Control over the Percentage, Shape and Size of the Graphite Particles in Martensitic White Castings Alloyed with Cr, Nb and Mg

This paper presents the results obtained regarding the control by manufacturers of the percentage, shape, and size of the precipitated graphite in the working layer of duplex work-rolls used in hot strip mill finishing stands. This working layer is manufactured in a martensitic white cast iron alloyed with Cr and Nb to promote the precipitation of M3C and MC carbides, which provide a high wear resistance. The thermal cycling behavior of this layer also has a decisive influence on its service life. In this context, the percentage of graphite and its morphology play a very important role against said thermal cycling. With the aim of studying their effect on the sphericity of graphite, the analyzed industrial manufacturing factors worth highlighting include the liquidus temperature, the %Si, the use of an FeSi inoculant with traces of Lanthanum, inoculation with different amounts of FeB and SiCaMn, and the addition of Mg. At the periphery of the working layer, it was found that the use of the FeSi inoculant with traces of La promoted an increase in the density of counts of graphite, and that inoculation with FeB and the addition of 0.02% Mg diminished the nodularity of the graphite. Furthermore, throughout the entire thickness of the working layer, an increase in the amount of SiCaMn of up to 0.6 kg/T produced an increase in the size of the graphite particles and a marked improvement in their nodularity.


Introduction
The failure mechanisms that arise in the surface of indefinite chill double-poured (ICDP) work rolls used in the finishing stands of hot strip mills involve phenomena of plastic deformation, abrasive wear, and cracking resulting from mechanical or thermal stresses [1]. These work rolls are manufactured by means of vertical centrifugal casting. The outer working layer is a martensitic white cast iron with the presence of graphite particles, while the core is a grey cast iron with a pearlitic matrix and dispersed spheroidal graphite. Its main alloying elements are Ni and Cr. The chemical composition may include Nb and Mo so as to improve the wear resistance of the working layer. The Nb forms carbides of the MC type with a hardness close to 2400 HV [2], while the Cr forms carbides of the M 3 C type [3], with a hardness close to 1200 HV [3], and the Ni and the Mo increase the hardenability of the material [4]. Thus, its microstructure will mainly be made up of proeutectic austenite, eutectic austenite, MC and M 3 C carbides, and graphite particles [5]. The austenite will be partially converted to martensite during air cooling after quenching at 1000 • C. During the rolling process, the working layer is heated on entering into contact with the sheet steel to be rolled. This heating is counteracted Table 1 shows the standard chemical composition range for these working layers. The working layer was cast first and then the core was subsequently cast in two stages. In the first stage, an intermediate layer was cast aimed at ensuring optimum binding with the working layer. The remainder of the core was then cast. The working layer was smelted in a medium frequency induction furnace. The inoculants were placed on the bottom of the ladle when "bleeding" the molten iron. This bleeding occurred at a temperature of 1420 • C. The casting was carried out at a temperature of around 60 • C above the liquidus temperature. Demoulding took place 4 or 5 days after casting. After quenching at 1000 • C with air cooling, the roll was subjected to tempering at 400 • C. The experimental procedure employed was based on the Design of Experiments statistical technique [28]. In this case, performing a total of 8 experiments, 6 industrial manufacturing Factors were analyzed, each Factor varying between 2 levels. Table 2 shows the analyzed Factors and levels, while Table 3 displays the Array of Experiments, together with the confounding pattern. These same manufacturing factors were studied in a previous work of the authors, where the resistance of the working layer against mechanical stresses was analyzed, correlating the results with the volume fraction of precipitated carbides [29]. The set of generators associated with this array of experiments is D = AB, E = AC, and F = BC, whose resulting relation definition is I = ABD = ACD = BCF, where I is a column formed only by some (+1) [28]. The resolution of this design is III, i.e., the main effects are confounded with the interactions of two Factors [28]. Table 4 shows the chemical composition of the inoculants employed in this study. Table 5 shows the main casting parameters of the 8 experiments (8 work rolls).   For the analysis of the volume fraction and morphology of the precipitated graphite, samples were obtained from two regions of the working layer. One of them, denominated Zone α, which comprised a thickness of 15 mm from the periphery of said working layer, and another, denominated Zone β, which comprised another 15 mm of thickness from a distance of 25 mm from the periphery up to a distance of 40 mm. The variables used as "responses" to characterize the graphite particles were:

•
The volume fraction of graphite, denoted as Vv.

•
The number of counts per mm 2 of graphite, denoted by N A .

•
The mean Feret diameter, denoted by F mean .

•
The roundness parameter, denoted by R and defined as [perimeter] 2 4π(area) . This parameter defines a perfect nodule when its value is between 1 and 1.4.

•
The percentage of counts of graphite with a roundness parameter equal to or less than 1.4.
These responses were determined using the Image ProPlus software (version 4.5.0.29) (Media Cybernetics, Rockville, MD, USA), together with its Materials-Pro analysis module. For this purpose, 10 micrographs were randomly obtained at 100× in each experiment and in each region (Zones α and β). The metallographic samples were in the polished state, without etching with a chemical reagent. The optical microscope employed was a NIKON Epiphot 200 (Nikon, Tokyo, Japan), the images being obtained using the Omnimet Enterprise Image Analysis System. A JEOL JSM-5600 scanning electron microscope (JEOL, Nieuw-Vennep, The Netherlands), equipped with the characteristic energy dispersive X-ray (EDX, JEOL, Nieuw-Vennep, The Netherlands) microanalysis system, was used to obtain representative micrographs of the general microstructure. A CAMEBAX SX-100 electronic probe (CAMECA, Gennevilliers, France), equipped with BSE detectors (CAMECA, Gennevilliers, France), was used to obtain images of backscattered electrons. In these latter two cases, the samples were in the polished state and were etched with the chemical reagent Nital 4. Figure 1 shows the general microstructure of alloys of this type. The presence of M 3 C and MC carbides and dispersed graphite in an austenitic matrix, which has been transformed into Martensite during its air cooling, can be appreciated. Figure 1a corresponds to a sample from Experiment 2 in an The standardized effects were compared on a normal probability plot using the Statgraphics Plus (version 5.1) program (Statgraphics Technologies, The Plain, VA, USA). The standardized effect is the ratio between the difference in the value of the response and its mean and standard deviation. This represents not only whether the value of the variable is above or below the mean but also how far it deviates from it. Those standardized effects that deviate from the straight line towards the ends on the normal probability plot are significant. Those that deviate to the left indicate that the value of the response increases at their −1 level and, analogously, while those that deviate to the right indicate that the value of the response increases at their +1 level [28]. Tables 6-9 show the mean values obtained for the studied responses and the standardized effects corresponding to the Factors and Interactions indicated in the column denominated "Confounding Pattern". The first row of this column refers to the mean value obtained for each of the analyzed responses considering the eight experiments. Figures 2-6 show the representation of the standardized effects on a normal probabilistic plot for the analyzed responses.   The standardized effects were compared on a normal probability plot using the Statgraphics Plus (version 5.1) program (Statgraphics Technologies, The Plain, VA, USA). The standardized effect is the ratio between the difference in the value of the response and its mean and standard deviation. This represents not only whether the value of the variable is above or below the mean but also how far it deviates from it. Those standardized effects that deviate from the straight line towards the ends on the normal probability plot are significant. Those that deviate to the left indicate that the value of the response increases at their −1 level and, analogously, while those that deviate to the right indicate that the value of the response increases at their +1 level [28]. Tables 6-9 show the mean values obtained for the studied responses and the standardized effects corresponding to the Factors and Interactions indicated in the column denominated "Confounding Pattern". The first row of this column refers to the mean value obtained for each of the analyzed responses considering the eight experiments. Figures 2-6 show the representation of the standardized effects on a normal probabilistic plot for the analyzed responses.       Figure 2 shows the standardized effects on the volume fraction of graphite on a normal probability plot. Figure 2a shows that, in the outermost region of the working layer (Zone α), none of the analyzed Factors has a significant effect on the volume fraction of graphite. However, Figure 2b shows that Factor D (%Si) has a significant effect on the volume fraction of graphite in the inner region of the working layer (Zone β): situating this Factor at its +1 level (1.1-1.15 %Si) leads to an increase in the volume fraction of precipitated graphite.   Figure 3a shows that Factor A (inoculation with FeSi-La) and Factor E (inoculation with SiCaMn) have a significant effect on NA in the outermost region of the working layer: inoculation with FeSi-La and a low level of SiCaMn favor an increase in this variable. However, Figure 3b shows that none of the analyzed Factors has a significant effect on NA in the inner region.  Figure 4 shows the Factors with a significant effect on the size of the counts of graphite, represented by the mean Feret diameter (Fmean). In Figure 4a, it can be observed that Factor E (inoculation with SiCaMn) and Factor B (addition of FeB) significantly influence this variable in Zone α of the working layer: both Factors at their +1 level increase Fmean. It, hence, follows that if Factor E at its −1 level increases the number of counts of graphite, it is because when it is placed at this level, it also reduces the size of said graphite. Interaction AF + BE + CD is also found to have a significant effect. Figure 4b provides a detailed analysis of this interaction, in which it is verified that it is interaction BE that has a significant effect, increasing the variable Fmean when both Factors are at their +1 level. In Figure 4c, it can be observed that Factor F (%Mg), Factor E (SiCaMn), and Factor D (%Si) have a significant effect on Fmean in the inner region of the working layer (Zone β): placing these Factors at their +1 level produces an increase in the value of this variable.      Figure 4 shows the Factors with a significant effect on the size of the counts of graphite, represented by the mean Feret diameter (Fmean). In Figure 4a, it can be observed that Factor E (inoculation with SiCaMn) and Factor B (addition of FeB) significantly influence this variable in Zone α of the working layer: both Factors at their +1 level increase Fmean. It, hence, follows that if Factor E at its −1 level increases the number of counts of graphite, it is because when it is placed at this level, it also reduces the size of said graphite. Interaction AF + BE + CD is also found to have a significant effect. Figure 4b provides a detailed analysis of this interaction, in which it is verified that it is interaction BE that has a significant effect, increasing the variable Fmean when both Factors are at their +1 level. In Figure 4c, it can be observed that Factor F (%Mg), Factor E (SiCaMn), and Factor D (%Si) have a significant effect on Fmean in the inner region of the working layer (Zone β): placing these Factors at their +1 level produces an increase in the value of this variable.     Figure 2a shows that, in the outermost region of the working layer (Zone α), none of the analyzed Factors has a significant effect on the volume fraction of graphite. However, Figure 2b shows that Factor D (%Si) has a significant effect on the volume fraction of graphite in the inner region of the working layer (Zone β): situating this Factor at its +1 level (1.1-1.15 %Si) leads to an increase in the volume fraction of precipitated graphite. Figure 3 shows the Factors with a significant effect on the number of counts of graphite per mm 2 , N A . Figure 3a shows that Factor A (inoculation with FeSi-La) and Factor E (inoculation with SiCaMn) have a significant effect on N A in the outermost region of the working layer: inoculation with FeSi-La and a low level of SiCaMn favor an increase in this variable. However, Figure 3b shows that none of the analyzed Factors has a significant effect on N A in the inner region.

Results
(%Mg) in Zone α of the working layer: if these Factors are respectively situated at their −1, +1, +1 levels, this parameter increases and the nodularity of the graphite decreases. The deleterious effect of the %Mg on the nodularity of the graphite is worth noting. This result could be justified by the fact that the residual Mg associated with the Mg added at Level +1 (0.02%) is insufficient to achieve a nodular geometry, promoting a "star-shaped" geometry that results in high roundness parameters [30]. Several graphite particles with this geometry can be observed in Figure 1a. Figure 5b shows the significant effect of Factor E (SiCaMn) in Zone β of the working layer. As occurred in Zone α of the working layer, if this Factor is situated at its −1 level, the roundness parameter increases. The addition of Mg does not have a significant effect in this region. Furthermore, Figure 6 shows the Factors with a significant effect on the % graphite with a nodular geometry, i.e., with a roundness parameter below 1.4. In Zone α of the working layer, Factor B (FeB) is significant: To increase the % nodular graphite, this Factor must be placed at its −1 Level; see Figure 6a. Similarly, in Zone β of the working layer, the effect of Factor E (SiCaMn) is found to be significant: placing this Factor at its +1 level will increase the percent of particles of graphite with a spheroidal morphology; see Figure 6b.

Conclusions
This paper presents a research method for manufacturers and users of duplex work rolls used in hot strip mill finishing stands whose working layers is manufactured in Ni-hard cast iron alloyed with Nb and Mg that enables them to control the volume fraction and morphology of precipitated graphite. As to the manufacturing Factors and their Levels analyzed in this paper, it is concluded that: 1. Inoculation with SiCaMn has a significant effect on the shape and size of the precipitated graphite. An increase of 0.3 to 0.6 kg/T will lead to: Furthermore, Figure 6 shows the Factors with a significant effect on the % graphite with a nodular geometry, i.e., with a roundness parameter below 1.4. In Zone α of the working layer, Factor B (FeB) is significant: To increase the % nodular graphite, this Factor must be placed at its −1 Level; see Figure 6a. Similarly, in Zone β of the working layer, the effect of Factor E (SiCaMn) is found to be significant: placing this Factor at its +1 level will increase the percent of particles of graphite with a spheroidal morphology; see Figure 6b.

Conclusions
This paper presents a research method for manufacturers and users of duplex work rolls used in hot strip mill finishing stands whose working layers is manufactured in Ni-hard cast iron alloyed with Nb and Mg that enables them to control the volume fraction and morphology of precipitated graphite. As to the manufacturing Factors and their Levels analyzed in this paper, it is concluded that: 1. Inoculation with SiCaMn has a significant effect on the shape and size of the precipitated graphite. An increase of 0.3 to 0.6 kg/T will lead to:  Figure 4 shows the Factors with a significant effect on the size of the counts of graphite, represented by the mean Feret diameter (F mean ). In Figure 4a, it can be observed that Factor E (inoculation with SiCaMn) and Factor B (addition of FeB) significantly influence this variable in Zone α of the working layer: both Factors at their +1 level increase F mean . It, hence, follows that if Factor E at its −1 level increases the number of counts of graphite, it is because when it is placed at this level, it also reduces the size of said graphite. Interaction AF + BE + CD is also found to have a significant effect. Figure 4b provides a detailed analysis of this interaction, in which it is verified that it is interaction BE that has a significant effect, increasing the variable F mean when both Factors are at their +1 level. In Figure 4c, it can be observed that Factor F (%Mg), Factor E (SiCaMn), and Factor D (%Si) have a significant effect on F mean in the inner region of the working layer (Zone β): placing these Factors at their +1 level produces an increase in the value of this variable. Figure 5 shows the Factors with a significant effect on the roundness parameter of the particles of graphite. Figure 5a shows the significant effect of Factor E (SiCaMn), Factor B (FeB), and Factor F (%Mg) in Zone α of the working layer: if these Factors are respectively situated at their −1, +1, +1 levels, this parameter increases and the nodularity of the graphite decreases. The deleterious effect of the %Mg on the nodularity of the graphite is worth noting. This result could be justified by the fact that the residual Mg associated with the Mg added at Level +1 (0.02%) is insufficient to achieve a nodular geometry, promoting a "star-shaped" geometry that results in high roundness parameters [30]. Several graphite particles with this geometry can be observed in Figure 1a. Figure 5b shows the significant effect of Factor E (SiCaMn) in Zone β of the working layer. As occurred in Zone α of the working layer, if this Factor is situated at its −1 level, the roundness parameter increases. The addition of Mg does not have a significant effect in this region. Furthermore, Figure 6 shows the Factors with a significant effect on the % graphite with a nodular geometry, i.e., with a roundness parameter below 1.4. In Zone α of the working layer, Factor B (FeB) is