S and O have a high capacity to react with nodularizing elements, while the resulting products could be useful as nucleation sites for graphite. Any consumption of these elements could change a nodular to compacted or just lamellar morphologies. The main objective is to evaluate the effects of inoculation and S- or O-addition after Mg-treatment and inoculation on the sensitivity of high Si ductile iron, solidified in mold media without S or O contribution, to graphite degeneration in the surface layer compared to the body of the casting. Mg-treated and ladle-inoculated cast iron is used, containing 0.031%Mgres, 3.37%C, 3.44%Si, 0.44%Mn, for 4.43% carbon equivalent. FeS2 or Fe2O3 powder is placed on the bottom of a standard ceramic cup, used in thermal analysis (7.3 mm cooling modulus). Structure characteristics are analyzed from the section of the ceramic cup samples. The skin thickness has different values, depending on the graphite nodularity or shape factors or matrix evaluation results. Inoculation has a beneficial effect on the graphite phase characteristics: with only 3% higher nodularity in the casting body, this treatment is very efficient to decrease the skin effect thickness. A stronger degenerative effect of S on the graphite morphology was found, compared to the stoichiometric equivalent O-addition after inoculation, not only in the casting but also as the thickness of the skin. At higher S- or O-addition, the casting skin thickness increased, but at different levels: from two to four times higher after S-addition, and up to 30% higher after O-addition.
Conversely, sulfur and oxygen appear to have an important role in graphite nucleation in both gray and ductile cast irons. It was found that with both graphite morphologies, complex compounds act as nucleation sites in commercial cast irons, in a general three-stage graphite formation, but in different sequences: oxide–sulfide—lamellar graphite (gray iron) and sulfide–oxide—nodular graphite (ductile iron), respectively.1
(Mn,X)S compounds (where X = Fe, Al, O, Si, Ca, Sr, Ti, etc.) are major sites for graphite nucleation in gray cast irons.1,2,3,4,5,6,7,8 According to Jacobs et al.,9 graphite nodules nucleate heterogeneously on particles formed in the melt, and exhibited a duplex sulfide [(Ca,Mg)S, (Sr,Ca,Mg)S, central seed]/oxide [outer shell, (Mg,Al,Si,Ti)Oxide, spinel] structure. Later, T. Skaland10 found complex hexagonal silicate phases of XO-SiO2 or XO-Al2O3–2SiO2 [X = Ca, Ba, Sr], as contributors to inoculation, on the surface of the previously formed Mg-silicates, after Mg-treatment, making them more favorable sites for subsequent graphite nucleation. For < 0.005%S in a base cast iron, complex nitrides mainly exist at the nodular nucleation sites, while if the sulfur content increases above 0.005% effective nodular graphite nuclei are sulfides.11,12,13,14 Simple silicates were observed in the matrix, while more complex silicates, such as Al, Ca, Ce, La, were found in conjunction with graphite that probably acted as graphite nucleation sites.15
Nodularizing elements consumed in a reaction with sulfur and oxygen during Mg-treatment leads to decreasing spheroidizing potential, resulting in a transition from nodular/spheroidal morphology to vermicular/compacted shapes to coral and lamellar morphologies. It was found that controlled addition of sulfur after Mg-treatment could be a solution to produce compacted graphite cast iron in foundry conditions.16,17,18,19
S and O present in the core and mold materials, or in the coating of core and mold (to improve the casting surface quality), have an additional contribution to graphite degeneration, mainly at lower contents of residual magnesium. It was found that iron-nodularizing potential, molding systems (with and without S), and the mold coatings (with included S-content, or conversely with a desulfurization capability) are important factors for graphite degeneration in the surface layer of castings. Sulfur is the major actor in the metal–mold interaction, resulting in graphite degeneration at the casting surface.20,21,22,23,24,25
The main objective of the present paper is to evaluate the effects of inoculation and S- or O-addition after Mg-treatment and inoculation on the sensitivity of high Si ductile cast iron, solidified in mold media without S or O contribution, to graphite degeneration in the surface layer compared to the casting body.
Table 1 shows the representative parameters of the experimental procedure. Electrically melted cast iron was subjected to Mg-treatment by tundish-cover technique and inoculation during the iron transfer from Mg-treated ladle to pouring ladle. A commercial Mg-treatment master alloy and Ca, Ba-bearing FeSi inoculant were used. Mg-treated and double-treated (nodularized and inoculated) cast irons were solidified in phenol-formaldehyde resin-coated sand (Croning process) mold (no S-content), typically ceramic cups used for thermal (cooling curve) analysis. FeS2 or Fe2O3 in a powder state was added, at two levels in stoichiometric equivalence for MgS and MgO formation, on the bottom of the ceramic cups (on a polystyrene plate), before pouring the inoculated cast iron. No flotation phenomena were observed.26
For structure analysis, in the surface layer and casting body, samples obtained in the thermal analysis ceramic cup were used. Structure was analyzed in the as-cast state, without etching (graphite analysis) and after nital etching (matrix analysis). To evaluate the structure characteristics in the casting body, three analysis directions were considered, each one with five analysis points, placed in the same positions, outside of the surface layer and casting center, to avoid their influence.
The formation of the surface layer (skin) was evaluated as the average thickness and structure characteristics. In both cases, the end effect was avoided, so 22 mm analysis length was used on each side (Figure 1). For each side, the layer thickness was measured in points at 100 μm apart and the structure characteristics at 1.0 mm [total 20 measurements], respectively. The final result is expressed as the average of all of measurements, for each parameter.
Results and Discussion
The chemical composition of Mg-treated and inoculated cast irons, as base and minor elements, was evaluated by the use of a performance spectrometer (SPECTROLAB M 10, Hybrid optic) capable of determining a high number of chemical elements. The experimental cast iron is characterized by high silicon content (3.44%Si) and is hyper-eutectic, with a carbon equivalent CE = 4.43%, at a normal carbon content. The medium manganese content (0.44%Mn) is typical for a pearlitic–ferritic ductile cast iron, without silicon as the alloying element. The active elements (wt%: 0.031 Mg, 0.007Ce, 0.004La, 0.013Ca) content characterizes a ductile cast iron, at the lower limit of nodularizing potential (Table 2). Minor elements, at low content (wt%: 0.075Cr, 0.05Ni, 0.01Mo, 0.067Cu, 0.0058Co, 0.0039 V, 0.0062As, < 0.0008Te, 0.0002Zn, 0.0053Sn, 0.0085 N, 0.0067Ti, 0.0013B, < 0.0002Pb, 0.0015Sb, < 0.0002Bi), have a low antinodularizing effect, expressed by the representative factor K (< 0.15).27 Pearlite formation sensitivity, characterized by Px factor27 is also influenced less by minor elements, with manganese and silicon contents having a dominant effect (Table 3).
Casting Body Structure Characteristics
The representative structure, as graphite and metal matrix, was evaluated (Figure 2). In the casting body, the Mg-treated and un-inoculated cast iron are characterized by a pearlitic–ferritic matrix, which mainly becomes a ferritic structure after inoculation. The graphite phase is characterized by different shape factors.
Convexity (Cv): the ratio of the square root of the convex perimeter (Pc) to the real perimeter (Pr) of the measured particle.
Elongation (A): The ratio of the maximum diameter (Dmax) to the equivalent rectangle shortest side (a) [the rectangle which has the same area and perimeter as the particle].
The area (AG), real (Pr) and convex (Pc) perimeter, maximum size (Dmax) and the maximum ferret (Fmax) of particles were considered.
The graphite phase characteristics (Figure 2) were assessed using Automatic Image Analysis [analySIS® FIVE Digital Imaging Solutions software] (particles larger than 5 μm, as maximum size). The criteria used to consider a graphite particle as a spheroid or a compacted particle are defined in terms of sphericity, roundness, convexity and elongation parameters.
Transition from lamellar (flake) morphology through vermicular (compacted) form up to nodular (spheroidal) morphology means increasing sphericity/circularity (SSF), roundness (RSF) and convexity (Cv) up to 1.0 level, as the maximum value for a spherical particle, while elongation (A) parameter decreased to 1.0 level, respectively. The higher the degree of graphite compactness [closer to a spherical shape], the greater all mechanical properties, especially elongation and impact strength values of ductile cast irons, for a given metal matrix.
A previous paper28 determined that in silicon-alloyed ductile cast iron, the sphericity/circularity shape factor, involving area (AG) and the real perimeter (Pr) of the measured particle, is more sensitive to contour irregularities, as a typical effect of silicon alloying. If the object contour is irregular, the perimeter can increase even if the overall shape is still close to a circle. A lower level of this shape factor typically translates into decreased mechanical properties, especially ductility and toughness, for a given matrix structure.
Inoculation increased graphite nodularity by 3% (Figure 3) and improved the graphite shape factors (Figure 4), at higher power for SSF, RSF, and Cv, and lower level of A, respectively. Active elements such as Ca, Ba, and Al protect Mg in its reaction with sulfur and oxygen, but there is limited nodularizing potential of Ca and Ba, so it is their role in an inoculation treatment that usually increases the quality of graphite phase.
It appears that the beneficial effects of inoculation treatment on the graphite phase characteristics, already known in conventional, un-alloyed ductile cast irons, are also active in silicon-alloyed materials. As sphericity/circularity (SSF) shape factor is less beneficially affected, it means that silicon-alloyed ductile cast iron is sensitive to spheroidal morphology changes, including lower silicon alloying grade (cca 3.5%Si). Graphite nodularity (NG) is defined by Eqn. 3:
where Aall particles—area of all graphite particles; Aparticles (RSF)—area of particles for a specific roundness shape factor (RSF).
Oxygen and sulfur addition after inoculation affected the cast iron structure, mainly as graphite phase characteristics, and, consequently, also the ferrite/pearlite ratio. Although these two elements were added in stoichiometric equivalence to form Mg—oxide and sulfide, their effect on the structure was different. Oxygen addition after Mg-treatment and inoculation has a limited effect as graphite nodularity decreases 2.5–3.5%. The treated cast irons remain in ductile (nodular graphite) iron family, but at lower limit of usually accepted graphite nodularity (cca. 70%).
In the same solidification conditions, sulfur addition drastically affected graphite nodularity, ten times more than a stoichiometric addition of oxygen, as nodularity decreased by 26–28%: from more than 70% nodularity in un-treated cast iron to less than 46% nodularity after sulfur addition (higher sulfur addition, lower graphite nodularity). The specific effect of added sulfur in decreasing graphite nodularity is 5–10%NG/0.01% S(add), as in MgS formation, 0.01%S combines with 0.0074%Mg.
As Figure 2 shows, added sulfur contributed to graphite morphology changing to vermicular/compacted morphology, without lamellar or degenerated graphite forms occurring. The sulfur effect is stronger than oxygen’s capacity to produce this transition. This behavior of sulfur addition after magnesium treatment could be applied in an un-conventional way to produce vermicular/compacted graphite cast iron, with good results in an industrial application.16,17,18,19
Graphite shape factors are also affected by oxygen or sulfur addition in inoculated ductile cast irons. In all cases, oxygen addition has a limited negative effect compared to sulfur, and also graphite nodularity is similarly affected. It appears that the sphericity/circularity shape factor (SSF), involving the real perimeter of graphite particles, is more visibly influenced compared to roundness (RSF), which depends on the area (AG) and the maximum ferret [Fmax] of the object (graphite particle). Fmax is the maximum distance of parallel tangents at opposing measured particle borders. A perfect circle assumes a value of RSF = 1, but it decreases when Fmax increases against the area.
Surface Layer (Skin) Structure Characteristics
The experimental program included a ceramic mold (coated resin sand), which does not contribute sulfur or oxygen, thereby avoiding the migration of these elements into the molten iron. These two elements are known to be the major factors in an interaction between liquid metal and mold media, to cause graphite degeneration in the surface layer of castings, before solidification.
Figure 2 shows a typical structure at the surface of the test castings, with and without inoculation and with or without addition of sulfur or oxygen in the final iron [Mg-treated + Inoculation]. Both graphite in un-etched sample and the metal matrix after a nital etch is shown. A typical casting surface layer, usually known as “skin,” is formed in all cases, but at different thicknesses [0.1–1.0 mm/100–1000 μm], representing 0.3–3.0% by comparing to the casting wall thickness [32 mm]. As Figure 5 shows, there is a relationship between the measurement values obtained in the un-etched condition for the graphite phase only and the metal matrix after a nital etch in the same sample. The values for the influence of the metal matrix are higher. The highest skin thickness is characteristic of the un-inoculated ductile iron casting, while the lowest level was obtained after inoculation, for both evaluation as graphite or metal matrix, respectively.
Sulfur or oxygen additions negatively affected the beneficial effect of inoculation, more noticeably for the action of sulfur action, very close to its elimination (Figure 6). The oxygen effect was very low. Although inoculation increased graphite nodularity in the casting body by only 3%, inoculation appears to be very efficient for reducing the tendency of Mg-treated cast iron to lose graphite compactness in the surface layer (Figure 7). The addition of oxygen inoculated ductile cast iron caused a limited decrease in the graphite nodularity in the casting body, which corresponds with a very low increase in the skin thickness. In contrast, a sulfur addition in the same conditions changed the ductile (nodular graphite) iron casting to a compacted graphite iron casting (at high nodularity type). As a result, the skin thickness is increased considerably.
In the present experiments, the mold media does not contain free oxygen or sulfur to migrate into the molten, and also no coatings were applied to the mold, to avoid influencing the metal–mold interaction, either by contribution of these active elements or by inhibiting their transfer from mold to liquid metal. Despite these conditions, the experimental casting displayed a degenerated graphite surface layer of limited thickness. Previous experiments22 demonstrated that a sulfur contribution, originating in the P-Toluenesulfonic Acid (PTSA) component of a furan resin sand—PTSA binder system, increased the skin thickness from 5 to 10 times compared to a coated resin sand mold, without any sulfur contribution.
In order to have a better evaluation on the thickness of the surface layer (skin) with degenerated graphite morphologies, the variation of the graphite nodularity and representative graphite shape factors were analyzed from the surface to the casting body where the graphite phase was unaffected, typically a level from the surface up the casting body (without graphite phase affectation), typically a 2 mm distance. Figure 8 shows the scheme of the casting skin thickness (δ) measurement by this way. The same technique was applied for all of the considered graphite parameters.26
The graphs included in Figure 9 show the variation of the graphite phase characteristics, as nodularity and considered shape factors on the section of the test castings, in Mg-treated cast irons, without inoculation and with inoculation, and without or with sulfur or oxygen addition after inoculation. The difference between these iron treatment variants is apparent for all of the considered evaluation parameters, considerably so at the surface of castings and up to 0.5–1.2 mm across the section. Generally, the lowest graphite phase compactness degree was obtained just at the casting surface, as NG = 5–25%, SSF = 0.3–0.5, RSF = 0.2–0.5, Cv = 0.6–0.8 and A = 3–5, typical for compacted graphite cast iron characteristics. These parameters are improved further into the casting, but depending on the applied treatment: NG = 45–75%, SSF = 0.6–0.8, RSF = 0.45–0.6, Cv = 0.8–0.9, A = 1.5–2.5. This evolution of the graphite phase parameters represents an increasing nodular/vermicular (compacted) graphite ratio.
The graphite characteristics variation is contrary to variation of the solidification cooling rate, which is typically higher at the metal–mold interface and lower to the casting center. In Mg-treated iron castings, increasing the solidification cooling rate generally leads to improved or greater graphite nodularity and an improved graphite particles compactness degree, as the graphite morphologies transition from compacted to nodular. The structure anomaly revealed here in these experiments is the result of graphite degeneration at the surface layer due to chemical reactions, which consume and reduce the availability of the nodularizing elements.
Graphite nodularity parameter appears to be the most affected by the chosen treatment to the iron melt. According to Figure 9a, the effects of two groups of treatments are noticeable in the section at the surface layer: Inoculation and oxygen addition after inoculation improve nodularity. Conversely an un-inoculated iron, or an inoculated iron with sulfur addition after inoculation, displays reduced graphite nodularity. The difference between these two groups of treatments (inoculation and O-addition after inoculation vs. un-inoculation and S-addition after inoculation) persists into the casting section, but to a lesser extent: difference from 10 to 70% nodularity in the casting skin and only from 50 to 70% nodularity in the casting section, respectively.
Of the considered shape factors, sphericity (circularity), involving the real perimeter of analyzed graphite particles, appears to be the most affected by the absence of inoculation or particularly by a sulfur addition after inoculation (Figure 9b). The obtained results confirm the data reported in a previous paper,28 as for silicon-alloyed ductile cast iron, this shape factor is more sensitive to graphite morphology changes, compared to the roundness shape factor (RSF) (Figure 9c), which involves the maximum ferret parameter of a graphite particle, and is usually used in graphite phase characterization, according to ISO 945. The convexity shape factor (Figure 9d) is less sensitive to graphite morphology variation across the iron casting section.
As the graphs in Figure 9 show, the thickness of the surface layer (skin) depends not only on the applied treatments on the Mg-treated cast irons, but also on the considered factors, such as graphite nodularity or different graphite phase shape factors. A summary of these effects is illustrated in Figure 9, which considers the casting skin thickness depending on the graphite characteristics [un-etched samples]. Generally, the affected skin thickness is greater by using the graphite phase parameters, compared to direct measurement values, and especially if nodularity [NG], sphericity (SSF] or roundness [RSF] is applied. The lowest skin thickness measurement resulted from employing the convexity (Cv) and elongation (A) parameters.
Inoculation increased graphite nodularity by 3% and improved all of the graphite shape factors in the casting body area.
In the casting body, an O-addition has a limited effect on decreasing graphite nodularity (by 2.5–3.5%).
S-addition drastically affected the graphite nodularity in the casting body [by 26–28%, 5–10%NG/0.01% S(add)], ten times more than a similar stoichiometric addition of oxygen, and it caused more compacted graphite without lamellar graphite formation.
Graphite shape factors are also affected by O/S-addition in inoculated irons, in the casting body, especially as sphericity shape factor (SSF). In all cases, an O-addition has a limited negative effect compared to an S-addition, including their effects on the graphite nodularity.
A typical casting surface layer (skin), at 0.1–1.0 mm thickness (0.3–3.0% of casting wall thickness), was found, with consistently higher values for the metal matrix, in comparison with the graphite phase, but in a good correlation.
The highest skin thickness characterizes the un-inoculated iron, while the lowest skin thickness was after inoculation, for both graphite and metal matrix evaluations. Although inoculation increased graphite nodularity in the casting body by only 3%, this treatment appears to be very efficient to control the tendency for skin formation.
S/O-addition decreased the beneficial effect of inoculation, more apparent for the S-action, very close to its elimination. An O-addition led to a limited decrease in graphite nodularity in the casting body, which corresponds with very low increase in the skin thickness. In contrast, an S-addition altered the ductile iron to a compacted graphite iron structure and, as a result, the skin thickness increased considerably.
Generally, the resulting skin thickness is higher by the use of graphite phase parameters, compared to a traditional direct measurement value on the casting section. The obtained values are higher if nodularity [NG], sphericity (SSF] or roundness [RSF] are applied, compared to convexity (Cv) and elongation (A) parameters use.
Graphite nodularity and sphericity (circularity) shape factors appear to be the most affected by no inoculation or, especially, if S-addition is applied after inoculation.
In general, there is the lowest sensitivity for inoculated iron (with or without O-addition), a medium response for un-inoculated and a low S-addition, and the highest degeneration level for the graphite phase with a higher S-addition after inoculation.
I. Riposan, T. Skaland, Modification and inoculation of cast iron, in ASM, Cast Iron Science and Technology Handbook, ed. by D.M. Stefanescu, pp 160–176 (2017)
I. Riposan, M. Chisamera, S. Stan, T. Skaland, M.I. Onsoien, AFS Trans. 109, 1151–1162 (2001)
I. Riposan, M. Chisamera, S. Stan, C. Hartung, D. White, Mater. Sci. Technol. 26(12), 1439–1447 (2010)
I. Riposan, M. Chisamera, S. Stan, Mater. Sci. Forum 925, 3–11 (2018)
A. Sommerfeld, B. Tonn, Int. J. Metalcasting 3(4), 39–47 (2009)
B.B. Gundlach, AFS Metalcasting Congr., Atlanta, GA, USA, Paper 08-158 (2008)
J.A. Campbell, Met. Mater. Trans. B. 40(6), 786–801 (2009)
G. Alonso, D.M. Stefanescu, M. Larranaga, E. De la Fuente, R. Suarez, AFS Trans. 124, 124–134 (2016)
M.H. Jacobs, T.J. Law, D.A. Melford, M.J. Stowell, Metals Technol., 490–500 (1974); 98–108 (1976)
T. Skaland, Nucleation Mechanism in Ductile Iron. Doctoral Thesis, Metallurgisk Institutt, Trondheim, Norway (1992)
J.K. Solberg, M.I. Onsoien, Mater. Sci. Technol. 17(10), 1238–1242 (2001)
V. Igoraski, S. Okade, Int. J. Cast Metal Res. 11(2), 83–88 (1998)
H. Nakae, Y. Igarashi, Mater. Trans. 43(11), 2826–2831 (2002)
G. Alonso, D.M. Stefanescu, P. Larranaga, R. Suarez, E. De la Fuente, AFS Trans., 125, Paper 17-031 (2017)
I. Riposan, M. Chisamera, S. Stan, P. Toboc, G. Grasmo, D. White, C. Ecob, C. Hartung, J. Mater. Eng. Perform. 20(1), 57–64 (2011)
I. Riposan, M. Chisamera, Giesserei Praxis, Issue 9/10, 155–162 (1991) and Casting, Forging and Heat Treatment (Japan), Issue 547, 17–23 (1993)
M. Chisamera, I. Riposan, S. Stan, M. Barstow, D. Kelly, R.L. Naro, AFS Trans. 104, 581–588 (1996)
M. Chisamera, I. Riposan, S. Stan, M. Barstow, D. Kelly, R.L. Naro, AFS Trans. 110, 851–860 (2002)
M. Chisamera, I. Riposan, S. Stan, M. Barstow, D. Kelly, R.L. Naro, AFS Trans. 111, 869–883 (2003)
I. Riposan, M. Chisamera, S. Stan, Int. J. Metalcasting 7(1), 9–20 (2013)
N. Ivan, M. Chisamera, I. Riposan, Int. J. Metalcasting 6(4), 61–69 (2012)
N. Ivan, M. Chisamera, I. Riposan, S. Stan, AFS Trans. 121, 379–390 (2013)
R. Danko, M. Gorny, M. Holtzer, S. Zymankowska-Kumon, ISIJ Int. 54(6), 1288–1293 (2014)
S. Boonmme, D.M. Stefanescu, AFS Trans. 121, 449–459 (2013)
J. Qing, S. Lekakh, V. Richards, AFS Trans. 121, 13–120 (2013)
D. Anca, Research on the loss modifying effect in the superficial layer phenomenon of the iron castings. Ph.D. Thesis, POLITEHNICA University of Bucharest, Romania, 2014
T. Thielemann, Giessereitechnik, Issue 1, 16–24 (1970)
S. Stan, I. Riposan, M. Chisamera, I. Stan, J. Mater. Eng. Perform. 28(1), 278–286 (2019)
The work has been funded by the Sectoral Operational Programme Human Resources Development 2007–2013 of the Romanian Ministry of Labour, Family and Social Protection through the Financial Agreement POSDRU/107/1.5/S/76903. The authors would like to recognize and thank Michael Barstow (Consultant) for reviewing and editing this paper.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
This paper is an invited submission to IJMC selected from presentations at the 2nd Carl Loper 2019 Cast Iron Symposium held September 30 to October 1, 2019, in Bilbao, Spain.
About this article
Cite this article
Anca, D., Chisamera, M., Stan, S. et al. Graphite Degeneration in High Si, Mg-Treated Iron Castings: Sulfur and Oxygen Addition Effects. Inter Metalcast 14, 663–671 (2020). https://doi.org/10.1007/s40962-019-00385-2
- ductile cast iron
- graphite degeneration
- casting skin
- shape factors