Primary Austenite Morphology and Tensile Strength in CGI for Different C Contents, Cooling Conditions and Nodularity

Abstract

Compacted graphite iron (CGI) is a good option for the blocks and cylinder heads in heavy duty engines due to their well-balanced thermal and mechanical properties. In this work, a remelting technique has been utilized for the production of CGI with different nodularity (10 and 20%), C contents (CE=3.5, 3.8, 4.2) and under different solidification and cooling rates. The employed experimental parameters had a sizeable influence on the morphology and fraction of the inter-dendritic structure and resulted in ultimate tensile strength (UTS) that ranged from 335 to 456 MPa and 371 to 521 MPa for the 10 and 20% nodularity, respectively. The results show that the UTS is linearly related to the solidification time and the microstructural parameter that express the scale length of the inter-dendritic region. Different CE and nodularity provide different relationships between UTS, solidification time and microstructure. Finally, an empirical model has been developed for the prediction of the UTS.

Introduction

Compacted graphite iron (CGI) is used for the production of cylinder blocks and heads in heavy duty engines due to it’s well-balanced thermal and mechanical properties. Improved performance and fuel economy can be obtained by increased pressure in the combustion chamber. The increased pressure is accompanied by higher thermal and mechanical loads that CGI must withstand. The understanding of microstructure–property relations in CGI is essential for the prediction of ultimate tensile strength (UTS) aimed to assist the development of CGI components with enhanced strength.

Carbon content and solidification rates have a great influence on cast iron UTS. Decreased carbon content and increased solidification rates result in improved UTS.¹ The graphite cracking or delamination from the matrix is the main source of strength reduction in cast iron alloys.^2,3 For this reason, the maximum graphite length has been considered as the maximum crack length in the modified Griffith equation for the prediction of UTS in fully pearlitic lamellar graphite iron (LGI).^4,5,6 However, has been shown,⁷ that the modified Griffith equation is resulted directly from the empirical relationship between D^Hyd_IR and UTS, indicating that D^Hyd_IR is the microstructure parameter that represents best the physical notion of the maximum crack length in LGI. Further, the obtained empirical equation relating UTS to D^Hyd_IR has been implemented in a full-scale casting simulation based on kinetic models.⁸

In this work, a range of solidification time and carbon content have been utilized in a remelting experiment for the production of CGI with different microstructure morphologies. The aim of this work is to study the influence of graphite nodularity and inter-dendritic morphology on the UTS of CGI. This model can be incorporated in a full-scale casting simulation, similar to the one developed for UTS prediction in CGI.⁸

Experimental Procedure

CGI samples with controlled nodularity were produced by a remelting experimental technique. The remelting experiment is described in more detail in.⁹ Briefly, several SGI cylindrical samples with dimensions Ø38×42 mm were melted in an experimental device and held at 1450°C under argon atmosphere, for a certain period of time aimed to produce samples with 10 and 20 % nodularity. The holding time, representing the fading time of Mg, is the single experimental variable that is modified to provide different nodularity. After the holding time was completed, the samples were solidified under different cooling rates aimed to produce samples with different microstructure coarseness. The cooling rates range between 0.6 and 4.8 K/min. Further, three different carbon contents were employed in this work; see Table 1.

Table 1 Chemical Composition (wt%) for the Remelting Experiments (CE=%C+%Si/3)

Flat miniaturized tensile bars with 15 mm gauge length and 2 mm thickness were used for the tensile test. The tensile bars were machined at a distance 9 mm from the center of the cylindrical samples. Miniaturized tensile tests were performed at room temperature in a Kammrath and Weiss testing module with a strain rate of 10 μm/s and with four specimens from each condition.

After the tensile test, the samples were etched in a picric acid-based solution to reveal the dendritic microstructure. The etching reagent contains picric acid, NaOH, KOH and distilled water at a weight ratio of 1:1:4:5, respectively. The etching was performed at 110 °C. Figure 1a shows an overview of etched microstructure in a tensile bar. Optical microscopy is utilized for the estimation of the hydraulic diameter of the inter-dendritic region (D^Hyd_IR). The microstructure feature D^Hyd_IR expresses the length scale of the inter-dendritic region and is estimated by the fraction of the inter-dendritic area (A_IR) over the perimeter of the dendrite–eutectic interface (P_γ).¹⁰ The primary dendritic structure was selected in the image manipulation software GIMP, and the microstructural parameters A_IR and P_γ were estimated from the binary image. Figure 1b presents a typical micrograph of the analyzed microstructure, and Fig. 1c shows the binary image where the dark areas are primary dendrites. A_IR is the white area, and P_γ is the periphery of the dark areas in Fig. 1c. The image analysis was performed with the open-source software ImageJ.

Figure 1

Overview of etched microstructure in a tensile bar (a) micrograph for microstructure analysis (b) binary image for microstructural analysis (c).

Results and Discussion

Typical micrographs of alloy 3 are shown in Figs. 2 and 3 for the samples with 10% nodularity. As can been seen, the experimental cooling conditions resulted in a sizeable variation of the microstructure coarseness. The samples that solidify with the fast-cooling rate have smaller eutectic grains and finer primary austenite dendritic structure. As the cooling rate is decreasing, both dendritic and eutectic microstructure becomes coarser. Further, the eutectic grains diameter decreases and the fraction of the dendrites increases as the CE decreases from 4.2 to 3.8.

Figure 2

Etched microstructure from different cooling rates. CE=3.8.

Figure 3

Etched microstructure from different cooling rates. CE=4.2.

Figure 4a shows the UTS as a function of solidification time for the 10% nodularity samples and different CE. As can be seen, UTS is increasing by decreasing solidification time and CE. Further, there are different linear relationships between solidification time and UTS for each CE. Similar relationships between UTS solidification time and CE have been observed in LGI.¹ The increase of D^Hyd_IR is a result of longer solidification time and higher carbon content. The UTS for 10 and 20% nodularity is plotted against D^Hyd_IR in Fig. 4b. In the same plot are presented results for 3% nodularity reported in.¹¹ As can be seen, the UTS increases with the reduction of D^Hyd_IR. Furthermore, the linear relationship between UTS and D^Hyd_IR differs with the nodularity.

The linear correlation between UTS and both D^Hyd_IR and solidification time appears to be different from the findings of previous investigations. Previous studies in CGI suggested that D^Hyd_IR coarsens is proportional to the cube root of time⁹ and the UTS of pearlitic LGI is proportional to the inverse square root of D^Hyd_IR.⁷ These discrepancies can be explained by the small number of data points for the UTS and solidification time of each CE and the fact that the relation between UTS and the inverse square root of D^Hyd_IR has been shown to be valid for fully pearlitic LGI. The UTS of CGI is influenced by different microstructure features such as nodularity and ferrite/pearlite ratio in addition to the D^Hyd_IR.

The majority of the experimental data from Fig. 4b were used to develop an empirical model (Eqn. 1) that expresses the UTS as a function of D^Hyd_IR and nodularity. The relationship between estimated and experimental UTS has a Pearson correlation coefficient (r) equal to 0.94 and a mean absolute percentage error (MAPE) equal to 5%.

Figure 4

UTS as a function of (a) solidification time and (b) hydraulic diameter of the inter-dendritic region.

$${\sigma }_{\mathrm{UTS}}=356-1.8\cdot {D}_{\mathrm{IR}}^{\mathrm{Hyd}}+8.9\cdot \mathrm{Nodularity}$$

(1)

The comparison between the calculated and the experimental UTS is demonstrated in Fig. 5. The graph data indicate a good correlation between experimental and calculated UTS.

Figure 5

Correlation between calculated and experimental UTS.

The derived model can be incorporated in a full-scale casting simulation for the UTS prediction in CGI. The casting simulation methodology, similar to the one developed in⁸ for LGI, must be capable of predicting the microstructural parameters of interest.

Conclusions

The present work utilized a remelting experimental technique to study the influence of the carbon content and solidification time on the primary austenite morphology and UTS of CGI. The results show that the UTS is linearly related to the solidification time and D^Hyd_IR. Different CE and nodularity provide different relationships between the UTS and D^Hyd_IR. The studied microstructural parameters have been incorporated in an empirical model for the prediction of the UTS.