Consequences and - Semantic Scholar
Panel 13-1320.pdf, Page 1 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
No-Bake S-Containing Mold-DI Metal Interactions: Consequences and Potential Application J. Qing, S. Lekakh, V. Richards Missouri University of Science and Technology, Rolla, Missouri Copyright 2013 American Foundry Society ABSTRACT The ductile iron liquid metal is very reactive with environmental gases and solid molds after a nodulizing treatment. Mold-metal reactions could alter the spheroidal graphite shape to vermicular or flake. In this research, the metal-mold reactions were enhanced by various sulfur additions to the no-bake mold inserts to study these phenomena and to artificially produce layered dual-graphite structure in the casting. Thermodynamic and solidification modeling together with sets of different experimental techniques were applied to a test casting with internal core passages representing the cylinder head of a diesel engine. The effects of insert surface shape and location in the mold; content and type of sulfur-containing agent; coating; pouring temperature and additional chilling were experimentally studied and discussed. These results could be used to better understand the metal-mold interaction, prevent surface defects and design dual-graphite structure processes.
atmosphere around the liquid metal surface (Fig. 2a) or the complex composition surface slag moved from the bulk metal on its surface (Fig. 2b). These products were observed on a levitation re-melted ductile iron droplet.
Fig. 1. Possible interactions between Mg-treated liquid metal and the environment are diagramed.
Keywords: Mg-treated cast iron, metal-mold interaction, dual graphite structure, sulfur-modified insert INTERACTIONS OF Mg-TREATED CAST IRON WITH ENVIRONMENT A magnesium treated ductile iron liquid metal is highly reactive with its surrounding gases or the solid environment (mold). Various possible interactions between Mg-treated metals and the environments are classified in Fig. 1. The negative effects of these interactions can be Mg-fading and consequent spherical graphite shape degeneration in the bulk casting or shape degeneration in specific surface regions, as the results of metal-mold interactions. Additionally, the reaction products (insulated inclusions or aggregated into slag) could have negative effects on casting performance and machinability. A better understanding of these complicated phenomena is important for day-to-day foundry practice. While holding magnesium-treated ductile iron in the ladle or automatic pouring devices, two different processes could take place: (1) vaporization of magnesium having a high partial pressure and (2) adsorption and diffusion of oxygen into the liquid metal and reaction of oxygen with magnesium dissolved in the metal. These two processes give different types of reaction products. In the first instance, the MgO whiskers could grow into the
(a)
(b) Fig. 2. These are reaction products that are observed on the surface of the levitation re-melted ductile iron droplet: (a) MgO whiskers above the re-melted droplet surface and (b) Mg-Si-O, Mg-Si-Ca-O, and REM-Si-O (white) phases on the droplet surface.
Panel 13-1320.pdf, Page 2 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
Steel
Molten steel and ductile iron metal have different reaction sequences with metalloid-type impurities (oxygen or sulfur). For example, when refining molten steel, alkali and rare earth metals will first deeply deoxidize the liquid steel. Then, further additions produce desulfurization. However, alkali and rare earth metals in the liquid cast irons have different sequences of refining reactions. Some will react with oxygen first (Mg) and some with sulfur (Ca). Figure 3 schematically illustrates the reaction sequences of liquid steel (low carbon) and liquid cast iron (high carbon) with Mg, Ca or REM additions.1, 2 In the case of the reverse process, when a Mg-treated liquid metal reacts with the environment, dissolved Mg will be oxidized first, based on thermodynamic equilibrium. However, multiple reaction products have been observed on a levitation re-melted droplet surface (in air) (Fig. 2b) because all active elements (Mg, Si, Ca, REM) can be oxidized simultaneously in a surface layer containing oxygen.
[O]
[S]
[O]
[S]
Mg [S]
Cast iron
[S]
[O]
[S]
(a)
Ca
[O]
[O]
casting process). This layer, however, was formed after solidification. It did not exhibit a degenerate graphite shape (Fig. 6) when compared with that for conventional sand mold casting processes (green or no-bake).
[O]
(b)
Ce
Fig. 4. This green sand mold casting skin with multiple features includes: (a) de-carburized and graphite degradation layers; (b) an iron oxide subsurface film and (c) a liquid slag/mold reaction product, also incorporated into the subsurface.
Ca
Mg [S]
(c)
Ce
Additions, wt.%
Fig. 3. Sequences of the refining reactions in the 1, 2 liquid steel and the cast iron are illustrated.
Another important group of interactions took place when Mg-treated cast iron was poured into molds (Fig. 1). A casting skin was formed when the iron contacted the external mold environments during solidification. Defects in this skin include a degenerate graphite shape, high temperature reaction products, porosity and exogenous inclusions from a reaction of surface slag/dross with molding sand or even non-reacted sand particles. Figure 4 illustrates different types of features observed on the skin of the ductile iron casting, solidified in a green sand mold.3 No-bake molds provide better casting surface quality. However, even in this case, the casting surface had a degenerate graphite layer (see flake graphite in Fig. 5). Boonmee and Stefanescu studied the mechanism of formation for casting skin in Mg-treated irons with spherical and compacted graphite shapes.4, 5 It was shown that the casting skin can have a negative influence on the mechanical properties of thin wall castings.6 The surface of the cast bar had a limited oxidized/decarburized layer when the Mg-treated metal solidified in a water-cooled graphite mold (continuous
Fig. 5. This photomicrograph shows the degenerate graphite phase layer on the casting surface from a no-bake mold.
In the industrial case described previously, the Mg-treated metal, gas or solid environment interactions had negative effects. These effects included decreased machinability3 or decreased mechanical properties of unmachined thinwall castings with as cast surfaces.4, 5 In some cases, however, castings with an enlarged, layered dual graphite structure may have advantages when compared to castings with only flake or spheroidal graphite. A dual graphite structure is referred to as a structure that contains both spheroidal graphite and flake graphite in specific casting regions. One potential application of this dual graphite structure in iron casting is on a diesel engine head, which requires high strength and excellent thermal conductivity in different regions for a single casting. Our previous studies succeeded in producing dual graphite structure in the casting by the in-stream de-nodulizing treatment of a particular portion of the Mg-treated metal
Panel 13-1320.pdf, Page 3 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
poured into the mold.7, 8 In this paper, a localized in-situ de-nodulizing treatment was studied to produce dual graphite structure in the casting.
Table 2. Average Chemistry (wt %) of Experimental Ductile Irons in Ladle Heat C Si Mn Cu S P Ni Mg 1 2 3
3.60 2.0-2.7 0.31 3.70 2.5 0.30 3.67 2.6 0.34
0.6 0.004 0.01 0.02 0.043 0.6 0.004 0.01 0.03 0.046 0.6 0.003 0.01 0.03 0.049
The objectives, designs and tested variables of experimental heats are summarized in Table 3. No-bake molds made of F-60 silica sand were used in all of the heats. The total sulfur content in the no-bake sand was tested with the C-S determinator as 0.02%. The special S-modified core insert were also produced from no-bake F-60 sand.
(a)
Table 3. Objective and Tested Variables in Experimental Heats Heat Objective Casting (mold) Variables 1
(b) Fig. 6. These photomicrographs show “skin” on continuous castings produced in a water cooled graphite mold ([a] un-etched and [b] - etched).
EXPERIMENTAL AND MODELING PROCEDURES EXPERIMENTAL HEATS Three sets of experimental heats, designated as Heat 1, Heat 2 and Heat 3, were prepared using base pearlitic ductile iron melted in a 200-lb (90.7-kg) induction furnace. Low sulfur containing charge materials included industrial ductile iron returns and plain carbon steel scrap (Table 1). The liquid metal was heated to 1500C (2732F), tapped at 1480C (2696F) and treated in a pocket ladle with Fe-Si-Mg followed by FeSi base inoculants. The chemistry of base ductile iron in each heat was poured into copper chill molds and analyzed with an Oxford Arc Spectrometer and a CS600 Leco C-S determinator. Table 2 displays the average chemistries of these heats in the ladle after treatment. Table 1. Charge for Melting Ductile Iron In charge Weight (lb) Plain-C steel disks
50
DI return
80
Induction Iron
20
Carbon riser
3
Cu Total induction furnace charge
0.8 153.8
Ladle treatment FSM (46% Si, 5.7% Mg, 1% Ca, 0.4% La, 1% max Al,), 1-10 mm Inoculants (75% Si, 4% Al, 1% Ca), 1-0.2 mm
2.4 0.8
“Skin” formation Wedge (no-bake mold) - Pouring T in no-bake - Inoculation casting - Cooling rate
2
Metal/Smodified mold interactions
“Star” casting (no-bake mold with Smodified core inserts inserts)
3
Dual graphite structure development
Plate casting with internal passages: - flat S-modified core inserts - “Muffin pan” chill plate with Smodified core inserts
(a) (b)
- Sulfur content - Sulfur agent type - Position - Coating - Sulfur content - Pouring T - Coating - Venting
In Heat 1, the variables’ effects, including pouring temperature, carbon equivalent (CE) and inoculation, on the formation of the surface skin in no-bake mold castings were verified at two casting sections with different cooling rates (Table 4). Base ductile iron was poured into six no-bake, keel block molds with thick (3 in. [7.62 cm]) and thin (1in. [2.54 cm]) sections. Table 4. Pouring Sequence and Studied Variables in Heat 1 Ladle CE Pouring T, Inoculation o C 1 4.30 (low) 1420 (high) 2 4.49 (high) 1400 (high) 3 4.30 (low) 1300 (low) 4 4.29 (low) 1300 (low) x 5 4.49 (high) 1320 (low) 6 4.48 (high) 1320 (low) x
In Heat 2 and Heat 3, the S-modified inserts were set in no-bake molds to study the mold-metal interactions and the possibility of developing a dual graphite structure. Two different sulfur reagents, a yellow elemental sulfur and iron pyrite, FeS2, were used to produce different sulfur contents (2% - 8% weight percent) in S-modified inserts. Sulfur reagents were mixed with dry sand and the no-bake binder was then added. Inserts were shaped as wedges (Fig. 7a) or flat plates. In Heat 2, 6 in. x 6 in. x
Panel 13-1320.pdf, Page 4 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
3 in. (15.2 cm x 15.2 cm x 7.6 cm) casting plates were surrounded by four wedge shape inserts, each of which was adopted to investigate the effect of multiple variables. These variables included a sulfur reagent type and sulfur content, an insert position in the mold and additional coating (Fig. 8). In the latter, a water based graphite coating with 0.2 mm (0.008 in.) thick surface layer and 0.5 mm (0.02 in.) penetration into the sand was used (Fig. 7b).
Microscopy/Energy Dispersive X-ray (SEM/EDX) Aspex system.
(a)
(a) (b) Fig. 7. These photos show (a) S-modified core inserts and (b) the structure of a coated core insert.
(b) (c) Fig. 9. These photos show: (a) a plate casting with internal passages; (b) a S-modified core insert tablet and (c) a “muffin pan” type chill plate with installed thermocouples.
MODELING Computational Fluid Dynamics (CFD) Fluent commercial software was used to model heat and mass transfer during pouring into the mold and casting solidification. Additionally, temperature fields in the core inserts and the solidified castings were analyzed to designate a period of time when sulfur could be vaporized and penetrate into the mold cavity. Fig. 8. “Star” casting used in Heat 2 with 4 wedge type core inserts is pictured.
In Heat 3, the casting design included three 10 x 10 x 1 in. (25.4 x 25.4.3 x 2.54 cm) plates separated by two 8 x 8 x 1 in. (20.3 x 20.3 x 2.54 cm) internal cores (Fig. 9a). This design simulated a diesel engine head with passages. Two types of S-modified core inserts were tested: (1) a flat insert (6 x 6 x1 in. [15.2 cm x 15.2 cm x 2.54 cm]) installed at the bottom of the mold cavity and (2) 2 in. (5.08 cm) diameter tablets (Fig. 9b) mounted into a specially designed “muffin pan” chill plate (Fig. 9c). The purpose of this design is described, subsequently. The thermal cooling curves from castings were obtained with K-type thermocouples, protected by quartz tubes and connected to a 24-bit National Instrument DAQ. Castings from each experimental heat were sectioned and surface ground for macro-observation. Metallographic samples were cut from different locations in the castings. Those samples were then polished and observed with optical microscopy and an automated Scanning Electron
FACTORS AFFECTING THE “SKIN” IN DUCTILE IRON CASTINGS FROM NO-BAKE MOLDS In Heat 1, high and low carbon equivalent (CE), high and low pouring temperature and with extra inoculation were varied in six molds as given in Table 4. The typical local solidification time obtained from the center of the thick and thin sections in poured casting was near 100 sec and 400 sec, respectively (Fig. 10a). Metallography samples perpendicular to the casting surface were prepared. The maximum thickness of the surface skin (L) was measured for each case, as given in Table 5. A comparison of casting surface layers on thick (Fig. 10b) and thin (Fig. 10c) sections indicated that cooling rate and solidification time had major impacts on the thickness of the degenerate graphite layer. The effects of tested variables on the maximal thickness of a casting surface “skin” are summarized in Table 6. Several of the variables, such as the carbon equivalent, had a minor effect on the thickness of the casting surface “skin” as well. The significant negative effect of an increased casting surface roughness was identified.
Panel 13-1320.pdf, Page 5 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
In Heat 1, the no-bake mold material contained less than 0.02% sulfur. This amount of sulfur and an oxidizing gas environment were enough for the formation of a limited thickness (less than 0.5 mm [0.02 in.]) degenerate layer. This layer can influence the mechanical properties of thinwall castings but was not thick enough to develop a dual graphite structure in medium and heavy sections. For the following experimental heats, inserts made of no-bake sand were artificially modified with mixing a de-nodulizing element (sulfur), to enhance the degradation of graphite nodules near the casting surface to the extent to produce an engineered layer for a controlled dual graphite structure in iron castings.
(c)
Fig. 10. This graphic shows: (a) cooling curves from the center casting (arrows indicate the local solidification time) and the microstructure of the surface layers with degenerate graphite in (b)thick and (c) thin sections of wedge casting.
Table 5. Variation in the Maximum Thickness of “Skin” in the Castings from No-Bake Molds (Heat 1) Ladle/Section 1/thick 1/thin 2/thin 3/thin 4/thin 5/thin 6/thin
Variables
Effect Strong Weak
CE Pouring T Inoculation Roughness Cooling rate
X X X X X
Higher CE gives thinner layer Lower T gives thinner layer Inoculation gives thinner layer Rough surface gives thicker layer Slower cooling gives thicker layer
DEVELOPMENT OF DUAL GRAPHITE STRUCTURE BY S-MODIFIED INSERTS THERMAL ANALYSIS OF S-MODIFIED NO-BAKE CORE Prior to the experimental heats, thermodynamic calculations (FactSage software) were performed to predict the decomposition temperatures of two sulfur containing reagents: an elemental yellow sulfur and a compound FeS2 (Fig. 11a). The predicted vaporization temperatures were experimentally verified for base and S-modified (5% S) no-bake sand with an in-house built large scale thermogravimetric analyzer (TGA) (Fig. 11b). The TGA curve for no-bake sand began dropping at 80C (176F) due to the binder decomposition, which made the other two curves also gradually drop near this temperature. The tested decomposition temperature of yellow sulfur began at approximately 200C (392F), while the decomposition of FeS2 began above 400C (752F). Based on this data, yellow sulfur was used as an addition to the inserts in the majority of tests.
(a)
(b)
Table 6. Effect of Different Factors on the Thickness of a Surface Layer with a Degenerate Graphite Shape in a No-Bake Mold (Heat 1)
Variables high T, low CE high T, low CE high T, high CE low T, low CE low T, low CE+ inoculant low T, high CE low T, high CE+ inoculant
Max L, µm 435 253 220 158 94 290 169
EFFECT OF VARIABLES ON DUAL GRAPHITE STRUCTURE The “star” casting in Heat 2 allowed a series of comparisons among variables in a single casting. For example, four wedge yellow S-modified core inserts were installed at the sides of the casting plate. Examples of macro-observations on the distributions of modified structure are shown in Fig. 12. In this test, the wedgeshaped S-modified core inserts served as an excellent reactant and part of the sulfur diffused from the insert into the nearby mold cavity filled by liquid metal. At the same time, the sulfur transported in the vertical direction from the bottom to the top of the casting and, to some extent, horizontally beyond the insert boundary. This movement indicated that gas phase transport of sulfur as bubbles in the liquid metal was involved in addition to S-transport by the melt convection. Different yellow sulfur contents in the cores were tested in this heat. As predicted, 5% S in the core produced a thicker modified structure than a 2% S in the core insert. The interface between the untreated and the S-treated metal was sharper and clearer when an insert with higher sulfur content was used (Fig. 13).
Panel 13-1320.pdf, Page 6 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
The effect of coating on thickness of the degenerate layer and casting surface finishing was investigated with 5% yellow S-modified core inserts. A significant improvement in surface quality with a coated core can be seen in Fig. 14. The thin layer of coating isolated the metal from direct contact with the S-modified core inserts. It also reduced metal penetration into the sand, which was observed in un-coated inserts. Two types of sulfur containing reagents, purified precipitated yellow sulfur powder and FeS2 (100 mesh), were also compared in the “star” casting with application of coated inserts. Yellow sulfur produced a larger volume of modified structure than did the FeS2. The internal interface between the unmodified graphite and the modified graphite for yellow S was more diffuse than with the FeS2 (Fig. 15).
(a)
(a)
(b) Fig. 11. These graphs plot (a) thermodynamic prediction and (b)TGA decomposition of S-modified no-bake sand.
(a)
(b)
(c) Fig. 12. These photos show (a) macrostructure in a vertical section and two horizontal sections: (b) near the bottom and (c) the top in contact with 5% yellow Smodified core inserts (Heat 2). Red lines indicate boundary.
(b) Fig. 13. These micrographs show the transition from spheroidal graphite to degenerate graphite: (a) 5% and (b) 2% of yellow S-modified core inserts.
(a)
(b)
(c) Fig. 14. This graphic shows the external casting/mold interface: (a) uncoated and (b) coated 5% yellow S-modified core inserts and (c) the comparison of modified area in the casting.
Panel 13-1320.pdf, Page 7 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
(a)
(b)
(a)
(c) Fig. 15. This graphic shows the boundary between the flake graphite layer and the matrix with spheroidal graphite: (a) diffuse (yellow S-modified core inserts) and (b) sharp (FeS2-modified core) and (c) comparison on modified area in the casting. Both modified cores were coated.
IN-SITU STRUCTURE MODIFICATION BY FLAT SMODIFIED CORE INSERTS The in-situ structure modification of Mg-treated metals by applying S-modified core inserts was studied in the Heat 3a set by using a plate-type casting with two internal cores simulating a diesel engine head. A flat 5% yellow S-modified core insert was installed into the no-bake mold to treat the bottom plate of the casting. The pouring temperature was varied from 1350C to 1420C (2462F to 2588F). Test results indicated that a higher pouring temperature gave a thicker, though less consistent modified zone (Fig. 16). Moreover, the modified structure was not localized in the bottom plate. Flake graphite was also found in the middle and even in the top casting plates for this higher pouring temperature case. To analyze the results on in-situ structure modification, experimentally obtained cooling curves from the core insert and the casting were compared to S-vaporization temperature from TGA data. Two arrows in Fig. 17 indicate the start of S-vaporization, when the insert temperature exceeded the equilibrium vaporization temperature and the end of the treatment process when casting developed a solid skin. However, even after formation of solid skin, part of treated liquid metal containing sulfur can travel within the mold cavity by liquid metal convection. This effect was more pronounced at a higher pouring temperature.
(b) Fig. 16. Distribution of the modified structure in the vertical sections of the castings with a bottom flat 5% yellow S-modified core insert, was poured at (a)1350 C (2462F) and (b) 1420C (2588F), respectively. Red lines outline the region with the modified structure.
Fig. 17. Cooling curves are obtained from the bottom and top plates of the casting and from near surface layer of flat S-modified core insert. The arrows indicate the period of possible mold-metal interactions.
COMBINING AN IN-SITU STRUCTURE MODIFICATION WITH LOCAL CHILLING A series of virtual cases were modeled using CFD to prevent liquid metal convection and localize the treated zone in the bottom plate of the casting. Modeling showed
Panel 13-1320.pdf, Page 8 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
that the combination of in-situ S-modified inserts with localized chilling could be used for process control. Based on CFD modeling, the so-called “muffin pan” concept of a chill plate with installed 2 in. (5.08 cm) diameter S-modified insert tablets was proposed (Fig. 18) and tested in the Heat 3b set. The experimental cooling curves obtained from the tablet inserts installed in the chill plate, the bottom plate and the top plate of the casting are shown in Fig. 19. The chilling effect shortened the duration of the sulfur reaction from 100 sec for the unchilled case (Fig. 17) to 50 sec. It also decreased the bottom plate’s solidification time compared to the middle and the top plates, as in Fig. 19. (a)
A series of castings were produced in Heat 3a with variations in yellow S-content and “muffin pan” design. At a high sulfur content (8%), the casting was overtreated. Flake graphite was observed even in the top casting plate (Fig. 20). In this case, mushroom-like modified zones were formed and sulfur gas bubbles transported sulfur in both vertical and horizontal directions. Decreasing the sulfur content in the core inserts to 5% resulted in a milder reaction. The reaction continued to occur, however, around the S-gas bubbles. The bubbles formed the protuberances in the modified zones seen in Figs. 21b and 21c. Many of the S-bubbles had sufficient time to dissolve in the metal, producing localized reaction “islands” as visible in Fig. 21d. The S-bubbles that survived led to gas pores observed in the casting (Fig. 21a). Segregated sulfur, stopped by the central core, led to a flat region of flake graphite below the central core. Vents were drilled through on each of the pockets on the “muffin pan” chill plate to release the gas pressure created by sulfur vaporization in the core inserts. In addition, the sulfur content was further decreased to 2 - 4% to minimize over-treatment. As shown in Fig. 22, the modified structure was localized primarily in the bottom plate. No degenerate graphite was found in the top plate. The last experimental conditions indicated the possibility of controlling layering structure by variations in studied parameters. Improving the layering of the dual-graphite structure and eliminating the porosity will be possible to accomplish in the future by optimizing these process parameters.
(b) Fig. 18. This is a CFD modeling prediction (10 sec after pouring): (a) liquid fraction (in casting volume, not filled image) and (b) temperature distribution in the vertical section of the casting and the bottom positioned “muffin pan” chill plate with S-modified inserts.
Fig. 19. These are the measured cooling curves at different locations in the casting and S-modified insert inside “muffin pan” chill plate. Arrows indicate a starting and ending time of a possible sulfur reaction with the Mg-treated metal.
Panel 13-1320.pdf, Page 9 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
CONCLUSIONS Casting “skin” layers with degenerate spheroidal graphite are considered deleterious to the mechanical properties and machinability of thin-wall ductile iron castings. The effects of CE, pouring temperature, inoculation, and surface roughness, on the ductile iron casting “skin” have been verified in this article. Improving inoculation and decreasing surface roughness by coating no-bake sand molds are possible methods for controlling the thickness of a casting skin in ductile iron casting, specifically castings made in no-bake molds.
Fig. 20. Excessive sulfur in the modified inserts (8%) over-treated the casting. Significant vertical and horizontal gas sulfur transportations were observed.
Conversely, the enhanced interaction kinetics of Mgtreated ductile iron with molds can be applied to produce iron castings with a dual graphite (spheroidal and flake) structure. An in-situ de-nodulizing treatment of ductile iron liquid metal by S-modified no-bake sand inserts was investigated to develop structure modification in local zones. The localized change of the graphite’s shape from spheroidal to flake was successfully achieved through this process. Forming a desired macro-modified layer throughout the casting, however, is a challenge. The parameters affecting this process, including sulfur type and content, coating, pouring temperature, position for Smodified core inserts and chilling effect were designed, modeled and experimentally tested. An observation of the protuberances and localized island-like areas with degenerate graphite shape indicated a “foot print” of sulfur transport. The localization of graphite morphological degeneration can be accomplished by the combination of S-modified core inserts with a chilling effect. Both sulfur gas transport and liquid metal convection, however, were obstacles that actually redistributed the treated metal and produced gas porosity. Improving the dual-graphite structure layering and eliminating the porosity can be achieved in the future by optimizing the process parameters.
Fig. 21. There are specific features of 5% yellow Smodified core inserts—metal interactions: (a) gas bubbles (possible from S2-gas) surrounded by flake graphite; (b and c) protuberances from the bottom surface flake graphite region and (d) localized flake graphite spot in the top plate.
a b
ACKNOWLEDGMENTS
Fig. 22. This graphic shows the (a) the distribution of the modified structure in the vertical sections of castings with a bottom “muffin pan” chill plate with 4% yellow S-modified core inserts and (b) sharp boundary between flake and spherical graphite zones. Black lines indicate boundaries.
This research was sponsored by Benet Laboratories on behalf of the US Army Contracting Command Joint Munitions and Lethality Contracting Center— accomplished under Cooperative Agreement Number W15QKN-11-2-0001. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of AFS or Benet Laboratories or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation here in stated. The authors would like to thank US Army ARDEC-Benet laboratories for funding this research. The authors wish to acknowledge technical support and discussion input of George Kokos, James Barlow and Zhiping Lin from Caterpillar. The authors wish to recognize the assistance
Panel 13-1320.pdf, Page 10 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
of Mingzhi Xu, Marc Harris, Forrest Huebner, Clinton Ratliff, John Stanek, Jeremy Robinson, Bradley Bromet, Ian Christian and James Smoot for heat pouring and sample preparation. REFERENCES 1.
2.
3.
4.
5.
6.
7.
8.
Lekakh, S., Richards, V., Peaslee, K., “ThermoChemistry of Non-Metallic Inclusions in DI. Proceedings,” The Carl Loper Cast Iron Symposium, Madison, WI (2009). Lekakh, S., Robertson, D., Loper, C., “Thermochemistry of Nodulization and Inoculation of Irons”, American Cast Iron Inoculation Conference Proceedings (2005). Peach, W., Lekakh, S., Richards V., Peaslee, K., Teague, J., “Effects of Near-Surface Metallurgy on Machinability of Gray Cast Iron,” AFS Transactions, vol. 116, paper 08-1 (2008). Stefanescu, D., Wills S., Massone, J., “Quantification of Casting Skin in Ductile and Compacted Graphite Irons and Its Effect on Tensile Properties,” AFS Transactions, vol. 117, pp. 587-606 (2009). Boonmee, S., Stefanescu, D., “The Mechanism of Formation of Casting Skin in CG Iron and its Effect on Tensile Properties,” Key Engineering Materials, vol. 457, pp. 11-16 (2011). Goodrich, G., Lobenhofer, R., “Effect of Cooling Rate on Pearlitic Ductile Iron Mechanical Properties,” AFS Transactions, vol. 115, paper 07045 (2007). Lekakh, S., Qing, J., Richards, V., “Investigation of Cast Iron Processing to Produce Controlled Dual Graphite Structure in Castings,” AFS Transactions, paper 12-024 (2012). Lekakh, S., Qing, J., Richards, V., “Modeling of Melt Mixing Phenomena in Cast Iron with Dual Graphite Structure,” TMS 2012, 141st Annual Meeting and Exhibition, Supplemental Proceedings, vol. 2, pp. 347-354 (2012).
No-Bake S-Containing Mold-DI Metal Interactions: Consequences and Potential Application J. Qing, S. Lekakh, V. Richards Missouri University of Science and Technology, Rolla, Missouri Copyright 2013 American Foundry Society ABSTRACT The ductile iron liquid metal is very reactive with environmental gases and solid molds after a nodulizing treatment. Mold-metal reactions could alter the spheroidal graphite shape to vermicular or flake. In this research, the metal-mold reactions were enhanced by various sulfur additions to the no-bake mold inserts to study these phenomena and to artificially produce layered dual-graphite structure in the casting. Thermodynamic and solidification modeling together with sets of different experimental techniques were applied to a test casting with internal core passages representing the cylinder head of a diesel engine. The effects of insert surface shape and location in the mold; content and type of sulfur-containing agent; coating; pouring temperature and additional chilling were experimentally studied and discussed. These results could be used to better understand the metal-mold interaction, prevent surface defects and design dual-graphite structure processes.
atmosphere around the liquid metal surface (Fig. 2a) or the complex composition surface slag moved from the bulk metal on its surface (Fig. 2b). These products were observed on a levitation re-melted ductile iron droplet.
Fig. 1. Possible interactions between Mg-treated liquid metal and the environment are diagramed.
Keywords: Mg-treated cast iron, metal-mold interaction, dual graphite structure, sulfur-modified insert INTERACTIONS OF Mg-TREATED CAST IRON WITH ENVIRONMENT A magnesium treated ductile iron liquid metal is highly reactive with its surrounding gases or the solid environment (mold). Various possible interactions between Mg-treated metals and the environments are classified in Fig. 1. The negative effects of these interactions can be Mg-fading and consequent spherical graphite shape degeneration in the bulk casting or shape degeneration in specific surface regions, as the results of metal-mold interactions. Additionally, the reaction products (insulated inclusions or aggregated into slag) could have negative effects on casting performance and machinability. A better understanding of these complicated phenomena is important for day-to-day foundry practice. While holding magnesium-treated ductile iron in the ladle or automatic pouring devices, two different processes could take place: (1) vaporization of magnesium having a high partial pressure and (2) adsorption and diffusion of oxygen into the liquid metal and reaction of oxygen with magnesium dissolved in the metal. These two processes give different types of reaction products. In the first instance, the MgO whiskers could grow into the
(a)
(b) Fig. 2. These are reaction products that are observed on the surface of the levitation re-melted ductile iron droplet: (a) MgO whiskers above the re-melted droplet surface and (b) Mg-Si-O, Mg-Si-Ca-O, and REM-Si-O (white) phases on the droplet surface.
Panel 13-1320.pdf, Page 2 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
Steel
Molten steel and ductile iron metal have different reaction sequences with metalloid-type impurities (oxygen or sulfur). For example, when refining molten steel, alkali and rare earth metals will first deeply deoxidize the liquid steel. Then, further additions produce desulfurization. However, alkali and rare earth metals in the liquid cast irons have different sequences of refining reactions. Some will react with oxygen first (Mg) and some with sulfur (Ca). Figure 3 schematically illustrates the reaction sequences of liquid steel (low carbon) and liquid cast iron (high carbon) with Mg, Ca or REM additions.1, 2 In the case of the reverse process, when a Mg-treated liquid metal reacts with the environment, dissolved Mg will be oxidized first, based on thermodynamic equilibrium. However, multiple reaction products have been observed on a levitation re-melted droplet surface (in air) (Fig. 2b) because all active elements (Mg, Si, Ca, REM) can be oxidized simultaneously in a surface layer containing oxygen.
[O]
[S]
[O]
[S]
Mg [S]
Cast iron
[S]
[O]
[S]
(a)
Ca
[O]
[O]
casting process). This layer, however, was formed after solidification. It did not exhibit a degenerate graphite shape (Fig. 6) when compared with that for conventional sand mold casting processes (green or no-bake).
[O]
(b)
Ce
Fig. 4. This green sand mold casting skin with multiple features includes: (a) de-carburized and graphite degradation layers; (b) an iron oxide subsurface film and (c) a liquid slag/mold reaction product, also incorporated into the subsurface.
Ca
Mg [S]
(c)
Ce
Additions, wt.%
Fig. 3. Sequences of the refining reactions in the 1, 2 liquid steel and the cast iron are illustrated.
Another important group of interactions took place when Mg-treated cast iron was poured into molds (Fig. 1). A casting skin was formed when the iron contacted the external mold environments during solidification. Defects in this skin include a degenerate graphite shape, high temperature reaction products, porosity and exogenous inclusions from a reaction of surface slag/dross with molding sand or even non-reacted sand particles. Figure 4 illustrates different types of features observed on the skin of the ductile iron casting, solidified in a green sand mold.3 No-bake molds provide better casting surface quality. However, even in this case, the casting surface had a degenerate graphite layer (see flake graphite in Fig. 5). Boonmee and Stefanescu studied the mechanism of formation for casting skin in Mg-treated irons with spherical and compacted graphite shapes.4, 5 It was shown that the casting skin can have a negative influence on the mechanical properties of thin wall castings.6 The surface of the cast bar had a limited oxidized/decarburized layer when the Mg-treated metal solidified in a water-cooled graphite mold (continuous
Fig. 5. This photomicrograph shows the degenerate graphite phase layer on the casting surface from a no-bake mold.
In the industrial case described previously, the Mg-treated metal, gas or solid environment interactions had negative effects. These effects included decreased machinability3 or decreased mechanical properties of unmachined thinwall castings with as cast surfaces.4, 5 In some cases, however, castings with an enlarged, layered dual graphite structure may have advantages when compared to castings with only flake or spheroidal graphite. A dual graphite structure is referred to as a structure that contains both spheroidal graphite and flake graphite in specific casting regions. One potential application of this dual graphite structure in iron casting is on a diesel engine head, which requires high strength and excellent thermal conductivity in different regions for a single casting. Our previous studies succeeded in producing dual graphite structure in the casting by the in-stream de-nodulizing treatment of a particular portion of the Mg-treated metal
Panel 13-1320.pdf, Page 3 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
poured into the mold.7, 8 In this paper, a localized in-situ de-nodulizing treatment was studied to produce dual graphite structure in the casting.
Table 2. Average Chemistry (wt %) of Experimental Ductile Irons in Ladle Heat C Si Mn Cu S P Ni Mg 1 2 3
3.60 2.0-2.7 0.31 3.70 2.5 0.30 3.67 2.6 0.34
0.6 0.004 0.01 0.02 0.043 0.6 0.004 0.01 0.03 0.046 0.6 0.003 0.01 0.03 0.049
The objectives, designs and tested variables of experimental heats are summarized in Table 3. No-bake molds made of F-60 silica sand were used in all of the heats. The total sulfur content in the no-bake sand was tested with the C-S determinator as 0.02%. The special S-modified core insert were also produced from no-bake F-60 sand.
(a)
Table 3. Objective and Tested Variables in Experimental Heats Heat Objective Casting (mold) Variables 1
(b) Fig. 6. These photomicrographs show “skin” on continuous castings produced in a water cooled graphite mold ([a] un-etched and [b] - etched).
EXPERIMENTAL AND MODELING PROCEDURES EXPERIMENTAL HEATS Three sets of experimental heats, designated as Heat 1, Heat 2 and Heat 3, were prepared using base pearlitic ductile iron melted in a 200-lb (90.7-kg) induction furnace. Low sulfur containing charge materials included industrial ductile iron returns and plain carbon steel scrap (Table 1). The liquid metal was heated to 1500C (2732F), tapped at 1480C (2696F) and treated in a pocket ladle with Fe-Si-Mg followed by FeSi base inoculants. The chemistry of base ductile iron in each heat was poured into copper chill molds and analyzed with an Oxford Arc Spectrometer and a CS600 Leco C-S determinator. Table 2 displays the average chemistries of these heats in the ladle after treatment. Table 1. Charge for Melting Ductile Iron In charge Weight (lb) Plain-C steel disks
50
DI return
80
Induction Iron
20
Carbon riser
3
Cu Total induction furnace charge
0.8 153.8
Ladle treatment FSM (46% Si, 5.7% Mg, 1% Ca, 0.4% La, 1% max Al,), 1-10 mm Inoculants (75% Si, 4% Al, 1% Ca), 1-0.2 mm
2.4 0.8
“Skin” formation Wedge (no-bake mold) - Pouring T in no-bake - Inoculation casting - Cooling rate
2
Metal/Smodified mold interactions
“Star” casting (no-bake mold with Smodified core inserts inserts)
3
Dual graphite structure development
Plate casting with internal passages: - flat S-modified core inserts - “Muffin pan” chill plate with Smodified core inserts
(a) (b)
- Sulfur content - Sulfur agent type - Position - Coating - Sulfur content - Pouring T - Coating - Venting
In Heat 1, the variables’ effects, including pouring temperature, carbon equivalent (CE) and inoculation, on the formation of the surface skin in no-bake mold castings were verified at two casting sections with different cooling rates (Table 4). Base ductile iron was poured into six no-bake, keel block molds with thick (3 in. [7.62 cm]) and thin (1in. [2.54 cm]) sections. Table 4. Pouring Sequence and Studied Variables in Heat 1 Ladle CE Pouring T, Inoculation o C 1 4.30 (low) 1420 (high) 2 4.49 (high) 1400 (high) 3 4.30 (low) 1300 (low) 4 4.29 (low) 1300 (low) x 5 4.49 (high) 1320 (low) 6 4.48 (high) 1320 (low) x
In Heat 2 and Heat 3, the S-modified inserts were set in no-bake molds to study the mold-metal interactions and the possibility of developing a dual graphite structure. Two different sulfur reagents, a yellow elemental sulfur and iron pyrite, FeS2, were used to produce different sulfur contents (2% - 8% weight percent) in S-modified inserts. Sulfur reagents were mixed with dry sand and the no-bake binder was then added. Inserts were shaped as wedges (Fig. 7a) or flat plates. In Heat 2, 6 in. x 6 in. x
Panel 13-1320.pdf, Page 4 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
3 in. (15.2 cm x 15.2 cm x 7.6 cm) casting plates were surrounded by four wedge shape inserts, each of which was adopted to investigate the effect of multiple variables. These variables included a sulfur reagent type and sulfur content, an insert position in the mold and additional coating (Fig. 8). In the latter, a water based graphite coating with 0.2 mm (0.008 in.) thick surface layer and 0.5 mm (0.02 in.) penetration into the sand was used (Fig. 7b).
Microscopy/Energy Dispersive X-ray (SEM/EDX) Aspex system.
(a)
(a) (b) Fig. 7. These photos show (a) S-modified core inserts and (b) the structure of a coated core insert.
(b) (c) Fig. 9. These photos show: (a) a plate casting with internal passages; (b) a S-modified core insert tablet and (c) a “muffin pan” type chill plate with installed thermocouples.
MODELING Computational Fluid Dynamics (CFD) Fluent commercial software was used to model heat and mass transfer during pouring into the mold and casting solidification. Additionally, temperature fields in the core inserts and the solidified castings were analyzed to designate a period of time when sulfur could be vaporized and penetrate into the mold cavity. Fig. 8. “Star” casting used in Heat 2 with 4 wedge type core inserts is pictured.
In Heat 3, the casting design included three 10 x 10 x 1 in. (25.4 x 25.4.3 x 2.54 cm) plates separated by two 8 x 8 x 1 in. (20.3 x 20.3 x 2.54 cm) internal cores (Fig. 9a). This design simulated a diesel engine head with passages. Two types of S-modified core inserts were tested: (1) a flat insert (6 x 6 x1 in. [15.2 cm x 15.2 cm x 2.54 cm]) installed at the bottom of the mold cavity and (2) 2 in. (5.08 cm) diameter tablets (Fig. 9b) mounted into a specially designed “muffin pan” chill plate (Fig. 9c). The purpose of this design is described, subsequently. The thermal cooling curves from castings were obtained with K-type thermocouples, protected by quartz tubes and connected to a 24-bit National Instrument DAQ. Castings from each experimental heat were sectioned and surface ground for macro-observation. Metallographic samples were cut from different locations in the castings. Those samples were then polished and observed with optical microscopy and an automated Scanning Electron
FACTORS AFFECTING THE “SKIN” IN DUCTILE IRON CASTINGS FROM NO-BAKE MOLDS In Heat 1, high and low carbon equivalent (CE), high and low pouring temperature and with extra inoculation were varied in six molds as given in Table 4. The typical local solidification time obtained from the center of the thick and thin sections in poured casting was near 100 sec and 400 sec, respectively (Fig. 10a). Metallography samples perpendicular to the casting surface were prepared. The maximum thickness of the surface skin (L) was measured for each case, as given in Table 5. A comparison of casting surface layers on thick (Fig. 10b) and thin (Fig. 10c) sections indicated that cooling rate and solidification time had major impacts on the thickness of the degenerate graphite layer. The effects of tested variables on the maximal thickness of a casting surface “skin” are summarized in Table 6. Several of the variables, such as the carbon equivalent, had a minor effect on the thickness of the casting surface “skin” as well. The significant negative effect of an increased casting surface roughness was identified.
Panel 13-1320.pdf, Page 5 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
In Heat 1, the no-bake mold material contained less than 0.02% sulfur. This amount of sulfur and an oxidizing gas environment were enough for the formation of a limited thickness (less than 0.5 mm [0.02 in.]) degenerate layer. This layer can influence the mechanical properties of thinwall castings but was not thick enough to develop a dual graphite structure in medium and heavy sections. For the following experimental heats, inserts made of no-bake sand were artificially modified with mixing a de-nodulizing element (sulfur), to enhance the degradation of graphite nodules near the casting surface to the extent to produce an engineered layer for a controlled dual graphite structure in iron castings.
(c)
Fig. 10. This graphic shows: (a) cooling curves from the center casting (arrows indicate the local solidification time) and the microstructure of the surface layers with degenerate graphite in (b)thick and (c) thin sections of wedge casting.
Table 5. Variation in the Maximum Thickness of “Skin” in the Castings from No-Bake Molds (Heat 1) Ladle/Section 1/thick 1/thin 2/thin 3/thin 4/thin 5/thin 6/thin
Variables
Effect Strong Weak
CE Pouring T Inoculation Roughness Cooling rate
X X X X X
Higher CE gives thinner layer Lower T gives thinner layer Inoculation gives thinner layer Rough surface gives thicker layer Slower cooling gives thicker layer
DEVELOPMENT OF DUAL GRAPHITE STRUCTURE BY S-MODIFIED INSERTS THERMAL ANALYSIS OF S-MODIFIED NO-BAKE CORE Prior to the experimental heats, thermodynamic calculations (FactSage software) were performed to predict the decomposition temperatures of two sulfur containing reagents: an elemental yellow sulfur and a compound FeS2 (Fig. 11a). The predicted vaporization temperatures were experimentally verified for base and S-modified (5% S) no-bake sand with an in-house built large scale thermogravimetric analyzer (TGA) (Fig. 11b). The TGA curve for no-bake sand began dropping at 80C (176F) due to the binder decomposition, which made the other two curves also gradually drop near this temperature. The tested decomposition temperature of yellow sulfur began at approximately 200C (392F), while the decomposition of FeS2 began above 400C (752F). Based on this data, yellow sulfur was used as an addition to the inserts in the majority of tests.
(a)
(b)
Table 6. Effect of Different Factors on the Thickness of a Surface Layer with a Degenerate Graphite Shape in a No-Bake Mold (Heat 1)
Variables high T, low CE high T, low CE high T, high CE low T, low CE low T, low CE+ inoculant low T, high CE low T, high CE+ inoculant
Max L, µm 435 253 220 158 94 290 169
EFFECT OF VARIABLES ON DUAL GRAPHITE STRUCTURE The “star” casting in Heat 2 allowed a series of comparisons among variables in a single casting. For example, four wedge yellow S-modified core inserts were installed at the sides of the casting plate. Examples of macro-observations on the distributions of modified structure are shown in Fig. 12. In this test, the wedgeshaped S-modified core inserts served as an excellent reactant and part of the sulfur diffused from the insert into the nearby mold cavity filled by liquid metal. At the same time, the sulfur transported in the vertical direction from the bottom to the top of the casting and, to some extent, horizontally beyond the insert boundary. This movement indicated that gas phase transport of sulfur as bubbles in the liquid metal was involved in addition to S-transport by the melt convection. Different yellow sulfur contents in the cores were tested in this heat. As predicted, 5% S in the core produced a thicker modified structure than a 2% S in the core insert. The interface between the untreated and the S-treated metal was sharper and clearer when an insert with higher sulfur content was used (Fig. 13).
Panel 13-1320.pdf, Page 6 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
The effect of coating on thickness of the degenerate layer and casting surface finishing was investigated with 5% yellow S-modified core inserts. A significant improvement in surface quality with a coated core can be seen in Fig. 14. The thin layer of coating isolated the metal from direct contact with the S-modified core inserts. It also reduced metal penetration into the sand, which was observed in un-coated inserts. Two types of sulfur containing reagents, purified precipitated yellow sulfur powder and FeS2 (100 mesh), were also compared in the “star” casting with application of coated inserts. Yellow sulfur produced a larger volume of modified structure than did the FeS2. The internal interface between the unmodified graphite and the modified graphite for yellow S was more diffuse than with the FeS2 (Fig. 15).
(a)
(a)
(b) Fig. 11. These graphs plot (a) thermodynamic prediction and (b)TGA decomposition of S-modified no-bake sand.
(a)
(b)
(c) Fig. 12. These photos show (a) macrostructure in a vertical section and two horizontal sections: (b) near the bottom and (c) the top in contact with 5% yellow Smodified core inserts (Heat 2). Red lines indicate boundary.
(b) Fig. 13. These micrographs show the transition from spheroidal graphite to degenerate graphite: (a) 5% and (b) 2% of yellow S-modified core inserts.
(a)
(b)
(c) Fig. 14. This graphic shows the external casting/mold interface: (a) uncoated and (b) coated 5% yellow S-modified core inserts and (c) the comparison of modified area in the casting.
Panel 13-1320.pdf, Page 7 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
(a)
(b)
(a)
(c) Fig. 15. This graphic shows the boundary between the flake graphite layer and the matrix with spheroidal graphite: (a) diffuse (yellow S-modified core inserts) and (b) sharp (FeS2-modified core) and (c) comparison on modified area in the casting. Both modified cores were coated.
IN-SITU STRUCTURE MODIFICATION BY FLAT SMODIFIED CORE INSERTS The in-situ structure modification of Mg-treated metals by applying S-modified core inserts was studied in the Heat 3a set by using a plate-type casting with two internal cores simulating a diesel engine head. A flat 5% yellow S-modified core insert was installed into the no-bake mold to treat the bottom plate of the casting. The pouring temperature was varied from 1350C to 1420C (2462F to 2588F). Test results indicated that a higher pouring temperature gave a thicker, though less consistent modified zone (Fig. 16). Moreover, the modified structure was not localized in the bottom plate. Flake graphite was also found in the middle and even in the top casting plates for this higher pouring temperature case. To analyze the results on in-situ structure modification, experimentally obtained cooling curves from the core insert and the casting were compared to S-vaporization temperature from TGA data. Two arrows in Fig. 17 indicate the start of S-vaporization, when the insert temperature exceeded the equilibrium vaporization temperature and the end of the treatment process when casting developed a solid skin. However, even after formation of solid skin, part of treated liquid metal containing sulfur can travel within the mold cavity by liquid metal convection. This effect was more pronounced at a higher pouring temperature.
(b) Fig. 16. Distribution of the modified structure in the vertical sections of the castings with a bottom flat 5% yellow S-modified core insert, was poured at (a)1350 C (2462F) and (b) 1420C (2588F), respectively. Red lines outline the region with the modified structure.
Fig. 17. Cooling curves are obtained from the bottom and top plates of the casting and from near surface layer of flat S-modified core insert. The arrows indicate the period of possible mold-metal interactions.
COMBINING AN IN-SITU STRUCTURE MODIFICATION WITH LOCAL CHILLING A series of virtual cases were modeled using CFD to prevent liquid metal convection and localize the treated zone in the bottom plate of the casting. Modeling showed
Panel 13-1320.pdf, Page 8 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
that the combination of in-situ S-modified inserts with localized chilling could be used for process control. Based on CFD modeling, the so-called “muffin pan” concept of a chill plate with installed 2 in. (5.08 cm) diameter S-modified insert tablets was proposed (Fig. 18) and tested in the Heat 3b set. The experimental cooling curves obtained from the tablet inserts installed in the chill plate, the bottom plate and the top plate of the casting are shown in Fig. 19. The chilling effect shortened the duration of the sulfur reaction from 100 sec for the unchilled case (Fig. 17) to 50 sec. It also decreased the bottom plate’s solidification time compared to the middle and the top plates, as in Fig. 19. (a)
A series of castings were produced in Heat 3a with variations in yellow S-content and “muffin pan” design. At a high sulfur content (8%), the casting was overtreated. Flake graphite was observed even in the top casting plate (Fig. 20). In this case, mushroom-like modified zones were formed and sulfur gas bubbles transported sulfur in both vertical and horizontal directions. Decreasing the sulfur content in the core inserts to 5% resulted in a milder reaction. The reaction continued to occur, however, around the S-gas bubbles. The bubbles formed the protuberances in the modified zones seen in Figs. 21b and 21c. Many of the S-bubbles had sufficient time to dissolve in the metal, producing localized reaction “islands” as visible in Fig. 21d. The S-bubbles that survived led to gas pores observed in the casting (Fig. 21a). Segregated sulfur, stopped by the central core, led to a flat region of flake graphite below the central core. Vents were drilled through on each of the pockets on the “muffin pan” chill plate to release the gas pressure created by sulfur vaporization in the core inserts. In addition, the sulfur content was further decreased to 2 - 4% to minimize over-treatment. As shown in Fig. 22, the modified structure was localized primarily in the bottom plate. No degenerate graphite was found in the top plate. The last experimental conditions indicated the possibility of controlling layering structure by variations in studied parameters. Improving the layering of the dual-graphite structure and eliminating the porosity will be possible to accomplish in the future by optimizing these process parameters.
(b) Fig. 18. This is a CFD modeling prediction (10 sec after pouring): (a) liquid fraction (in casting volume, not filled image) and (b) temperature distribution in the vertical section of the casting and the bottom positioned “muffin pan” chill plate with S-modified inserts.
Fig. 19. These are the measured cooling curves at different locations in the casting and S-modified insert inside “muffin pan” chill plate. Arrows indicate a starting and ending time of a possible sulfur reaction with the Mg-treated metal.
Panel 13-1320.pdf, Page 9 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
CONCLUSIONS Casting “skin” layers with degenerate spheroidal graphite are considered deleterious to the mechanical properties and machinability of thin-wall ductile iron castings. The effects of CE, pouring temperature, inoculation, and surface roughness, on the ductile iron casting “skin” have been verified in this article. Improving inoculation and decreasing surface roughness by coating no-bake sand molds are possible methods for controlling the thickness of a casting skin in ductile iron casting, specifically castings made in no-bake molds.
Fig. 20. Excessive sulfur in the modified inserts (8%) over-treated the casting. Significant vertical and horizontal gas sulfur transportations were observed.
Conversely, the enhanced interaction kinetics of Mgtreated ductile iron with molds can be applied to produce iron castings with a dual graphite (spheroidal and flake) structure. An in-situ de-nodulizing treatment of ductile iron liquid metal by S-modified no-bake sand inserts was investigated to develop structure modification in local zones. The localized change of the graphite’s shape from spheroidal to flake was successfully achieved through this process. Forming a desired macro-modified layer throughout the casting, however, is a challenge. The parameters affecting this process, including sulfur type and content, coating, pouring temperature, position for Smodified core inserts and chilling effect were designed, modeled and experimentally tested. An observation of the protuberances and localized island-like areas with degenerate graphite shape indicated a “foot print” of sulfur transport. The localization of graphite morphological degeneration can be accomplished by the combination of S-modified core inserts with a chilling effect. Both sulfur gas transport and liquid metal convection, however, were obstacles that actually redistributed the treated metal and produced gas porosity. Improving the dual-graphite structure layering and eliminating the porosity can be achieved in the future by optimizing the process parameters.
Fig. 21. There are specific features of 5% yellow Smodified core inserts—metal interactions: (a) gas bubbles (possible from S2-gas) surrounded by flake graphite; (b and c) protuberances from the bottom surface flake graphite region and (d) localized flake graphite spot in the top plate.
a b
ACKNOWLEDGMENTS
Fig. 22. This graphic shows the (a) the distribution of the modified structure in the vertical sections of castings with a bottom “muffin pan” chill plate with 4% yellow S-modified core inserts and (b) sharp boundary between flake and spherical graphite zones. Black lines indicate boundaries.
This research was sponsored by Benet Laboratories on behalf of the US Army Contracting Command Joint Munitions and Lethality Contracting Center— accomplished under Cooperative Agreement Number W15QKN-11-2-0001. The views and conclusions contained in this document are those of the authors and should not be interpreted as representing the official policies, either expressed or implied, of AFS or Benet Laboratories or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation here in stated. The authors would like to thank US Army ARDEC-Benet laboratories for funding this research. The authors wish to acknowledge technical support and discussion input of George Kokos, James Barlow and Zhiping Lin from Caterpillar. The authors wish to recognize the assistance
Panel 13-1320.pdf, Page 10 of 10 AFS Proceedings 2013 © American Foundry Society, Schaumburg, IL USA
of Mingzhi Xu, Marc Harris, Forrest Huebner, Clinton Ratliff, John Stanek, Jeremy Robinson, Bradley Bromet, Ian Christian and James Smoot for heat pouring and sample preparation. REFERENCES 1.
2.
3.
4.
5.
6.
7.
8.
Lekakh, S., Richards, V., Peaslee, K., “ThermoChemistry of Non-Metallic Inclusions in DI. Proceedings,” The Carl Loper Cast Iron Symposium, Madison, WI (2009). Lekakh, S., Robertson, D., Loper, C., “Thermochemistry of Nodulization and Inoculation of Irons”, American Cast Iron Inoculation Conference Proceedings (2005). Peach, W., Lekakh, S., Richards V., Peaslee, K., Teague, J., “Effects of Near-Surface Metallurgy on Machinability of Gray Cast Iron,” AFS Transactions, vol. 116, paper 08-1 (2008). Stefanescu, D., Wills S., Massone, J., “Quantification of Casting Skin in Ductile and Compacted Graphite Irons and Its Effect on Tensile Properties,” AFS Transactions, vol. 117, pp. 587-606 (2009). Boonmee, S., Stefanescu, D., “The Mechanism of Formation of Casting Skin in CG Iron and its Effect on Tensile Properties,” Key Engineering Materials, vol. 457, pp. 11-16 (2011). Goodrich, G., Lobenhofer, R., “Effect of Cooling Rate on Pearlitic Ductile Iron Mechanical Properties,” AFS Transactions, vol. 115, paper 07045 (2007). Lekakh, S., Qing, J., Richards, V., “Investigation of Cast Iron Processing to Produce Controlled Dual Graphite Structure in Castings,” AFS Transactions, paper 12-024 (2012). Lekakh, S., Qing, J., Richards, V., “Modeling of Melt Mixing Phenomena in Cast Iron with Dual Graphite Structure,” TMS 2012, 141st Annual Meeting and Exhibition, Supplemental Proceedings, vol. 2, pp. 347-354 (2012).
