The microstructures, mechanical properties, and temperature distributions in nodular cast iron friction-welded joint

The paper focuses on an experimental topic relating to the field of friction welding process of a nodular cast iron. The microstructures, phase transformation, temperature distributions, microhardness, and tensile test are all studied within the framework of the paper in question. Maximum temperature measurements in the axial center and periphery of the analyzed joints were equal to 950 and 840 °C, respectively. Both temperature and increasing temperature gradient at the axial center were higher than those at the periphery. The maximum tensile strength of the examined friction-welded nodular cast iron joints was 53% of that of the parent metal. The welding region was composed of deformed graphite nodules, coarse pearlite, proeutectoid ferrite, and acicular martensite. Highly deformed graphite nodules were distributed along the weld interface due to the material flow in the thermo-mechanically affected zone (TMAZ). In the central zone, graphite displayed a striped configuration and ferrite transformed into a martensite structure. In the peripheral region, graphite surrounded by martensite kept the form of individual granules. Maximum hardness at the interface in the TMAZ and the heat-affected zone reached 603 HV and 345 HV, respectively. The executed microstructure analysis showed that the cracks started occurring mostly at the interface of the deformed graphite nodules and then spread through the grain boundaries of metal matrix. The fracture surface appearance showed a cleavage fracture in the peripheral zone and a little dimple fracture around graphite nodules in the central zone.


Introduction
Nodular cast iron (also called ductile or spheroidal graphite cast iron) is much stronger and has a notably higher elongation than other grade cast irons. A relatively high strength and toughness of nodular cast iron are advantageous when it comes to numerous structural applications such as crankshafts, steering knuckles, brackets, valves, truck axles, hubs, water pipes, and many others [1].
Mechanical properties of nodular cast irons are directly associated with their metal matrix microstructure. The ascast matrix microstructure of nodular cast irons may be completely ferritic, completely pearlitic, or be a combination of pearlite and ferrite, with a spheroidal graphite distributed in the matrix. The microstructural properties are affected by the solidification-cooling rate associated with the section size of castings, as well as that of alloying elements [2]. Many techniques and special materials are available that allow for repair welding of nodular cast iron castings, for joining nodular cast iron to itself, as well as to other ferrous materials. The aforementioned welding methods include shielded metal arc welding, flux cored arc welding, gas metal arc welding, gas tungsten arc welding, submerged arc welding, diffusion bonding, impact-electric current discharge joining, laser welding, oxyacetylene powder welding, and lately-friction stir processes [3].
When nodular cast iron castings are being repaired or joined by means of traditional fusion welding technologies, their high carbon content (more than 3.5%) may cause the formation of hard brittle phases, namely the cementite carbides and martensite in the fusion zone and heat-affected zone (HAZ). Carbon atoms of nodular cast iron diffuse into austenite during welding and form martensite and carbides at the weld interface while cooling [4]. These give rise to poor elongation properties and high hardness values, as described in another paper [5]. The structural transformations result in the reduction of ductility to a level where the susceptibility to cracking is so high that either a spontaneous post-welding cracking of the joint takes place, or cracks are generated when the first operational load is applied [6]. Therefore, welding of nodular cast iron, likewise welding other cast irons, requires special precautions to obtain optimum properties in the welded metal and the adjacent HAZ. Brittle martensite can be tempered to a lower strength, yet more ductile structure through preheating or post weld heat treatments. Some welding procedures are designed to reduce the size of the HAZ and thus minimize cracking [7].
While the main goal is to avoid the formation of excess cementite in the metal matrix, which makes the bonded region brittle, in nodular cast iron an additional objective based on retaining a spheroid form of graphite, is of almost equal importance [8]. To minimize the formation of massive carbides and high-carbon martensite, the most beneficial solution is to utilize carbon in the form of spheroids which have a low surface-to-volume ratio [9]. Friction welding is an example of welding process in which both similar and dissimilar metals can be welded by means of solid-state diffusion processes to overcome metallurgical complications associated with fusion welding. The welding method produces a weld when two or more workpieces, rotating or moving reciprocally, come into contact under pressure to produce heat and plastically displace material from the weld interface [10].
The discussed friction welding process is beneficial in terms of reducing the cost of complex forgings or castings, for example, of welding of a spindle or shaft to a cast/forged head [11,12]. However, in the case of friction welding of steels, numerous defects occur, including: center defects, restraint cracks, weld interface carbides, hot-shortness cracks, and porosity, which may lead to the failure of the friction welds [13].
The main problem when it comes to friction welding of nodular cast iron is free graphite precipitations which have lubricating properties that reduce the efficiency of the welding process. During the friction welding process, graphite nodules are deformed or fragmented, thus creating an unfavorable microstructure [14][15][16]. Consequently, all the properties relating to strength and ductility decrease as the proportion of nonnodular graphite increases. The form of nonnodular graphite is important, because thin graphite flakes with sharp edges have an adverse effect on crack initiation [13].
Another problem is the absence of interatomic bonds in the case of friction welding of nodular cast iron caused by the formation of continuous graphite between contact surfaces. Said phenomenon is also accompanied by a large decrease of the friction coefficient, which in turn decreases the amount of heat generated at the interface. Higher pressure applied while rotating the components should support heating the joint to the melting point. However, an increase in pressure is accompanied by the formation of a network of cracks in the weld zone, mainly due to the small plasticity of nodular cast iron [17].
In recent years, a lot of studies relating to continuous friction welding [18][19][20][21][22] and friction stir welding processes [23,24] of nodular cast iron to nodular cast iron or ductile to other low carbon metals have been conducted by various researchers. Knowledge on friction weld processes is oriented toward structural and mechanical properties [25,26] and metallurgical phase transformation [27,28], or welding parameter optimization [29,30]. Microstructure and mechanical properties of friction nodular cast iron joints hitherto are poorly described in the literature. Only other materials, such as mild steels, have been considered in a friction welding of a truck axle-related application, where spindles are welded to axle housings [31].
The main purpose of this paper is to show the effect of temperature and welding parameters on microstructures and mechanical properties observed in friction-welded joints. As such, the tensile test, Vickers hardness test, direct temperature measurement, optical and the electron microscopy techniques were taken advantage of to conduct the discussed study.

Experimental
The chemical composition and mechanical properties of a parent material selected for the study are shown in Table 1. Figure 1 shows the microstructure of the as-cast nodular with graphite nodules in a matrix. The specimens having the shape of bars of 20 mm in diameter and 100 mm in length were cut for friction welding process. The surface for friction welding was prepared by means of using an abrasive cut-off machine. The process of joining was carried out on a continuous drive friction machine of the ZT4-13 type (ASPA, Wroclaw, Poland). Friction welding process of nodular cast iron is shown in Fig. 2. Friction welding parameters used in the experiment are shown in Table 2. During welding the primary parameters: friction pressure (FP), upsetting pressure (UP), friction time (FT), and upsetting time (UT) were continuously monitored and recorded. Few more trials have been carried out with different parameters in order to get defect free welds. The    main parameters employed in this study were FP, UP, and FT. Additionally, constant rotational speed (RS = 1450 rpm) and upset time (UT = 3 s) are used in this study. The geometry of specimens used for welding process and schematic representation on the arrangement of fixing thermocouples are shown in Fig. 3. These thermocouples were installed in 1.2 mm holes at the periphery and the axial center of a joint-2, 4, and 6 mm from the interface. The thermocouples were beaded at the tip and stuck at the measuring points with a silicate high temperature glue (TECHNICQLL, NALMAT, Trzebinia, Poland). Temperature identifiable in a close proximity to the interface of the joints was obtained with a TP202K1b2001 K-type thermocouple (NiCr-NiAl, CZAKI THERMO-PRODUCT, Raszyn, Poland) with the accuracy of ± 0.1°C. Temperature reading was performed with a UT 325 digital thermometer (Uni-Trend Technology, Hu Men Town, China) with the requisition frequency of 1000 Hz during friction welding. Room temperature tensile tests were carried out as per ASTM:E8/E8M-11 [32] standard specimen configuration. A tensile strength test was carried out on a 100-kN servo-controlled universal testing machine (Instron). Moreover, Vicker's microhardness measurements were made across the bondline with the load of 500 g and hold time of 15 s. The microstructure of joints was examined by means of light metallography (LM), as well as a JOEL JSM-500 scanning electron microscope (SEM). The specimens were mechanically polished by means of using emery special sheets with the help of disk and bench polishing machine. The prepared specimens were etched by applying 3% Nital for inspecting the metallurgical behavior of the welded joints. The fracture surfaces of the specimens were observed in the SEM using the BEI COMPO mode, applying backscattered electrons (BSE). Figure 3 shows the heat-affected zone formulated during the friction stage of welding process. In the early phase of the welding process (Fig. 3a), the surfaces of the analyzed  samples end were abraded during welding. As a result of this process, metal filings were thrown out of the joining plane. In the next phase of the welding process nodular cast iron (Fig. 3b), the graphite particles were removed from the outer surface by means of application of a high friction pressure. Oxides and other foreign materials present on the original surface were broken up and dispersed. Many sparks can be seen during the friction phase (see Fig. 3b). Moreover, during this stage, one can also observe axial displacement occurring at a slow rate, as the material at the interface starts to heat up, soften and is expelled as weld flash (Fig. 3b).

Heating phase
Time-temperature profiles at the axial center and the periphery are shown in Fig. 4a, b. Moreover, time-temperature curves were separated on the heating and cooling phases. As indicated by all of the thermocouple measurements, the analyzed metal experienced a slow rise in temperature to a peak value during welding. The peak heating temperature at x = 2 mm from the interface was 948°C (see Fig. 4a) for welding time t = 95 s. Temperature reached peak value, but did not exceed the melting point of nodular cast iron, i.e., 1130°C. Moreover, the increasing temperature at the axial center was about 100°C higher than the one at the periphery specimen (see Fig. 4b).
Difference in temperatures occurred due to the change of the yield strength stress of nodular cast iron with temperature (see Fig. 5). The temperature of high strength was lower than the one of low strength when the nodular cast iron base metal was deformed under the applied friction pressure. During the friction welding process, wear generated heat and consequently increased the temperature at the weld interface. Temperature increase resulted in the decrease of yield stress, which in turn allowed for an intensive plastic deformation of the welded metal at the interface. Additionally, reaction products, the microhardness distribution of the joints, and fracture morphology are all going to be discussed in the next sections.

Cooling phase
After the friction stage, the forging stage began and the metal cooled down gradually. During the cooling phase, heat was conducted away from the weld interface and lost to the surroundings through the ends of weld specimens. At 101 s, temperature started gradually dropping to 650°C after 120 s, while at the location of x = 6 mm from the interface, the peak heating temperature was 850°C. It can be seen from Fig. 4a that all three curves show a very similar cooling behavior below approximately 800°C. The temperature curves of the weld center and peripheral zone during cooling phase (Fig. 4a, b) clearly show that temperature of the centerline is higher than that of the peripheral zone at beginning of cooling phase. The cooling curve appears to be broader than heating curve, since rate of cooling is relatively slow. The similar curve profile received in the paper [33]. The heating and cooling rates of the welding process are approximately calculated around 14 and 5°C at the interface, respectively.

Mathematical modeling
The empirical models used to describe the temperature curves of welding process during heating and cooling phase are shown in Tables 3 and 4. The empirical model was formulated by the authors of this study for: a. Heating phase where a, b, and c are the constants of the models; t is the welding time, s; T is calculated temperature,°C. The goodness of fit of tested mathematical models to the experimental data was evaluated with the coefficient of determination R 2 . The higher R 2 the better is the goodness of fit the models. A nonlinear regression analysis was conducted to fit the models by genetic algorithm (GA) using computer program MATLAB 7.0 software (Math-Works Inc., Natick, MA). Genetic algorithms (GA) are a search optimization technique based on natural selection and heredity mechanism. The GA has been applied in many complex optimization and search problems, outperforming traditional optimization and search methods. The basic genetic algorithm contains the following steps: (1) selection of the initial population chromosome, (2) fitness function calculation, (3) checking stopping criterion, (4) selection of chromosomes, (5) application of the genetic operators, (6) creating a new population, (7) return to step 2, and (8) the best chromosome presentation.
The fitness functions formulated in Eqs. (1) and (2) are then used to measure the fitness value for each chromosome in the GA procedure. The critical parameters in GAs are the population size, number of generations, crossover probability, and mutation rate. In this study, a population size of 60, crossover probability of 0.8, mutation rate of 0.2, and the number of iteration of 500 were employed. Figure 6a shows the appearance of the as-welded joints with different welding parameters (see Table 2). There were noticeable differences in the width of the heat-affected zones and the amount of flash produced in the welds by opting for different welding conditions. As expected, the weld made using the long time welding condition had the most flash and widest HAZ, especially while compared to welds made using other welding conditions. Symmetrical weld flashes formed on both sides of the joints can be seen. Moreover, the size of the weld flashes was dependent on friction welding parameters. Figure 6 shows that for flashes S3, S6, and S7, specimens underwent a greater deformation resulting from a high axial pressure while compared to S1, S4, and S8 flash specimens welded with less axial pressure. Furthermore, it is evident that flash scale for 60 s (see S2 specimen) was much greater than the one for 30 s (see S1 specimen) which was consistent with the longer burn-off length (upset metal). The accumulated friction heat during the welding process was sufficient to soften the base metal. When the base metal reached the plastic state, upset metal formed due to the squeezing action with friction and upsetting pressures. Additionally, burn-off length increased rapidly with the increase of friction time. Similar phenomenon has already been observed by other authors [27,28].

Surface appearance
A nodular cast iron friction-welded macrograph of the specimen 6 is shown in Fig. 6b. Zones with different microstructures, such as weld interface, TMAZ and HAZ as the consequence of applying the friction welding process parameters to nodular cast iron. The microstructure variations of different zones of nodular cast iron joints are discussed in Sect. 3.7. Figure 7 shows the macroscopic fractography of tensile specimens for varied welding parameters. All the specimens failed at the bondline during the executed tensile test without obvious plastic deformation. Thus, fracture surfaces were relatively smooth and consisted of some continuous and shining regions near the periphery of the specimen. The phenomenon of incomplete bonding was detected at the center of the joint, as shown in specimens S4, S7, S10, and S11. These fractures were characterized by spiral lines of deformation distinctly, lighter plastic deformation, and insufficient heat generated in the central zone. Moreover, with increasing pressures, the fracture surface was more and more smooth as in specimens S3, S6, and S7 and relatively rough in the case of low pressure, as it can be observed in specimens S4 and S5 (Fig. 7).  During friction welding of nodular cast iron, metal is heated above the eutectoid transformation temperature A 1 (see Fig. 4a) and transforms into austenite. At the end of the forging phase, said metal cools down below the eutectoid temperature and austenite decomposes into ferrite, pearlite, bainite, or martensite, which are daughter products (see binary phase diagram Fig. 8) [33 -35].

Microstructure observation
Microstructure observations of the welded joints given in the form of a function of distance from the weld interface are given in Fig. 9. A thin proeutectoid ferrite layer S1 2,5mm formed close to the weld interface in both central and peripheral zones (see Figs. 9, 10b). The peak temperature during welding reached 950°C, thus in TMAZ, the process temperature was high enough to austenitize and coarsen the grains. Then the successive rapid cooling led to the facts that pearlite grains became coarse (Figs. 9, 10) and less ferrite precipitating at the pearlite grain boundary as reported in the papers [36,37]. And this situation was more serious in TMAZ near the weld interface where the process temperature was higher, which suggested the microstructure was quite heterogeneous in this area [38]. Moreover, ferrite structures in the original ductile iron had transformed into acicular martensite structures (Figs. 9, 11a) through rapid cooling from the high temperature state [34,35]. When the temperature of nodular cast iron exceeds the temperature of eutectoid, carbon in graphite spreads out to speed up the microstructure transformation by increasing carbon concentration in the base metal. While rapidly cooling, austenite may transform into a martensite structure (Fig. 11) as demonstrated in other papers [39][40][41][42][43].
The identified microstructures of nodular cast iron in the HAZ included irregular and deformed graphite precipitates and a mixture of pearlite and ferrite. The proportion of pearlite was small in comparison to the one of ferrite. Ferrite was mainly present around the region of graphite nodules (Figs. 9, 12a). The outcomes of microscopic observations show that graphite morphology (size and shape) was changed in an increasing distance from the weld interface. In some areas of the micrograph, ultra-refined graphite particles (0.5-1 lm) were uniformly distributed in a pearlitic matrix (see Fig. 9). The spheroidal graphite morphology changed into almost flake-shaped graphites (Fig. 12) located along the weld interface, because of the heat and extensive mechanical deformation generated during the friction process [23,24]. The length of flake-shaped graphite was approximately 50 lm (see Fig. 9). Using a high magnification, as given in Fig. 11b, numerous dense acicular martensite structures around the graphite nodules were distinctly presented. Martensite particles were approximately (2-6 lm) in size. Based on the phase transformation to martensite, the temperature of the HAZ during welding reached eutectoid temperature (see Fig. 8). Because the remaining time of eutectoid temperature was extremely short, only structures around graphite transformed into martensite while cooling as suggested by the authors of another paper [22]. As it can be seen from Fig. 12, the process of friction welding is difficult, because graphite particles, distributed in metal matrix without ductility, deteriorate the deformation plastic flow in high temperature and act as a lubricant, not allowing for a sufficient frictional heat to process the material. Because the plastic deformation of spheroidal graphite during friction welding was significant, deformation layers of spheroidal graphite were observed (Fig. 12a). Authors [14] reported that an early softening of nodular cast iron and a constant transfer of graphite nodules to the joint plane, led to the formation of a new graphite film (see Fig. 12b). However, the appropriate welding conditions reduced the amount of deformation layers of spheroidal graphite in the joint of nodular cast iron, as demonstrated in another paper [18].
The increase in the amount of deformation in thickness (about 40%) led to the elongation of graphite through the metal matrix. It is clear that, at this high amount of reduction, the matrix was deformed around graphite, especially at both sharp ends of angular graphite. Consequently, in the vicinity of graphite clusters areas of striations were created. These striations passed through many grains without deviation, resulting in a severe localized micro-cracking. The micro-cracks were often observed along matrix grain boundaries. Similar trend was reported in yet another paper [44]. Figure 13a shows the observed cracks. It can be seen that there are multiple cracks and deformed graphite nodules in the path of the cracks. The cracks are visible around the graphite and some cracks are actually in the graphite (circle in Fig. 13a). Extensive investigation of the situation when there was a lower crack severity showed that the graphite indeed deformed nodule particles were initiation sites for the cracks (arrows in Fig. 13a). However, when thermal stress was further developed in the adjacent areas, another crack initiated from the graphite particle and propagated in the metal matrix [44]. The connection of these minor cracks led to the formation of large major cracks, which run through martensite structures. Similar crack mechanism has been observed in case of powder welding process in other papers [45,46]. Figure 13b shows a schematic representation of cracking in the HAZ of a nodular cast iron in the case of friction welding process. As it can be seen, the marked areas are critical, because probably the highest stress can be identified there. The ellipsoidal or even flake-shaped graphite in the HAZ acts as a stress concentration, which may prematurely cause a localized plastic flow at low stress and initiate fracture in the matrix at higher stress (Fig. 13a). Moreover, graphite flakes deflect a passing crack and initiate countless new cracks as the material breaks. As a result, this graphite morphology exhibits no elastic behavior and fails in tension without a significant plastic deformation [13, 47]. Figure 14 shows the Vicker's hardness profile of the nodular cast iron cross-section following welding process.

Microhardness
The hardness of the TMAZ, HAZ, and parent material region was measured. The hardness value of the nodular cast iron parent material was approximately 185-195 HV. According to the microhardness curve of the peripheral zone, the maximal hardness value reached 605 HV in the TMAZ because numerous martensite structures were observed in this region (see Fig. 11a). Martensite was generated because the material was intensively heated and rapidly cooled down during the welding process [23,24]. The high hardness of the peripheral zone suggests that the carbon content of the martensite structures was relatively high. The result shows that the hardness of the welded sample decreased more rapidly when the phase changed from a fully deformed zone to the HAZ. As can be observed in Fig. 14, hardness decreases much slower in the HAZ and extends to 7 mm from the weld interface. The hardness value of this region ranged between 210 and 310 HV, which remained higher than that of the parent material. The hardness value of the central zone was slightly lower than the one of the peripheral zone, and the hardness of the parent material was the lowest. The maximum hardness value of the central zone, indicated in the TMAZ, was equal to 516 HV. As expected, hardness close to the weld zone was much higher than the hardness of the HAZ and the parent material (see Table 1). Variations of hardness in the welding interface were directly associated with the changes in the microstructure subjected to a high temperature and plastic deformation [3]. Earlier studies have indicated that when a carbon alloy is used in the welding process, refined grains are generated at the weld because of the dynamic recrystallization caused by a severe plastic deformation at high temperatures. After cooling and applying the subsequent pressure, the process of recrystallization and growth takes place that results in fine grain structure as demonstrated authors in the papers [33,48]. Moreover, according to the Hall-Petch relation, this would result in higher strength of the welded joint as reported [49].The structure of these refined grains comprised a lot of dislocation phenomena as reported by authors [47,50].

Effect of welding parameters on tensile strength of welded joints
A tensile test was applied after machining weld flashes formed during the friction welding process. The effect of friction time on the ultimate tensile strength (UTS) of welded joints and flash diameter can be seen in Fig. 15a, b. Friction time was changed from 30 to 240 s. In Fig. 15a it can be seen that the tensile strength of the joints increased together with increasing friction time for nodular cast iron samples. Increase in friction time leads to more amount of heat generation at the weld interfaces as reported in the papers [51,52]. However, the results of tensile strength obtained for the welds were not satisfactory. The maximum tensile strength was increased up to the value of 234 MPa, which was 53% efficiency of nodular cast iron strength base material. In addition, it can be seen that the UTS increased quickly with increasing upsetting pressure.
The effect of friction time on flash diameter is presented in Fig. 15b. It can be seen that flash diameter increased together with increasing friction time. Similar results for the relationship between welding parameters, flash diameter, and tensile strength were achieved for different joints in other papers [25,26].
The effect of friction time on axial shortening is presented in Fig. 15c. It is clearly seen that axial shortening increased together with increasing friction time. A similar trend was observed by authors [49,53]. The combination of high friction welding parameters produced more axial shortening that played an important role in the mechanical properties of the joints, as reported by authors of other works [27,28].

Fractography
The fracture surfaces of the tensile tested specimens were identified by means of scanning electron microscopy so as to understand failure patterns. The failure of the joints occurred only in the form of a fracture along the interface. Figure 16 shows a fracture surface micrograph of the tensile specimen fabricated using optimum parameters. It can be seen that two dissimilar fracture morphologies were distinguished in nodular cast iron welds. The fracture observation of the peripheral zone sample showed a cleavage fracture with river markings on the facets (Fig. 16a, b). River markings on the facets resulted from the propagation of the crack on a number of planes of different levels [46]. The cleavage planes of {100} or {110} type were clearly observed in the immediate vicinity of the graphite nodule. Additionally, martensite also cleaved along the {100} planes. Due to significant differences in orientations of single martensite blocks, crossing grain boundaries by a crack was hindered. In this case, the river pattern could be barely noticed (Fig. 16b). The cleavage planes are characteristic for ferrite, which forms a specific shell around graphite nodule [27]. In the case of brittle fracture, little to no visible plastic deformation precedes the fracture. Generally, a brittle fracture propagates through the grains. However, in the case of highstrength material, the crack follows grain boundaries [47].
The fracture observed in the central zone sample showed the ductile mode of the fracture surface (Fig. 16c). Cavities arose from inclusions or coarser precipitates were enlarged and-during further yielding-the material between them was necked and sheared. Moreover, a microvoid coalescence (MVC) seems to be the dominant form of the fracture region (Fig. 16c). The dimple pattern around graphite nodules showed the deformation of the surrounding ferrite during the final period of straining up to fracture. Additionally, Fig. 16c, d demonstrates a relatively large cavity size, especially in comparison to the graphite nodule size. According to the authors of yet another publication, [29] the formation of these cavities may be attributed to decohesion of graphite and the surrounding matrix, as demonstrated in Fig. 15d.

Conclusions
The following conclusions can be drawn from this paper:  Fig. 16 Fracture surface micrographs showing a mixture of dimple and a cleavage pattern mode for tensile test specimen with the optimum parameters a, b cleavage fracture with river markings; c, d shallow dimples with microvoid coalescence visible in the central part of the micrograph significantly affected microstructure transformation. Final microstructures in welding zones of friction weld were composed of pearlite, proeutectoid ferrite, and acicular martensite around the graphite nodules. 4. Graphite in the surface zone exhibited a striped configuration and distinct martensite structures formed in the metal matrix. Graphite in the central zone remained in the form of individual granules and acicular martensite was observed outside the graphite nodules. 5. Ellipsoidal or even flake-shaped graphite in the HAZ acts as a stress raiser, which may cause a localized plastic flow prematurely and initiate fracture in the metal matrix. Cracks occur mostly at the interface of deformed graphite nodules and then spread through grain boundaries of the metal matrix.