in castings. Thus separation to an interface without complete removal from the system, can be deleterious to steel quality. 4. Removal From an Interface Complete separation inclusions from liquid steel includes removal from the interface. There is one major technique that is used to completely remove inclusions and that is dissolution into a liquid slag. For liquid inclusions this step is not a problem as most liquid are fully miscible in other like liquid, i.e., liquid slag inclusions tend to be completely miscible in the covering slags found in the ladle tundish and mold. Solid inclusions; however, tend to have limited solubility in the covering slags and the production of clean steel, where the inclusions are solid, is dependent upon slag chemistry, mixing in the slag, slag temperature gradient and volume. To date most models of inclusion removal have assumed that inclusion transport to an interface is the rate determining step and that separation to the interface and removal from the interface are trivial; however, although inclusion separation to and emersion from the interface is thermodynamically favored transport into the slag phase is diffusion controlled and strongly dependent upon slag chemistry. Thus, in a manner similar to solute partition into a slag, inclusion separation can be controlled in a mixed mode where diffusion in the slag layer and the solubility of the dissolving particle may become important. If one views this process kinetically, ideally the flux of material away from the slag/metal interface should be greater than the flux of solid inclusions to the slag metal interface so that the interfacial concentration is less than the saturation limit for the slag. Once saturation concentrations in the slag are reached, inclusion will pile up at the interface, agglomerate and then dissolve at a rate determined by diffusion and the amount of mixing in the slag. Temperature gradients in the slag are also important as solubility decreases with temperature and the total capacity of the slag to dissolve inclusions and remain liquid is defined by slag chemistry and temperature. References 1. L. S. Darken and R. W. Gurry: Physical Chemistry of Metals, McGraw-Hill, 1953 2.. N. A. Gockcen: Trans AIME, 206, 1558 (1956) 3. A. Muan: Am. Ceram. Soc. Bull., 37, [2], 81, (1958) 4. R. J. Fruehan and E. T. Turkdogan: "Physical Chemistry of Iron and Steelmaking" , Making Shaping and Treating of Steel, USS, 1984. 5.. C. Itoh:" Thermodynamics of the formation of Magnesium Aluminate Spinel in Liquid Iron", PhD Thesis, Tohoku University, Japan 1996

6. C. Benedicks and Helge Lofquist: Non -Metallic Inclusions in Iron and Steel, John Wiley, 1931. 7. A. W. Cramb and I. Jimbo: "Interfacial Considerations in Continuous Casting", Iron and Steelmaking, Vol. 16, No.6, 1989, p 43 - 55. 8. M. Byrne, A. W. Cramb and T. W. Fenicle: "The Sources of Exogenous Inclusions in Continuous Cast, Aluminum Killed Steels," Iron and Steelmaker, June 1988, Vol. 15, p 41 50. 9. A. W. Cramb and M. Byrne: "Tundish Slag Entrainment at Bethlehem's Burns Harbor Slab Caster," Transactions of ISS, Vol 10, (1989), p 121 - 128. 10. M. Byrne and A. W. Cramb: "Operating Experiences with Large Tundishes", Transactions of ISS, Vol 10, (1989), p 91 - 100. 11. A. W. Cramb: "Directions in the Production of Clean Steels", Trans AFS, (1994) , p 3 - 9. 12. W. H. Emling, T. A. Waugman, S. L. Feldbauer and A. W. Cramb:"Subsurface Mold Slag Entrainment in Ultra Low Carbon Steels", Steelmaking Conference 1994, p 371 - 379. 13. K. Wasai and K. Mukai: J. Japan Ist. Metals, 52, (1988), p 1088 14. K. Wasai, A. Miyanaga and K. Mukai: " Observation of Non-Metallic Inclusion in Aluminum-deoxidized Iron Alloy", Proceedings of the Joint US-Japan Seminar - Clean Steel for the 21 st Century", ed. Y. Iguchi, Tohoku University, 1996, p 87-90. 15. K. Schwerdtfeger: "Rates of Movement of Solid Particles, Drops and Bubbles in Static Liquids", Kinetics of Metallurgical Processes in Steelmaking, Verlag StahlEisen, 1975, p 192 - 218. 16. R. K. Iyenger and W. O. Philbrook: "Motion of Droplets and Inclusions in Liquid Steel", ibid, p 219 - 233 17. J. Szekely and V. Stanek: Met. Trans., 1, 1970, p 119. 18. S. Joo, R. I. L. Guthrie and C. J. Dobson: "Modelling of Heat Transfer, Fluid Flow and Inclusion Transportation in Tundishes", Steelmaking Conference Procedings, 1989, p 401 408. 19. E. T. Turkdogan: "Ladle Deoxidation, Desulfurization and Inclusions in Steel - Part 1 : Fundamentals" Archiv fur Das Eisenhuttenwessen, 1983. No. 1., Vol 54, p 1-10. 20. E. T. Turkdogan: "Ladle Deoxidation, Desulfurization and Inclusions in Steel - Part 2 : Observations in Practice" Archiv fur das Eisenhuttenwessen, 1983, No. 2, Vol 54, p 45 - 52. 21. K. Schwerdtfeger: "Present State of Oxygen Control in Aluminum Deoxidized Steel", Archiv fur das Eisenhuttenwessen, 1983, No. 3, Vol 54, P 87 - 98. 22. S. Tanginuchi and A. Kikuchi: Tetsu-to-Hagane, 78, (1992), p 527

23. S. Tanginuchi: "Kinetics of Inclusion Agglomeration in Liquid Steel, Tetsu-toHagane,1996, p 81 - 111. 24. Y. Kikuchi, H. Matsuno, S. Maeda, M. Komatsu, M. Arai, K. Watanabe and H. Nakanishi: Proc. 8th Japan-Germany Seminar, (Oct 1993, Sendai, ISIJ, Tokyo), p 66. 25. D. Apelian and R. Mutharasan: J. Metals, 32, 1980, p 14-18. 26. D. Apelian, S. Luk, T. Piccone, R. Mutharasan: Steelmaking Conference Proceedings, Vol 69, 1986, p957-968. 27. P. F. Weiser, Steelmaking Conference Proceedings: ISS-AIME, Vol. 69, p969-976. 28. L. S. Aubrey, J. W. Brockmeyer and M. A. Mahaur: Steelmaking Conference Proceedings, ISS-AIME, Vol. 69, p977-991. 29. P. Kozakevitch and L. D. Lucas: Rev. Metallurg., 65, 1968, S. 589-98. 30. P. Kozakevitch and M. Olette: Revue de Metallurgie, October 1971, p 636 -646. 31. P. Kozakevitch and M. Olette: in "Production and Application of Clean Steels", 42, 1972, London, The Iron and Steel Institute. 32. P. V. Riboud and M. Olette: Proc. 7th ICVM, 1982, Tokyo, Japan, p 879 - 889.

Front Page Introduction Inclusion Stability

The Formation of Macro-Inclusions

There are four major methods of forming macro-inclusions:


Reoxidation Interaction between liquid steel and liquid slags o Vortexing o Ladle or mold filling o Argon Stirring o Pouring through a slag layer Erosion/corrosion during steel pouring Inclusion agglomeration due to clogging during steel pouring.

Reoxidation

Reoxidation is probably the most common cause of macroinclusion formation in casting [13]. To understand reoxidation one must understand that liquid iron is not stable in the presence of oxygen and that the spontaneous reaction that occurs leads to the formation of iron oxide. As deoxidizers are added the steel remains unstable in the presence of oxygen as a gas but now the inclusions which form include the oxides of the deoxidants. Some deoxidants, notably aluminum, magnesium and calcium, form very stable oxides which are more stable than some slag chemistries and some refractory chemistries. In this circumstance the steel reacts with the less stable oxides. The reoxidation can occur by reaction with:

The ambient atmosphere (air) The slag components less stable than the oxide of the deoxidant The refractories which are less stable than the oxides of the deoxidant

Source of Reoxidation

Inclusion Type

Comments

Contact with air

Inclusion chemistry can Due to the high be quite variable temperatures involved depending upon the mass flux of reoxidation in steel processing, and the residence time natural convection of the inclusion in the drives oxygen transport to liquid steel from the liquid steel after surrounding air reoxidation. It is not atmosphere. Mass flux unusual to see inclusions containing all of oxygen during reoxidation can be very oxidizable elements from the steel, e. g. iron high. oxide, manganese

oxide, alumina, silica etc..

In aluminum killed Contact with slags steels reaction with containing high levels slags generally causes of FeO, MnO and Silica alumina growth.

Although the reaction rates between slags and steels can be lower than that found during air reoxidation, during high levels of liquid steel turbulence, such as that found during stirring or pouring, reoxidation levels from this source can be quite high. In casting the water content of molds can be a significant issue, leading to high reaction rates

Reaction with Refractories

Although generally a low reaction rate, this can be a significant issue in the casting of ultra-clean steels.

Interaction between Liquid Steel and Liquid Slag

Emulsification of liquid slags or scums on the surface of liquid steels is a major source of macroinclusion formation. All of these types of defects are practice related and can be solved by practice changes. The issue in understanding emulsification is to understand the source of the energy that allows a buoyant droplet to become submerged. Generally this energy comes from the interaction of a flowing steel stream and a liquid slag. There are four major sources of this energy:

Open stream pouring onto or through a liquid slag (common during lip pouring) Filling a ladle or mold at too high a fill rate in the presence of slags or scums Vortexing during steel pouring from a ladle Stiring in the ladle with gas at too high a stir rate

Sankaranarayanan and Guthrie [4,5] studied vortexing during drainage in a water model of a ladle and reported that the initial rotational velocity at the surface of the vessel is extremely important in determining the height at which the vortex will form and that increased rotational velocities caused increased vortex initiation depth. Entrainment due to fluid flow at the interface has been examined by Noguchi et al. [6] who attempted to decrease the

entrainment of slag in low carbon Ti-Al killed steels. They noted that entrainment decreased as the casting speed was decreased. In a study that was conducted by Nakamura et al. [7], it was found that defects that contained mold slag increased in ultra-low carbon grades as the casting rate was increased. They also reported using as low argon flow rates as possible in their submerged entry nozzles to avoid entrainment. Manabu, et. al. [8] have also documented the existence of a critical gas flow rate for entrainment in both a silicon oil - water and a slag - steel system. These authors mention that the slag depth, slag properties, and gas bubble diameter play a role. The oil depth was found to be directly proportional to the flow rate needed to cause entrainment. The gas bubble size was found to be inversely proportional to the flow rate needed to cause entrainment. Manabu et al. investigated the effect of kinematic viscosities of the oils on emulsification and found though the kinematic viscosity was varied by a factor of 10, very little change was seen in the fluid velocity needed to cause entrainment. Harmon [9] documented the effect of interfacial tension and slag viscosity on emulsification phenomena. Details of the types of exogenous incusions that can be found in steels is given in references 10 and 11
Source Inclusion Type Comments

Tapping from furnace into vessel

A major problem during Small macroinclusions a tapping proctice that (20 - 100 microns) with allows slag from the a chemistry related to furnace to flow into the the ladle or the mold the ladle slag before the steel If fill rates are too high, shear at the liquid steel surface can lead to Larger macroinclusions emulsification of slags or ( greater than 40 scums on the surface of microns) with chemistry liquid steel. This can related to covering slag lead to very large or scum. macroinclusions if emulsification occurs close to solidification. Small macroinclusions Can be solved by gentle of ladle slag related stir rates (< 5 scfm) chemistry Large macroinclusion of chemistry related to Can be solved by bottom pouring flux or reducing the rate of rise reoxidation scum

Vessel Filling

Gas Stirring in Ladles

Filling of Bottom poured ingots

Ladle draining

Large macro-inclusions Can be solved by due to vertexing at end stopping pouring before of pour onset of vortex Generally form when there is reoxidation at Large macro-inclusions the same time as mold filling

Mold filling

Erosion/corrosion during steel pouring

This is a very common source of large defects which are typically solid and related to the materials of the mold itself. Some steel grades are quite corrosive (high manganese and grades that are barely killed and have high soluble oxygen contents) and attck the binder or the mold sand itself leading to large entraped sand particles. Reoxidation of steel leads to FeO based inclusions which are very reactive and wet the materials of the mold and lead to erosion of the mold in areas of high fluid turbulence. Of course sand which is not pressed, sintered or bonded in any way can easily be entrapped in turbulent fluid flow. Mold binders can also decompose at temperature and release mold particles that can be entrapped. Expansion due to the high thermal gradients associated with casting can also cause sand to loosen. Source
Loose Sand Sand

Defect

Comments
Always occurs in areas of high fluid turbulence for example in the gate or mold cavity

Reoxidation

Sand particles that may FeO in reoxidation have rims of products is very reoxidation corrosive. High liquid steel temperatures increase the corrosion rate in molds Certain steel grades react with refractories

High pouring temperatures

Sand

Steel Chemistry

Sand

Inclusion agglomeration due to clogging during steel pouring.

The formation of clogs when steels containing solid inclusions are cast can result in quite large macro- incluion defects if the clogs are released during teaming. All solid inclusions tend to agglomerate due to surface tension effects. Clogging of pourng nozzles can be the source of large macro-inclusion defects when steels are dirty and pouring times are long. A short discussion of nozle clogging follows: 1) The Early Days Snow and Shea (1949) in their study of erosion mechanisms reported the occurrence of corundum (Al2O3)-glass tufts covering the bore surface of nozzles used to teem fine-grained ingot steels treated with aluminum. Duderstadt et. al. (1967) found that nozzle blockage occurred with high levels of Al (0.0036%) and that nozzle sectioning revealed dendritic growth of alumina from the nozzle wall onto the bore. They also looked at ways to minimize heat loss in the nozzle by developing protruding nozzles and they also observed that a high preheat temperature is essential for a successful cast. Farrell and Hilty (1971) studied the effects of deoxidizing elements on the flow of liquid steel through nozzles. They observed that Al, Zr, Ti and the rare earths were all refractory at steelmaking temperatures and caused clogging. However, this was not the case with Si + Mn as a deoxidizer.

(2) Morphology of Alumina Ogibayashi et. al. (1992 and 1994) found two different forms of alumina in the case of an Alkilled steel. One consisting of chalky alumina which is a powdery white substance and the other consisting mainly of metal containing a large number of alumina clusters. The differences in morphology were attributed to steel cleanliness. Farrell et. al. (1970) found that deoxidation or indigenous inclusions form by precipitation and growth and there is an effect of cooling rate upon their size. The faster the cooling rate the smaller their size and the effect of very slow cooling rates is to increase the size of the deoxidation products until a limit is reached. In general reoxidation inclusions are larger than deoxidation types and tend to be richer in the weaker oxide-forming elements e.g. Si and Mn. Work carried out by Steinmetz and Lindberg (1977) in the laboratory and work carried out by Tiekink et. al. (1994) in industry found that Supersaturation of Al and O in steel leads to a high speed growth of alumina. High speed growth is characterized by dendrites and needles and occurs because the supply of reactants causes morphological instability in the growing dendrites. Hiraga et. al. (1995) looked at Ti added and Al-killed steel (Ti-AK) and found that the inclusions were coarse and granular which differed from the dendritic alumina in Al-killed steel. Ikemoto et. al. (1995) found an a-Al2O3 powdery buildup below the powder line of ZCG (ZrO2-CaO-C) nozzles. The buildup occurred on the reaction layer which forms on the hot face of the nozzle and is composed of two layers. The authors found that the buildup comprises of two layers with the first corresponding to a CaO-Al2O3 matrix and the second layer corresponding to Al2O3 aggregate. At the position at which buildup occurred, the portion of matrix in the

reacted zone was less than when there is no buildup. In addition , holes were formed in ZC coarse grains of the interior domain. This could be attributed to the extraction of CaO by diffusion out of the refractory and also from the coarse grains of the interior. The first layer of the buildup adhered onto Fe particles in the reacted zone of the surface of the nozzle and it contained a very high percentage of Si (4-10wt.%). This Si is considered to originate from the Si component in the nozzle. This Si as before could have most likely originated from the SiC from the refractory which oxidized to SiO2 and subsequently reacted further. (3) Mechanisms of Clogging Singh (1974) found that alumina was a light loose powder which was very friable and could be easily removed by the touch of a finger and proposed a boundary layer theory for the mechanism of clogging. The high contact angle of alumina in steel (134-146 degrees), will enable an inclusion to approach a refractory and attach itself in order to minimize contact with steel. Since the fluid velocity in the boundary layer (nearer the refractory wall) is zero, the alumina does not get washed away once it becomes attached. Impaction of inclusions from a turbulent flow can occur onto surfaces which are parallel to the mean flow because transverse eddy velocities acquired by a inclusion may project it onto the wall [Friedlander, 1957], [Wells, 1967]. The high temperatures of 1530C enable sintering of the alumina to occur. Dawson (1990) found that the accumulation of inclusions usually predominated at certain preferred locations within the nozzle and suggested the existence of flow separation as a cause. The flow within the separation zone is highly turbulent and the frequent flow reversals and stalls allow eddies to carry inclusions to the nozzle wall. Separation theory explains why loosely held friable powder is not washed away by the erosion forces of high velocity steel. Nakamura et. al. (1995) and Ichikawa et. al. (1994) established the effects of surface roughness on alumina refractories and found that the sample with coarse powder had greater quantities of alumina deposited in comparison to the fine powder sample Work carried out by Schwerdtfeger and Schrewe (1970) and Saxena et. al. (1978) suggested that the silica in the nozzle material reacts with the excess Al in the steel and forms alumina. Poirer et. al. (1995) found a deposit forming on alumina-graphite samples when low carbon steel was used (40-60ppm) and not when high carbon steel was used. They also found that clogging occurred with all types of C-refractories but it didn't with non C-refractories. The mechanism of formation of alumina was caused by reduction of the silica by the graphite in the nozzle to gaseous species which then transfers to the molten steel and further reacts to form alumina. Sasai and Mizukami (1994) suggested reduction of silica in the nozzle by graphite to gaseous species which reacts with Al in the steel to form alumina. They also suggested a similar mechanism for Ti-killed steel. Fukuda et. al. (1992) observed alumina deposition on the alumina graphite rods even when no Al was present in the iron melt. The suggested mechanism of formation of alumina was attributed to dissociation of the suboxide species (generated from reduction of silica by graphite) to the elemental form, which then reoxidizes to form alumina. Gao and Sorimachi (1993) suggested the formation of Ti-oxide as a reaction between titanium nitride and oxygen for inclusion formation in the case of Tikilled steel. Tsujino et. al. (1994) found deposition of alumina on zirconia-lime-graphite refractories as a consequence of reducing reactions by graphite. (4) Methods to reduce Nozzle Clogging Faulring et. al. (1980) found that calcium additions can prevent nozzle blockage and improve castability. However, the success of this is dependent on obtaining a 22% CaO in the

inclusion and also Ca can react with the sulfur present and form solid calcium sulfides. Larsen et. al. (1990) looked at calculating the Al and S concentration ranges in which CaS will form or oxide inclusion modifications will occur. They found that in order to achieve shape control to liquid calcium aluminates, the sulfur content must be low. Benson and Robinson (1993) looked at the addition of lime to the refractory in the form of lime stabilized zirconia. It was found that clogging was reduced in steel plants which only had marginal clogging and laboratory analysis of the refractories found that the transport of lime from the refractory to the refractory/steel interface was rate limiting. Akechi Ceramics (TYK Akechi Works) originally developed ZrO2-graphite based lining materials for the inner bore of immersion nozzles for effective clogging prevention. Ozeki et. al. (1987) found that clogging was reduced during field trials when they lined the inside of alumina graphite nozzles with zirconia-graphite. Nakamura et. al. (1991) found that by lining an alumina graphite nozzle with a material of composition CaO-ZrO2-SiO2-C, the alumina deposition in field trials were less than the conventional nozzles. Ichikawa et. al. (1991) found similar findings as above but they also looked at a ZrO2-BN-Si3N4 combination and found less alumina deposition. Kasai et. al. (1991) found from various compositions based on the ZG (ZrO2-C) with various combinations of CaO, MgO and SiO2, it was important to optimize the ratio of CaO or MgO in the refractory in order for a low melting point compound to be formed on the refractory surface. Benson and Robinson [1993] looked at lime-stabilized zirconia-graphite materials with addition of destabilizers to enable the extraction of lime. More recently Miyagawa et. al. (1995) found the decomposition of the CZ grain is accelerated by the presence of both Al2O3 and SiO2in CZC (CaO-ZrO2-C) materials. Work carried out by Miyagawa et. al. ( 1996) looked at the addition of b-Al2O3 on nozzle performance and also the composition Al2O3-C-SiC. Field trials were carried out and they observed that there was hardly any formation of a reactive layer for the b-Al2O3 in comparison to the a-Al2O3. Singh (1979) recognized the importance of fluid flow in respect to the existence of preferential sites of inclusion deposition and developed a annular nozzle which minimized the physical contact between liquid steel and the refractory surface. However, the application of the annular nozzle was restricted by operational concerns. More recently Tsukamoto et. al. (1991 & 1995) looked at annular step nozzles in more detail from Hydraulic water modeling tests. They observed that there was a decrease of stagnant area in molten steel, improvement in injected gas dispersion and breakage of vortex current at the outlet port. Argon injection techniques are used extensively in both tundish nozzles and submerged entry nozzles and shrouds since their introduction by Meadowcroft and Milbourne (1971). Unlike the chemical methods (nozzle material, calcium treatment) of inhibiting inclusion deposition, the effectiveness of argon injection is entirely independent of inclusion composition or steel temperature. This technique can therefore be used universally on a large variety of steel grades, however, there are problems with argon injection such as increased quality defects, bubble entrapment and nozzle cracking because of the high back pressure. Some improvements to the above problems were carried out by Kobayashi et. al. (1995) who improved the insert nozzle and the sliding plate to reduce clogging. Yokoyama et. al. (1996) looked at the permeability characteristics of the argon injection type nozzle on the steel quality by carrying out bubbling tests on submerged nozzles cut in half. The decreasing grain size decreases both the permeability and micro-pore size.

Duderstadt et. al. (1967) looked at designing nozzles which would minimize heat loss by having the upward zirconia nozzle protrude in order that the molten metal surrounding it would supply sufficient heat. This however, wasn't successful as clogging still occurred. Cramb mentioned that previous company practices were to insulate the upper nozzles which enabled longer casting times than when the nozzles weren't insulated. Preheating nozzles has been used to prevent the nozzles from acting as heat sinks during the beginning of a cast sequence. Louhin et. al. (1991) recommended that it is important to optimize the nozzle preheating cycle in order to avoid thermal shock and clogging caused by freezing. The use of ceramic fiber insulation has been also recently suggested by Nakamura et. al. (1995) who analytically modeled heat loss from various nozzles. They found that a large amount of heat loss can be decreased by the ceramic fiber slit. They also tried a nozzle with a void slit, but the effect was much less. Ogibayashi (1994) and Bannenberg (1995) have found that vacuum degassing of the ladle has found improved castability. It was found that 68 heats can be cast without Ar injection in the nozzle in the case of highly clean steel at 5.8 ppm oxygen (Ogiybayashi 1994). Recent work carried out by Wang et. al. (1996) have developed a new process for metal refining.

References (1) J. A. Horwath and G. M. Goodrich: "Micro-inclusion Classification in Steel Casting", AFS Transactions, 1995, p 495- 510 (2) J. M. Svoboda, R. W. Monroe, C. E. Bates and J. Griffin: "Appearance and Composition of Macro-inclusions in Steel Castings", AFS Transactions, 1987, p 187- 202 (3) C. R. Wanstall, J. Griffin and C. E. Bates: "Clean Steel Cast Technology", Steel Founders Society of America, Research Report No. 106. (4) R. Sankaranarayanan, R. Guthrie, "Slag Entrainment Through a 'Funnel' Vortex During Ladle Teeming Operations". Proceedings International Symposium on Developments in Ladle Steelmaking and Continuous Casting, CIM, Aug. 26-29, 1990, Hamilton, Ontario, p 66 - 87. (5) R. Sankaranarayanan, R. Guthrie, "A Laboratory Study of Slag Entrainment During the Emptying of Metallurgical Vessels". Steelmaking Conference Proceedings, 1992, Toronto, Ontario, p 655 - 664. (6) K. Noguchi et al: Zairyo to Purosesu (Current Advances in Materials and Processes), 1991, 4(4), p 1194-1197. (7) H. Nakamura, S. Kohira, J. Kubota, T. Kondo, M. Suzuki, Y. Shiratani, "Technology for Production of High Quality Slab at High Speed," 75th ISS Steelmaking Conference, Toronto, ON, 1992 (8) I. Manabu, S. Yutaka, O. Ryusuke, M. Zen-ichiro, "Evaluation of the Critical Gas Flow Rate Using Water Model for the Entrapment of Slag into a Metal Bath Subject to Gas Injection". Tetsu - To - Hagane, 1993, Vol. 79, No. 5, p 33

(9) J. M. Harman and A. W. Cramb: "A Study on the Effect of Fluid Physical Properties on Droplet Emulsification", Steelmaking Conference Proceedings, Vol. 79, 1996, p 773 - 784. (10) J. F. Farrell, P.J. Bilek, D. C. Hilty: "Inclusions Originating from Reoxidation of Steel", 28th Electric Furnace Conference Proceedings, Dec 9-11, 1970, AIME, 1971. (11) M. Byrne, A. W. Cramb and T. W. Fenicle: "The Sources of Exogenous Inclusions in Continuous Cast, Aluminum Killed Steels," Transactions of ISS, Vol 10, (1989), p 51 - 60. Clogging Bibliography

Bannenberg, N.; Seelmaking Conference Proceedings, (1995), Vol. 78, pp. 457-463. Benson, P.; Robinson, Q.; Steelmaking Conference Proceedings, (1993), Vol. 77, pp. 533-539. Dawson, S.; Ph.D. Thesis (1990), Toronto University. Dawson, S.; Steelmaking Conference Proceedings, (1990), Vol. 73, pp. 15-31. Duderstadt, G. C.; Iyengar, R. K.; Matesa, J. M.; AIME Proc. Electric Furnace Conference, (1967), pp. 61-66. Farrell, J. W.; Hilty, D. C.; AIME Proc. Electric Furnace Conference, (1971), pp. 31-46. Farrell, J. W.; Hilty, D. C.; Bilek, P. J.; AIME Proc. Electric Furnace Conference, (1970), pp. 64-88. Faulring, G.; Farrell, J.; Hilty, D.; I&SM, February, (1980), pp. 14-20. Friendlander, S.K.; Johnstone, H.F.; Industrial and Engineering Chemistry, (1957), Vol. 49, No. 7, pp. 1152. Fukuda, Y.; Ueshima, Y.; Mizoguchi, S.; ISIJ International, Vol. 32, (1992) N 1, pp. 164168. Gao, Y.; Sorimachi, K.; ISIJ International, Vol. 33, (1993) N 2, pp. 291-297. Hiraga, Y. et. al.; (1995), Taikabutsu Overseas, Vol. 15, No. 1, pp. 22-27. Ichikawa, K. et. al.; (1991), Taikabutsu Overseas, Vol. 11, No. 1, pp. 41-43. Ichikawa, K. et. al.; (1994), Taikabutsu Overseas, Vol. 14, No. 2, pp. 52-53. Ikemoto, T.; Sawano, K.; (1995), Taikabutsu Overseas, Vol. 15, No. 1, pp. 15-21. Kasai, N. et. al.; (1991), Taikabutsu Overseas, Vol. 11, No. 1, pp. 22-33.

Khalafalla, S.; Haass, L.; J. Am. Ceram. Soc., 55, (1972), p. 414. Klug, F.; Pasco, W.; Borom, M.; J. Am. Ceram. Soc., 65, (1982), p. 619 .Kobayashi, S. et. al.; (1995), Taikabutsu Overseas, Vol. 15, No. 1, pp. 33-37. Larsen, K.; Fruehan, R.J.; (1990), Steelmaking Conf. Proc., pp. 497-506. Loushin, K.M.; Suzuki, Y.; (1991), Steelmaking Conf. Proc., pp. 737-742. Meadowcroft, T.; Milbourne, R.; J. Metals, June, (1977), pp. 11-17. Miyagawa, N. et. al.; (1995), Taikabutsu Overseas, Vol. 15, No. 1, pp. 50-53. Miyagawa, N. et. al.; (1996), Taikabutsu Overseas, Vol. 16, No. 1, pp. 16-23. Nakamura, M. et. al.; (1995), Taikabutsu Overseas, Vol. 15, No. 1, pp. 28-32. Nakamura, N. et. al.; (1995), Taikabutsu Overseas, Vol. 15, No. 1, pp. 39-42. Nakamura, T. et. al.; (1991), Taikabutsu Overseas, Vol. 11, No. 1, pp. 38-40. Ogibayashi, S.; (Overseas edition), 46, 4, (1994), pp. 166-178. Ogibayashi, S.; Uchimura, M.; Maruki, Y.; Mizukoshi, D.; Tanizawa, K.; Seelmaking Conference Proceedings, 19992),Vol. 75, pp. 337-344. Okamoto, K.; Nakamura, T.; Kondo, M.; I&S Eng., December, (1982), pp. 47-52. Poirer, J.; Thillou, B.; Guiban, M.; Provost, G.; Steelmaking Conference Proceedi ngs, (1995), Vol. 78, pp. 451-456. Rackers, K. G.; Thomas, B. G.; Steelmaking Conference Proceedings, (1995), Vol. 78, pp. 723-734. Sasai, K.; Mizukami, Y.; ISIJ International, (1994), Vol. 34, N 10, pp. 802-809. Saxena, S. K.; Sandberg, H.; Waldenstrom, T.; Persson, A.; Steensen, S.; Scand. J. Metallurgy, (1978), 7, pp., 126-133. Schwerdtfeger, K.; Schrewe, H.; AIME Proc. Electric Furnace Conference, (1970), pp. 95-104. Singh, S. N.; Met. Trans, (1974), Vol. 5, pp. 2165-2178. Singh, S.; Steelmaking Conference Proceedings, (1979), Vol. 62, pp. 3-10.

Snow, R. B.; Shea, J. A.; J. Am. Ceram. Soc., (1949), Vol. 32, N 6, pp. 187-194. Steinmetz, E.; Lindenberg, H.; Morsdorf, W.; Hammerschmid, Arch. Eisenhuttenwes, (1977), 48, N 11, pp. 569-574. Tiekink, W.; Pieters, A.; Steelmaking Conference Proceedings, (1994), Vol. 77, pp. 423-427. Tsujino, R.; Tanaka, A.; Imamura, A.; Takahashi, D.; Mizoguchi, S.; ISIJ International, (1994), Vol. 34, N 11, pp. 853-858. Tsukamoto, N. et. al.; (1991), Steelmaking Conf. Proc., pp. 803-808. Tsukamoto, N. et. al.; (1995), Taikabutsu Overseas, Vol. 15, No. 1, pp. 43-49. Wang, L.; Lee, H.; Hayes, P.; ISIJ International, (1996), Vol. 36, N 1, pp. 17-24. Wells, A.C.; Chamberlin, A.C.; Brit. J. Appl. Phys., (1967), Vol. 18, pp. 1793-1799. Yamada, Y. et. al.; (1994), Taikabutsu Overseas, Vol. 14, No. 2, pp. 25-31. Yokoyama, Y. et. al.; (1996), Taikabutsu Overseas, Vol. 16, No. 1, pp. 24-29. Yoshida, T.; Koike, Y.; (1989), Taikabutsu Overseas, Vol. 9, No. 2, pp. 29-31. Zeze, M.; Private Communication, (1996). Front Page Introduction Inclusion Stability Microinclusion Formation Macroinclusion Formation Inclusion Atlas
o o

Micro-inclusions Macro-inclusions

Non-metallic inclusions
From Wikipedia, the free encyclopedia This article needs additional citations for verification. Please help improve this article by adding citations to reliable sources. Unsourced material may be challenged and removed.
(February 2011)

Non-metallic inclusions are chemical compounds and nonmetals that are present in steel and alloys. They are the product of chemical reactions, physical effects, and contamination that occurs during the melting and pouring process. These inclusions are categorized by origin as either endogenous or exogenous.[1] Endogenous inclusions, also known as indigenous, occur within the metal and are the result of chemical reactions. These products precipitate during cooling and are typically very small.[2] Exogenous inclusions are caused by the entrapment of nonmetals. Their size varies greatly and their source can include slag, dross, flux residues, and pieces of the mold.[2]

Contents
[hide]

1 Sources of inclusions formation 2 Classification of non-metallic inclusions 3 Influence of non-metallic inclusions to the properties of steel and alloys 4 See also 5 References

[edit] Sources of inclusions formation


Non-metallic inclusions arise because of many physical-chemical effects that occur in molten and consolidated metal during production. Non-metallic inclusions that arise because of different reactions during metal production are called natural or indigenous. They include oxides, sulfides, nitrides and phosphides. Apart from natural inclusions there are also parts of slag, refractories, material of a casting mould (the material the metal contacts during production) in the metal. Such non-metallic inclusions are called foreign, accidental or exogenous.

Most inclusions in the reduction smelting of metal formed because of admixture dissolubility decreasing during cooling and consolidation. The present-day level of steel production technology allows the elimination of most natural and foreign inclusions from the metal. However its general content in different steels can vary between wide limits and has a big influence on the metal properties.

[edit] Classification of non-metallic inclusions


Non-metallic inclusions, the presence of which defines purity of steel, are classified by chemical and mineralogical content, by stability and by origin. By chemical content nonmetallic inclusions are divided into the following groups:
1. oxides (simple FeO, MnO, Cr2O3, SiO2, Al2O3, TiO2 and others; compound FeOFe2O3, FeOAl2O3, FeOCr2O3, MgOAl2O3, 2FeOSiO2 and others; 2. sulfides (simple FeS, MnS, Al2S3, CaS, MgS, Zr2S3 and others; compound FeSFeO, MnSMnO and others); 3. nitrides (simple ZrN, TiN, AlN, CeN and others; compound Nb(C, N), V(c, N) and others), which can be found in alloyed steel and has strong nitride-generative elements in its content: titanium, aluminium, vanadium, cerium and others; 4. phosphides (Fe3P, Fe2P and others)

The majority of inclusions in metals are oxides and sulfides since the content of phosphorus is very small. Usually nitrides are present in special steels that contain an element with a high affinity to nitrogen. By mineralogical content, oxygen inclusions divide into the following main groups:

Free oxides FeO, MnO, Cr2O3, SiO2 (quartz), Al2O3 (corundum) and others Spinels compound oxides formed by bi and trivalent elements

Ferrites, chromites and aluminates are in this group.

silicates, which are present in steel like a glass formed with pure SiO2 or SiO2 with admixture of iron, manganese, chromium, aluminium and tungsten oxides and also crystalline silicates. Silicates are the biggest group among non-metallic inclusions. In liquid steel non-metallic inclusions are in solid or liquid condition. It depends on the melting temperature.

By stability, non-metallic inclusions are either stable or unstable. Unstable inclusions are those that dissolve in dilute acids (less than 10%concentration). Unstable inclusions are iron and manganese sulfides and also some free oxides. Present-day levels of steel production allow to move off from the metal different inclusions. However in general the content of inclusions in different steels varies within wide limits and has a big influence on the metal properties.

[edit] Influence of non-metallic inclusions to the properties of steel and alloys

Present-day methods of steel and alloy production are unable to attain completely pure metal without any non-metallic inclusions. Inclusions are present in any steel to a greater or lesser extent according to the mixture and conditions of production. Usually the amount of nonmetallic inclusions in steel is not higher than 0.1%. However, the number of inclusions in metal is very high because of their extremely small size. Non-metallic inclusions in steel are foreign substances. They disrupt the homogeneity of structure, so their influence on the mechanical and other properties can be considerable. During deformation, which occurs from flatting, forging, and stamping, non-metallic inclusions can cause cracks and fatigue failure in steel. When investigating the influence of non-metallic inclusions on the quality of steel, of great importance are the properties of these inclusions: size, shape, chemical and physical characteristics. All these properties depend on the chemical composition of steel, method of smelting and for certain steel grade. These properties can change within wide limits even within the same mode of production. Different methods for analysis of non-metallic inclusions have been developed and are now in use. These make it possible to determine content, structure and amount of non-metallic inclusions in steel and alloys with high accuracy.

[edit] See also


Aluminium alloy inclusions Entrainment defect Longitudinal facial crack

[edit] References
1. ^ Vander Voort, George F. (1984). Metallographic, Principles and Practice. McGraw-Hill. ISBN 0-07-066970-8. 2. ^ a b Beeley, P.R. (1972). Foundry Technology. The Butterworth Group. ISBN 0-408-70348-2.

Analysis of Non-Metallic Inclusions during Steel Manufacturing Using a PICA 1020 Inclusion Analyzer

Topics Covered
Introduction Monitoring Production of Ferrous and Non-Ferrous Alloys Analysis of Inclusions in Steels Classifying Inclusions Electron Beam Analyzers for Inclusion Characterization Routine Inclusion Analysis Improves Steel Quality Sample Preparation for Inclusion Analysis Analysis and Reporting of Inclusions during Steel Manufacturing Making Metallurgical Production Changes Following Inclusion Investigations

Introduction
One common phenomenon observed by steel manufacturers is the formation of non-metallic inclusions during the production process. Non metallic inclusions in steel are the cause for dangerous and serious material defects such as brittleness and a wide variety of crack formations. However, some of these inclusions can also have a beneficial effect on steel properties by nucleating acicular ferrite during the austenite to ferrite phase transformation especially in low carbon steels.

Monitoring Production of Ferrous and Non-Ferrous Alloys


By characterizing individual inclusions, determinations can be made as to the cleanliness of the steel and steel producers can consequently increase their throughput, better control their process, reduce their re-melts, and improve the quality of their product. In an effort to provide metal producing companies with a more direct means of monitoring their manufacturing process, Aspex and industry partners have developed methods which can routinely monitor the production of ferrous and non-ferrous products. In this TechNote, non-metallic inclusions were analyzed to illustrate the functionality of the Aspex PICA 1020 Inclusion Analyzer for

process monitoring and investigatory studies.

Analysis of Inclusions in Steels


The composition, and hence the properties, of the inclusions can be controlled through the chemistries of the metal and slag. By controlling the properties of the inclusions, higher quality steel can be made. However, despite the major advances in inclusion control, there has been no rapid and accurate method for determining the type, size and number of inclusions present in steel samples. Historically, a post-mortem analysis is performed on samples using manual SEM methods or conventional metallographic techniques long after the steel has been fully processed. Thus, the ability to make modifications to the process would typically require scientific studies that were intensive in-terms of personnel and cost. Also, analysis and interpretation of a limited number of inclusions using manual techniques may not be representative of the cleanliness or the type of inclusions present.

Classifying Inclusions
In general, inclusions are subdivided into macro- (>20 m), micro- (1-20 m), and submicroinclusions (<1 m). As a result, multiple analysis methods would have been conducted to examine the different inclusion size ranges. Figure 1 illustrates general methods currently in use for examining steel inclusions based on their applicable size ranges overlaid with a typical inclusion frequency distribution.

Figure 1. Comparison of applicable size ranges for various inclusion analysis tools. In addition, a typical frequency distribution is displayed.

Electron Beam Analyzers for Inclusion Characterization


Based on Figure 1, the electron beam analyzers provide a very broad dynamic range in both the size and number of small detectable inclusions. In addition to size and shape characteristics, an electron beam is particularly unique because it also provides the operator with an ability to determine the elemental composition of each individual inclusion, or feature. This is accomplished through the use of Aspexs Automated Feature Analysis algorithms (see TechNote#100). One of the real benefits of compositional data is the identification of solid and liquid inclusions at steel making temperatures.

Routine Inclusion Analysis Improves Steel Quality


Inclusion analysis has provided various avenues to explore the steel making process which, until now, has been fairly difficult to obtain enough data points to solve a problem in a timely manner. With the advent of Aspexs integrated electron beam equipment with NEW Silicon Drift Detectors, routine sample inspections can be perform in approximately 30-60 minutes allowing for process alterations to be made between heats, if necessary.

Sample Preparation for Inclusion Analysis


Metallographic preparation can be performed according methods given in ASTM standard E3 and E768. Typically, a sample of the molten metal is withdrawn (using standard samplers) from either the ladle or tundish and at different varying intervals, depending on the information required for process control. In addition, the final product can be monitored for quality control purposes. Prior to analysis, polishing of the surface must reveal the inclusions without interference from artefacts and foreign matter. Polishing must not alter the true appearance of the inclusions by excessive relief, pitting, and pull-out. Use of automated grinding and polishing equipment is recommended.

Analysis and Reporting of Inclusions during Steel Manufacturing


When a sample is prepared and polished for analysis, the sample is loaded directly into the Aspex analyzer. The area to be analyzed is then programmed into the instrument (between 50 and 200 mm2 area), and based on the inclusion types of interest, various analysis parameters can be chosen. (Please contact an Aspex representative for details concerning standard set-up parameters and alterations to standard protocols of the Aspex analyzer to optimize it for your needs.) Once the Aspex analyzer has completed its analysis, reports can be easily generated using AFA DataViewerTM, Tabular ReporterTM and Excel based software (See Figures 2 & 3).

Figure 2. Detected inclusions from a Tundish sample that resulted in clogging of the nozzle at the end of casting due the high Solid inclusion content. Modifications to the Calcium treatment were performed to alleviate the problem.

Figure 3. Size distribution versus the inclusion type (left) and ternary diagrams (right) aid in the evaluation of heats and determining the route cause of manufacturing problems and maintaining production control. Based on the location of the inclusions in a ternary diagram, the user can easily identify solid and liquid inclusion phases.

Making Metallurgical Production Changes Following Inclusion Investigations


Inclusion investigations allow for the acquisition of inclusion content, size, shape, and elemental composition data which can identify dominant chemical compounds in steel and permit appropriate metallurgical production changes. Differences in chemical composition, ladle treatment, casting conditions and differences in samples taken at different manufacturing levels can be shown statistically and provide operators with unprecedented insight into their process. Based on the general method described here, various applications of the technology have been successfully investigated in the steel industry (Table 1) and, as a result, quality assurance personnel and plant managers are incorporating this information directly into their Quality Management Systems. In the end, operators are given the capability to predict if there will be stable, eroding or clogging conditions during casting, and the ability to conduct routine root cause analysis investigations successfully. Table 1. Common investigatory topics currently being addressed through complete inclusion characterization. Topics Currently Being Addressed Nozzle clogging due to calcium aluminates or magnesium spinels Varying calcium additions to optimize castability Steel cleanliness versus casting time Comparison of inclusion content to standard analytical methods Hardware changes such as tundish linings and their overall effects Interstitial free degasser grade slag conditioning practice Tundish clogging and erosion minimization Degasser circulation and optimization Quality assessment steel cleanliness Slab billet or bloom disposition Alloy development

English Espaol Franais Deutsch Italiano Portugus Svenska Nederlands Polski esk Rom n Magyar Suomi Trke The Worlds Most Comprehensive METALS Database

Login Order now Home


Key Benefits Product Overview Resource Center Demo Contact How It Helps Unique Features Fact Sheet FAQ Articles News and Updates

Free Demo

Experience the full power of KEY to METALS for FREE

Contact us

Order Now

KEY to METALS Resource Center Articles Article

You want to be always up-to-date?

Apply to KEY to METALS eNews and receive fresh, leading-edge technical information and knowledge from the World's Most Comprehensive Metal Database.

Search Knowledge Base


Enter a phrase to search for:

Search by

Full text Keywords

Headings Abstracts

Documents download

Click on the links to download documents.


KEY to METALS Fact Sheet KEY to METALS Leaflet Extended Range Leaflet SmartComp Leaflet KEY to METALS Price List

Focus on Reliability

The reliability of KEY to METALS Database and services is ensured through the use of dedicated superstructure and ISO 27001:2005 certification.

Click here to learn more about unique KEY to METALS quality and reliability.

Clean Steel: Part Two


Abstract:
Non-metallic inclusions, which are undesirable components of all steels, play an important role with respect to their effect on the steel properties. Controlling inclusions in steel is closely connected with the concept of "clean steel". The improvement in steel properties by control of non-metallic inclusions plays an important part in defending the applications of steel against newer competitive materials.

Non-metallic inclusions, which are undesirable components of all steels, play an important role with respect to their effect on the steel properties. Controlling inclusions in steel is closely connected with the concept of clean steel. The improvement in steel properties by control of non-metallic inclusions plays an important part in defending the applications of steel against newer competitive materials. The aims of the metallurgist are to eliminate undesirable inclusions and control the nature and distribution of the remainder to optimize the properties of the final product.

Generally, non-metallic inclusions in steel normally have a negative contribution to the mechanical properties of steel, since they can init iate ductile and brittle facture. Among various types of nonmetallic inclusions, oxide and sulphide inclusions have been thought harmful for common steels.

All steels contain non-metallic inclusions to a greater or less extent. The type and appearance of these non-metallic inclusions depends on factors such as grade of steel, melting process, secondary metallurgy treatments and casting of steel. Because of this, it is of particular significance to determine how pure the steel is. The term steel cleanness is relative one, since even steel with only 1 ppm each of oxygen and sulfide will still contains 10 9 -1012 non-metallic inclusions per ton. From the viewpoint of cleanness all steels are dirty.

Non metallic inclusions in steel are the cause for dangerous and serious material defects such as brittleness and a vide variety of crack formations. However, some of these inclusions can also have a beneficial effect on steels properties by nucleating acicular ferrite during the austenite to ferrite phase transformation especially in low carbon steels. According to definition, the non-metallic inclusions are chemical compounds of metal with nonmetal which are present in steel and alloys like separated parts.

Classification of non-metallic inclusions


Non-metallic inclusions are divided by chemical and mineralogical content, by stableness/stability and origin. By chemical content non-metallic inclusions are divided into the following groups:

Oxides (simple: FeO, MnO, Cr2O3, TiO2, SiO2, Al2O3 etc.; compound: FeOFe2O3, FeOAl2O3, MgOAl2O3, FeOCr2O3 etc.) Sulphides (FeS, MnS, CaS, MgS, Al2S3 etc.; compound: FeSFeO, MnSMnO etc.) Nitrides (simple: TiN, AlN, ZrN, CeN etc.; compound: Nb(C,N), V(C,N) etc, which can be found in alloyed steels and has strong nitride-generative elements in its content: titanium, aluminum, vanadium, cerium etc.) Phosphides (Fe3P, Fe2P etc.)

The majority of inclusions in steels are oxides and sulphides. Among various types of nonmetallic inclusions, oxide and sulphide inclusions have been thought harmful for common steels. Usually, nitrides are present in special steels (stainless steels, tool steels) which have elements with a strong affinity for nitrogen (e.g. chrome, vanadium), which create nitrides.

Figure 1 shows sulfides and oxides of non metallic inclusion in steel.

Figure 1: Non-metallic inclusion in steel: oxides-dark gray and sulfides-light gray

By mineralogical content oxygen inclusions are divided into the following groups:

Free oxides FeO, MnO, Cr2O3, SiO2 (quartz), Al2O3 (corundum) etc.

Spinels-compound oxides which are formed by bi- and tri-valent elements as a ferrites, chromites and aluminates. Silicates which are presented in steel like a glass formed with pure SiO2 or SiO2 with admixture of iron, manganese, chromium, aluminum and tungsten oxides and also crystalline silicates.

Depending on the melting temperature, in liquid steel non-metallic inclusions are in solid or liquid condition.

As mentioned above the majority of inclusions in steels are oxides and sulfides. Sulfides in steel have been paid much attention because their treatment is an important problem in the steelmaking process. They affect on the properties of the final products by their deformation during the steel working process; especially their morphology has a significant effect on the steel properties.

According to analysis based on the steel ingots containing 0.01-0.15% S, the morphology of MnS can be classified into three types:

1) Type I is a globular .MnS with a wide range of sizes, and is often duplex with oxides. 2) Type II has a dendritic structure and is often called grain-boundary sulfide because it is distributed as chain-like formation or thin precipitates in primary ingot grain boundaries. 3) Type III is angular sulfide and always forms as monophase inclusion.

Most of the above mentioned sulfides are formed both during the process of secondary metallurgy or the solidification process. Recently, with the development of steelmaking technology, the sulfur concentration in steel was lowered drastically. Also, the continuous casting technology of steels with higher cooling rate than the ingot casting almost replaced the ingot casting.

So, the sulfides in the modern commercial steel are usually formed on solidification process or in solid steel during the subsequent cooling process. For example, the Widmansttten plate-like MnS2, is formed in solid steel and Figure 2 shows the common morphology of MnS in conventional continuously casting steel, including the globular duplex oxidesulfide (particle A, B and C) and the Widmansttten plate-like MnS (particle D).

Figure 2: Typical duplex oxidesulfide inclusion (particle A, B and C) and plate-like MnS (particle D) in conventional continuous casting silicon steel.

Numerous examples of the effect of non-metallic inclusions on steel properties show the importance of the behavior of the inclusions as well as of surrounding metal matrix during plastic working of steels. The aims of the metallurgist are to eliminate undesirable inclusions and control the nature and distribution of the remainder to optimize the properties of the final product.

An attempt by using program ABACUS was performed to model the behavior of slag inclusions and their surrounding matrix material during hot rolling and hot forging of hardenable steels. It is shown that it can be helpful for studying the behavior of inclusions, which is difficult or even impossible to obtain from a conventional experiment.

Figure 3 shows the effective strain contour during plastic deformation. Three regions of strain concentration (red) can be seen and a trihedral void (white region) close to the round inclusion is formed. The strain concentrations arise at the inner surface of the matrix. Another interested thing is that two edges of the pore tend to emerge and a bonding is formed. The difference in mechanical properties between the matrix and the inclusion is found to be the primary reason to create a void. The weak bonding at the interface between the matrix and the inclusion seems to facilitate to open the void.

Figure 4 shows the effect of rolling temperature on the relative plasticity index during hot rolling of steels. The relative plasticity index of inclusion increases while the rolling temperature rises. There exists a transition region, where the relative plasticity index changes rapidly. This trend agrees with the existing experimental results.

Figure 3: Void formation close to the inclusion.

Figure 4: Effects of rolling temperature on the relative plasticity index.

How it Helps | Unique Features | Fact Sheet | Articles | News and Updates | Terms of Use | Site map

Show pagesource Login

Main page About us Advertising Opportunities

Materials Engineering
Metals Ceramics Polymers Composites Fluids Materials Data Materials Links Contents

Collaboration
10 most popular articles Latest articles Create a new article Edit articles Contact us Terms of use Discuss the article and ask questions in our Materials Forum to Metals to Steel making

Non-metallic inclusions in steel


Dr. Dmitri Kopeliovich Non-metallic inclusions are chemical compounds of metals (Fe, Mn, Al, Si, Ca) with nonmetals (O, S, C, H, N). Non-metallic inclusions form separate phases. The non-metallic phases containing more than one compound (eg. different oxides, oxide+sulfide) are called complex non-metallic inclusions (spinels, silicates, oxysulfides, carbonitrides). Despite of small content of non-metallic inclusions in steel (0.01-0.02%) they exert significant effect on the steel properties such as:

Tensile strength Deformability (ductility) Toughness Fatigue strength corrosion resistance Weldability Polishability Machinability

Depending on the source, from which non-metallic inclusion are derived, they are subdivided into two groups: indigenous and exogenous inclusions. Indigenous inclusions are formed in liquid, solidified or solid steel as a result of chemical reactions (deoxidation, desulfurization) between the elments dissolved in steel. Exogenous inclusions are derived from external sources such as furnace refractories, ladle lining, mold materials etc. Amount of exogenous inclusions and their influence on the steel properties are insufficient.
Types of non-metallic inclusions

Oxides

FeO, Al2O3, SiO2, MnO, Cr2O3 etc. Al2O3*SiO2, Al2O3*FeO, Cr2O3*FeO, MgO*Al2O3, MnO*SiO2 etc.

Sulfides

FeS, MnS, CaS, MgS, Ce2S3 etc.

Oxysulfides

MnS*MnO, Al2O3*CaS, FeS*FeO etc.

Carbides

Fe3C, WC, Cr3C2, Mn3C, Fe3W3C etc.

Nitrides

TiN, AlN, VN, BN etc.

Carbonitrides

Titanium carbonitrides, vanadium carbonitrides, niobium carbonitrides etc.

Phosphides

Fe3P, Fe2P, Mn5P2


Formation of non-metallic inclusions

There are three stages of inclusions formation: 1. Nucleation At this stage nuclei of new phase are formed as a result of supersaturation of the solution (liquid or solid steel) with the solutes (eg. Al and O) due to dissolution of the additives (deoxidation or desulfurization reagents) or cooling down of the metal. The nucleation process is determined by surface tension on the boundary inclusion-liquid steel. The less the surface tension, the lower supersaturation is required for formation of the new phase nuclei. The nucleation process is much easier in the presence of other phase (other inclusions) in the melt. In this case the new phase formation is determined by the wetting angle between a nucleus and the substrate inclusion. Wetting condition (low wetting angle) are favorable for the new phase nucleation. 2. Growth Growth of a separate inclusion continues until the chemical equilibrium is achieved (no supersaturation). Growth of inclusions in solid steel is very slow process therefore a certain level of nonequilibrium supersaturation may be retained (eg. martensite). 3. Coalescence and agglomeration Motion of the molten steel due to thermal convection or forced stirring causes collisions of the inclusions, which may result in their coalescence (merging of liquid inclusions) or agglomeration (merging of solid inclusions). The coalescence/agglomeration process is driven by the energetic benefit obtained from decrease of the boundary surface between the inclusion and the liquid steel. Inclusions with higher surface energy have higher chance to merge when collide. Large inclusions float up faster than the smaller ones. The floating inclusions are absorbed by the slag. The floating process may be intensified by moderate stirring. Vigorous stirring will result in breaking the large inclusions into small droplets/fragments. Gas bubbles moving up through the molten steel also promote the inclusions floating and absorption by the slag. Blowing inert gas in Ladle Furnace (LF) or Vacuum ladle degassing of steel in result in obtaining cleaner steel.

Morphology of non-metallic inclusions

Globular shape of inclusions is preferable since their effect on the mechanical properties of steel is moderate. Spherical shape of globular inclusions is a result of their formation in liquid state at low content of aluminum. Examples of globular inclusions are manganese sulfides and oxysulfides formed during solidification in the spaces between the dendrite arms, iron aluminates and silicates. Platelet shaped inclusions. Steels deoxidized by aluminum contain manganese sulfides and oxysulfides in form of thin films (platelets) located along the steel grain boundaries. Such inclusions are formed as a result of eutectic transformation during solidification. Platelet shaped inclusions are most undesirable. They considerably weaken the grain boundaries and exert adverse effect on the mechanical properties particularly in hot state (hot shortness). Dendrite shaped inclusions. Excessive amount of strong deoxidizer (aluminum) results in formation of dendrite shaped oxide and sulfide inclusions (separate and aggregated). These inclusions have melting point higher than that of steel. Sharp edges and corners of the dendrite shaped inclusions may cause local concentration of internal stress, which considerably decrease of ductility, toughness and fatigue strength of the steel part. Polyhedral inclusions. Morphology of dendrite shaped inclusions may be improved by addition (after deep deoxidation by aluminum) of small amounts of rare earth (Ce,La) or alkaline earth (Ca, Mg) elements. Due to their more globular shape polyhedral inclusions exert less effect on the steel properties than dendrite shape inclusions.

Distribution of non-metallic inclusions

Besides of the shape of non-metallic inclusions their distribution throughout the steel grain structure is very important factor determining mechanical properties of the steel.

Homogeneous distribution of small inclusions is the most desirable type of distribution. In some steels microscopic carbides or nitrides homogeneously distributed in the steel are created by purpose in order to increase the steel strength. Location of inclusions along the grain boundaries is undesirable since this type of distribution weakens the metal. Clusters of inclusions are also unfavorable since they may result in local drop of mechanical properties such as toughness and fatigue strength.

Distribution of non-metallic inclusions may change as a result of metal forming (eg. Rolling). Ductile inclusions are deformed and elongated in the rolling direction. The less ductile inclusions are breaking forming chains of fragments. Elongated inclusions and chains result anisotropy of mechanical and other properties. Mechanical properties in transverse direction are lower than those parallel to the rolling direction. Discuss the article and ask questions in our Materials Forum

Related internal links


Steel making (introduction) Basic Oxygen Furnace (BOF) Electric Arc Furnace (EAF) Ladle refining Deoxidation of steel Desulfurization of steel Structure of killed steel ingot Macrosegregation in steel ingots Fabrication of large steel ingots Electroslag Remelting Steel strip processing Solidification Crystallization

Related external links

http://web.mst.edu/~karakus/Acers.March21%202002_files/frame.htm http://lib.tkk.fi/Diss/2004/isbn951227423X/isbn951227423X.pdf http://www.scielo.br/scielo.php?pid=S1516-14392004000400002&script=sci_arttext http://www.saimm.co.za/events/0401slags/downloads/105_Dekkers.pdf to Metals to Steel making


non-metallic_inclusions_in_steel.txt Last modified: 2009/02/04 by dmitri_kopeliovich Show pagesource Except where otherwise noted, this work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 License

You might also like