Journal Information
Vol. 8. Issue 3.
Pages 2481-3388 (May - June 2019)
Download PDF
More article options
Vol. 8. Issue 3.
Pages 2481-3388 (May - June 2019)
Original Article
DOI: 10.1016/j.jmrt.2019.01.028
Open Access
Thermal stability of Al–Fe–Ni alloy at high temperatures
Zeyu Biana,b, Shihan Daia,b, Liang Wua,b, Zhe Chena, Mingliang Wanga,
Corresponding author

Corresponding authors.
, Dong Chenb,c,
Corresponding author

Corresponding authors.
, Haowei Wanga,b
a School of Materials Science & Engineering, Shanghai Jiao Tong University, Shanghai, China
b State Key Laboratory of Metal Matrix Composites, Shanghai Jiao Tong University, No. 800 Dongchuan Road, Shanghai 200240, China
c Anhui Aluminium Matrix Composites Engineering Research Centre, Huaibei, China
This item has received

Under a Creative Commons license
Article information
Full Text
Download PDF
Figures (11)
Show moreShow less
Tables (4)
Table 1. The composition of Al–Fe–Ni alloy (wt%).
Table 2. The K values at the corresponding temperatures.
Table 3. The solid solubility of Fe and Ni in the Al matrix (wt%).
Table 4. The g values at the corresponding temperatures.
Show moreShow less

The Al–Fe–Ni eutectic alloy was produced by gravity cast, and the thermal stability of this eutectic alloy was studied for the first time in this work. The phase composition was characterized by X-ray diffraction, and the morphology evolution of eutectic phase was analyzed by differential scanning calorimeter and scanning electron microscopy. According to the microstructure evolution of the eutectic phase exposed to different temperatures, three stages can be divided to distinguish coarsening behaviors of the eutectic phase, including the stable stage, transient coarsening stage and fully coarsening stage. In each stage, the coarsening behavior of eutectic phase was studied. Firstly, it was found that the stable stage can be maintained up to at least 673K. Subsequently, the eutectic phase should go through coarsening and spheroidization in the transient coarsening stage. Among this stage, both the shape change and diffusion factors have played the vital roles in affecting the coarsening rate and spheroidization process. Finally, in the fully coarsening stage, the volume fraction of eutectic phase was deemed to be the main reason accounted for the difference between the results given in literature and this work.

Al–Fe–Ni eutectic alloy
Thermal stability
Ostwald ripening
Full Text

The eutectic Al alloys possess excellent liquidity and good hot tear sensitivity, and are usually the best choices for casting Al alloys [1,2]. The Al–Si eutectic alloys have become the most widely used cast Al alloys in the past decades, owing to their good castability, high corrosion resistance, and good wear resistance [3,4]. Through the addition of Cu and Mg elements, the Al–Si matrix can be highly strengthened via the precipitation of Al2Cu and Mg2Si phases [5]. With the further demands of improved elevated temperature performance, Al–Si alloys are difficult to afford because of the high diffusivities of Cu and Mg in the Al, and the lower eutectic temperature of Si phase [6–8]. Moreover, the Fe, Ni and RE elements can diffuse slowly in the Al, and are able to form high-temperature stable phases. Therefore, these elements have been used to develop novel eutectic Al alloys [9,10]. For instance, the Al–Fe–Ni, Al–Ni and Al–Ce eutectic alloys [11–13] and Al-based entropic alloys [14] have been investigated in recent years.

The Al–Fe–Ni system has been subjected to numerous investigations due to its important technological applications, including high-temperature alloys, magnetic materials and shape memory devices [15–17]. Since the low solubility of Ni and Fe in Al, the Al–Fe–Ni alloys are considered as promising materials for various electronic and electric applications to substitute conventional low conductivity Al casting alloys [18,19]. On account of the high resistance to corrosion at high temperatures of Al–Fe–Ni alloys, Leenaers et al. [20,21] applied it as the nuclear fuel cladding material and obtained good results. Canté et al. [22] studied the effect of cooling rate on the microstructure and hardness of Al–1.0Fe–1.0Ni alloy by directional solidification. They found that the microstructure was refined with the increasing cooling rate, and provided the relationship between hardness and spacing of eutectic phase. In order to control the bulky eutectic phase formed in solidification, the powder metallurgy and rapid solidification were used [23,24]. Průša et al. [25] examined the compression performance of Al–Fe–Ni alloy prepared by centrifugal atomization and hot extrusion at higher temperatures. They found that the mechanical performance of Al–Fe–Ni alloy was better than Al–Si alloy.

The Al–Fe–Ni alloy acting as the candidate for heat-resistant Al alloy has been widely accepted [26,27]. However, recent researches are still focusing on the adjustment of preparation technologies for the alloy. The study of microstructure and secondary phase evolution at elevated temperature is still scanty. The secondary phases have played the major role in improving the high temperature performance of the alloy. Thus, the acknowledge of their behaviors in high temperature environment is of great significance [28]. The behaviors of secondary phases at elevated temperatures are usually described by Ostwald ripening [29]. During such progress, the smaller phases should be swallowed by the larger phases derived by the decrease of total surface free energy. Lifshitz, Slyozov, and Wagner (LSW) provided the primary theory of coarsening kinetics [30,31]. However, this theory has merely described an ideal steady state coarsening, thus required modification in the practical use. The existed modification mainly concentrated on the volume fraction, distribution of secondary phase or transient progress before the steady state [32–34]. Speich and Oriani [35] have extended the LSW theory to the case of nonspherical particles, and they introduced a shape factor to describe the effect of morphology. Nevertheless, they did not consider the change of diffusion caused by morphologic changes of the secondary phases. Premkumar et al. [36] investigated the coarsening behavior of the secondary phase produced by powder metallurgy in the elevated temperature. They suggested that the coarsening rate coefficient was dependent on the volume fraction of the secondary phase. However, the system they studied was close to the ideal state, in which the shape of phase was near-spheroidal. While in the common cast circumstance, the shape of Al9FeNi eutectic phase is usually needle-like [37]. Thus, there is the obvious difference from their research.

The aim of this research is to investigate the basic feature, microstructure and property of Al–1.75Fe–1.25Ni eutectic alloy prepared by the traditional gravity cast method. Specifically, the thermal stability of the Al9FeNi eutectic phase for a long-time holding at elevated temperatures was studied, and the relationship between coarsening rate coefficient and temperature was put forward. Moreover, the effect of shape change of Al9FeNi phase on the alloy performance was also discussed.

2Experiments2.1Materials preparation

The Al–1.75wt% Fe–1.25wt% Ni alloy was prepared by commercially pure grade Al ingots (99.99% purity), Al–20wt% Fe and Al–10wt% Ni master alloys. The melting process was performed by an electric resistance furnace in a clean iron crucible, which was coated with the boron nitride paint. The alloy was melted at 1033K initially. Then, the melt temperature was heating up to 1123K, and held for 30min to ensure the fully dissolution of Fe and Ni elements. Afterwards, the melt temperature was reduced to 1033K, and the alloy melt was purified by the refining agent. In the next step, the melt was degassed in the vacuum furnace for 10min. Finally, the melt was poured at ∼1013K by gravity cast into a cast iron mold that was preheated to 473K. The schematic of the iron mold which is the key equipment used in the gravity casting is shown in Fig. 1.

Fig. 1.

The schematic of the iron mold used in the gravity casting.


The elemental composition of the as-cast alloy was measured by the inductively coupled plasma-atomic emission spectrometry (ICP-AES, iCAP7600). In Table 1, it is seen that the nominal composition of the Al–Fe–Ni alloy is in good accordance with the ICP result, indicating the successful casting control for the alloy.

Table 1.

The composition of Al–Fe–Ni alloy (wt%).

Element  Fe  Ni  Si  Al 
Nominal composition  1.75  1.25  –  bal. 
ICP result  1.67  1.26  0.028  bal. 

ICP = Inductively Coupled Plasma.

2.2Microstructure characterization

The phase composition was detected by X-ray diffraction (XRD, D8 ADVANCE Da Vinci) from 20° to 90° at a speed of 2°/min. Furthermore, the as-cast samples were metallographic grinded by the 400 #, 800 #, 1200 #, 2500 # emery papers (Starcke, 991A), and polished by the diamond suspension. Some polished samples were deeply etched in 5wt% NaOH for 1h to emphasize the microstructure features. The scanning electron microscopy (SEM, TESCAN MAIA3) was applied to characterize the microstructure features of the alloy. The SEM equipped with energy dispersive spectroscopy (EDS) was used for the compositional analysis for Al–Fe–Ni alloy.

The thermal analysis was carried out using a differential scanning calorimeter (DSC, Q20 type, TA Instrument) from room temperature (RT) to 1073K under a protective Ar atmosphere. The hardness was measured using a Vicker's hardness (HV) testing machine (EZ-mat, CARAT 930) with a load of 10kg and dwell time for 15s. At least five random dispersive points, on the circular section of a cylinder with a diameter of 10mm, were selected to reflect the overall hardness level of the sample. The average size and volume fraction of the second phase were obtained from SEM micrographs by the ImageJ software. At least five photos taken at different areas were used to estimate the average values of the sample at different situations. The gray scale contrast was used to distinct the eutectic phase from matrix, about 2000–5000 data points were utilized in one sample.

3Results and discussion3.1Typical microstructure of Al–Fe–Ni alloy

Fig. 2a shows the typical microstructure of Al–Fe–Ni alloy. The areas with symbols a, b and c are α-Al, the needle-like eutectic structure and grain boundary, accordingly. In Fig. 2h, the composition of Al–1.75wt% Fe–1.25wt% Ni alloy has located right in the Al–Al9FeNi two-phase region [18]. In this composition, the eutectic phase should be both Al and Al9FeNi. The eutectic temperature is about 923K as reported in literature [1], though it is not provided in Fig. 2h. Furthermore, the Al9FeNi phase has further been confirmed by the XRD analysis (Spectrum I, Fig. 2g). In the XRD pattern of as-cast sample, the Al phase and Al9FeNi phase can be well indexed to JCPDS No. 04-0787 and the reference [36], respectively. According to the calculated XRD pattern of Al9FeNi (Spectrum II, Fig. 2g) using the CrystalDiffract software, it is found that some unmarked diffraction peaks in literature should be ascribed to Al9FeNi as well. Besides, the studied Al–Fe–Ni alloy has the same composition with the alloy in literature [1], in which the intermetallic phase was deemed as Al9FeNi. Due to the typically gravity casting method in both cases, the intermetallic phase should confirm to be Al9FeNi, which has complied well with the predicted phase in the phase diagram [18].

Fig. 2.

(a) Typical microstructure of Al–Fe–Ni alloy; (b) local area of grain boundary; (c) local area of grain boundary after deep-etching; (d) a near boundary area of a grain; (e) Ni distribution map of (d), and (f) Fe distribution map of (d); (g) Spectrum I: XRD pattern of as-cast sample; Spectrum II: the calculated XRD pattern of Al9FeNi; (h) Isothermal section of the Al–Fe–Ni system at 627°C with Al content above 50at.% [18].


In this as-cast alloy, the average size of Al9FeNi needle-like eutectic phase is measured to be ∼320nm in diameter on the cross-section, with the length ranging from ∼5 to ∼50μm. Furthermore, the length distribution of the Al9FeNi phase is asymmetrical, and most of the lengths are concentrated at ∼20μm at the interior of eutectic grains. The detailed structure of the Al/Al9FeNi eutectics is analyzed in Fig. 2b, as the labeled rectangular area in Fig. 2a. The needle-like eutectic phases are both seen in the eutectic grains (Area d) and grain boundaries (Area e) (Fig. 2b). This phenomenon can be observed more clearly on the deep-etched surface, as displayed in Fig. 2c. Through the EDS analysis, it is verified that these phases are all Al9FeNi (Fig. 2d), which is in agreement with those mentioned in the literature [26].

The distribution of Fe and Ni element can be observed in the EDS maps of Fig. 2e and f, respectively. Specifically, it is found that Ni has a higher content on the grain boundary, while Fe has a higher content inside the grains through the EDS analysis (Fig. 2d). This may attribute to the non-equilibrium solidification during the actual casting process. It can be explained that the Fe element has been engulfed more inside the eutectic grains during the solidification of the Al–Fe–Ni alloy, since the Fe element has a lower diffusivity in Al matrix. Meanwhile, the abundant Ni element has been pushed into the solidifying front of the liquid phase, and finally clustered on the grain boundary.

3.2Basic performance of Al–Fe–Ni alloy

Fig. 3a shows the HV values of the as-cast alloy in relation with the holding temperatures (RT to 773K) for 1h. It can be seen that the alloy shows the almost stable HV values from 473 to 773K with the interval of 100K, which is slightly lower than that of the as-cast alloy. Fig. 3b exhibits the HV values of the as-cast alloy on dependence of the exposure duration from 0 to 24h at 598K with a very slight reduction in comparison with those of the as-cast alloy without treatment. Therefore, it should be sure that this alloy must have higher thermal stability than the typical cast Al–Si or Al–Cu alloys [2,6,7]. For instance, the high temperature tensile strength of Al–25Si based alloy has dropped from ∼155MPa to ∼10MPa, when the temperature increased from 653K to 808K [38].

Fig. 3.

(a) The HV values of alloy kept 1h in different temperatures; (b) the HV values of alloy kept in 598K after different time.


Nevertheless, there is still a gentle strength reduction of this alloy experiencing the thermal exposure, and it may be ascribed to the coarsening of eutectic structures. Based on these points, the microstructure evolution of Al/Al9FeNi eutectics under higher temperatures has attracted our further interests.

3.3Thermal stability of Al9FeNi phase3.3.1The microstructure evolution of Al9FeNi at different temperatures

As the main second phase in Al–Fe–Ni alloy, the Al9FeNi eutectic phase has played a critical role in determining the alloy performance. In order to further investigate the thermal stability of this eutectic alloy, the variation of microstructure of Al9FeNi phase has been studied in combination with the thermal exposure of the alloy at higher temperatures and longer durations.

Fig. 4 shows the microstructures of Al–Fe–Ni alloy under 673K, 773K and 898K for 24h, 48h and 72h, accordingly. After being exposed at 673K, the Al9FeNi phase exhibits the uniform distribution inside the eutectic grains during the whole duration. Meanwhile, the average equivalent diameter on the cross section (dcs) of the Al9FeNi phase can be analyzed to show the two-stage coarsening behavior. (1) At the beginning 24h thermal exposure, it is observed that the dcs value has remained ∼320nm of the Al9FeNi. (2) From 24 to 72h thermal exposure, the dcs value increased from ∼320 to ∼360nm, indicating a very small coarsening behavior. Furthermore, it is worth noting that there is little change in the length direction of the Al9FeNi phase, as shown in Fig. 5a. Generally, the Al9FeNi phase has kept stable at 673K exposure.

Fig. 4.

The sections of alloys after thermal exposure durations (0–72h) in the temperatures (573–898K).

Fig. 5.

The length direction of Al9FeNi phase holding at different temperatures after 72h: (a) 673K, (b) 773K, and (c) 823K.


During the thermal exposure at 773K, some obvious changes can be identified through the microstructure observation. Initially, some particles free areas (PFAs) are found in the eutectic grain interior (labeled in the rectangular area in Fig. 4), this phenomenon can be called as the local ablation phenomenon. In this progress, some eutectic phases are ablated into segmented phases even faded away to support their coarsening, reflecting the beginning of Ostwald ripening. However, the majority of eutectic phases have remained unchanged in length direction, apart from some PFAs formed by the phases ablated wholly or partly to a minor extent (Fig. 5b). These have indicated that only the dcs values of the Al9FeNi phase changed. Overall, the dcs values show the roughly increasing tendency with the exposure duration at 773K in a linear manner.

When the temperature is rising to 898K, approaching to eutectic temperature of this alloy, the shape of the Al9FeNi phase has generally changed from needle-like to spherical during the first 24h. This shape change progress, specific for the long needle-like phase ruptured into several segments, can be observed in a lower temperature (823K), which is exhibited in Fig. 5c. Together with this process, the ablation of eutectic phase should turn from local into global, and PFAs became even larger. The dcs values have exhibited a faster coarsening rate at the 0–24h exposure, and a slower coarsening rate at the 24–72h exposure. With the prolonging holding period, the coarsening of the Al9FeNi phase is fairly obvious compared to lower temperatures, as exhibited in the statistic diagram.

The Ostwald ripening, also known as coarsening, is usually used to describe the coarsening behavior of precipitated phase and eutectic phase at higher temperatures in the previous works [32,33]. In this procedure, the larger particles should become greater in size, while the smaller particles are dissolved by the reduction of total interface energy. Lifshitz, Slyozov and Wagner have put forwards the relation of the average particle size versus the holding time in a specific temperature [30,31], which can be described as:


  • d¯(t) is the average size of particle at certain time;

  • d¯(0) is the initial average size of particles;

  • K is a constant independent of time.

Using this formula, the relationships between dcs3 values for the Al9FeNi phase and thermal exposure duration in the temperatures (573–898K) have been linearly rationalized, as shown in Fig. 6.

Fig. 6.

Simulated results for the relationship between dcs3 and thermal duration.


The K values of corresponding to the temperatures have been listed in Table 2. It is clear that the coefficient K values are obviously increasing with the rising temperatures. In detail, the K values can be divided into three groups, including: (1) K(573K) and K(673K), (2) K(773K) and K(823K), and (3) K(873K) and K(898K). In each group, the K values are at the similar level with each other, indicating their similar coarsening behaviors for the Al9FeNi phase.

Table 2.

The K values at the corresponding temperatures.

Temperature (K)  K (μm3/h) 
573  1.4071×10−4 
673  2.783×10−4 
773  5.266×10−4 
823  9.439×10−4 
873  0.00921 
898  0.014 

Fig. 7 shows the DSC curve of as-cast Al–Fe–Ni alloy heated over 973K under the heating rate of 20K/min. Two inflection points are exhibited during the whole heating process, i.e., Ta=770K and Tb=840K. According to the variation tendency of the curve, both inflection points should correspond to the endothermic peaks, indicating that the local ablation, spheroidization and coarsening behaviors may occur in the Al9FeNi phase. Nevertheless, it is worth pointing out that the DSC has manifested a dynamic process of temperature change, and it is used to illustrate the existence of various behaviors of eutectic phase during the energy accumulation process here. In comparison, the alloy for the thermal exposure has experienced a quasi-static process with the constant temperature. Thus, the microstructure evolution for the eutectic phase during the thermal exposure cannot be fully described by the DSC experiment. Practically, the DSC experiment can emphasize the microstructure evolution from the aspects of energy fluctuation, similar to the thermodynamic prediction.

Fig. 7.

Thermograms of samples heated under the heating rates of 20K/min.


According to LSW theory, the corresponding relation between K and temperature is given in the following formula:


  • γ is the interfacial energy between eutectic phase and matrix;

  • Vm is molar volume of the eutectic phase;

  • Ce is equilibrium solubility of rate-controlling solute in the matrix.

D is diffusion coefficient of rate-controlling solute in the matrix, which can be expressed as:

Take logarithm of both sides, Eq. (2) can be simplified into:
where a and b are constants independent of temperature, respectively.

The lnK values at certain temperatures are scattering plotted in the diagram as showed in Fig. 8a. The solid curve drew in Fig. 8a is used to match the relationship between lnK and T, which is referring to the theoretical values provided by Premkumar et al. [36]. There is an obvious difference between the relationship calculated from literature and measured values in this experiment. This should be related to different behaviors of needle-like Al9FeNi in different temperature ranges as mentioned above.

Fig. 8.

(a) The diagram of lnK and temperature; (b) variation of lnK with temperature and three stages among it.


According to the microstructure observation and the LSW equation analysis, the evolution of the Al9FeNi phase at different temperatures can be roughly divided into three stages (Fig. 8b). In order to separate these stages, we define the first turning point as T1 between 673K and 773K, the second turning point as T2 between 773K and 873K. The Al9FeNi phase should display different behaviors at different temperature ranges:

  • Stable stage: When the temperature is lower than T1, the Al9FeNi phase has kept stable with a very small coarsening rate with the extended holding time.

  • Transient coarsening stage: When the temperature is between T1 and T2, the local ablation for Al9FeNi phase can partly occur, and the PFAs have emerged obviously. Simultaneously, several eutectic phases may rupture to some segments. Overall, the Al9FeNi phase has shown a medium coarsening rate.

  • Fully coarsening stage: When the temperature is higher than T2, the rupture for Al9FeNi phase should wholly occur. All the eutectic phases are segmented into small phases, and then spheroidized and coarsened quickly.

Critical temperatures of the two adjacent stages are T1 and T2, respectively. Each stage is discussed in detail below.

3.3.2The behavior of Al9FeNi in each stage3.3.2.1The stable stage

In the stable stage, the Al9FeNi phase has kept stable as well as the performance of matrix. In this stage, the enlargement of Al9FeNi size induced by Ostwald ripening is not dominant. The main reason of the enlargement should be owing to the precipitation of Fe and Ni from supersaturated solid solution formed during solidification. The actual solid solubility of Fe and Ni in the cast Al matrix can be found using EDS in Table 3.

Table 3.

The solid solubility of Fe and Ni in the Al matrix (wt%).

Element  Fe  Ni 
Primary Al  0.19  0.06 
Eutectic Al  0.39  0.31 

The basic difference between the stable stage and transient coarsening stage is that at the transient coarsening stage, the local ablation can occur, leading to the growth of the larger phase by consuming the smaller phase. The critical temperature divided these two stages has been estimated by a simplified model exhibited in Fig. 9.

Fig. 9.

The simplified model of local ablation.


It is considered that the distribution of eutectic phase is closely packed observed from the cross section of the phase. Therefore, the ablation of a phase can affect the adjacent six phases, and a coarsening phase should be influenced by three ablative phases at least. In order to simplify calculation, the length of eutectic phase is supposed to be approximately equal. Therefore, the enlargement of phases corresponding to local ablation can be described as:

where d0 is 320nm as mentioned above.

In this experiment, the longest duration for discernible local ablation of eutectic phase is 72h in this work. Therefore, taking d¯ and d0 into formula (1), the value of lnK can be calculated as −7.873. Besides, the relationship between lnK and T at lower temperatures (i.e., 573K, 673K and 773K) can be described by a certain tendency which is matched by dashed Line 2 showed in Fig. 8a. Finally, the corresponding T1 is found to be ∼716K, which is marked by triangle symbol in Fig. 8a. This value is estimated by tendency described by dashed Line 2. Indeed, this calculated T1 has fallen into the range of 673K and 773K. Meanwhile, if the precipitation of Fe and Ni from the Al matrix is also included for consideration, the practical critical temperature should be a little higher. fully coarsening stage

This stage has to be described prior to the transient coarsening stage, since it is a requisite to give the theoretical relationship between lnK and T beforehand. In this stage, the shape of eutectic phase has been segmented, and then transformed completely to the round shape (Figs. 4 and 5c). When the shape of eutectic phase has changed into spherical, the variation-tendency of lnK with the temperature should be in accordance with the theoretical prediction [36].

However, the tendency between lnK and T are observed to offset down along y axis in this stage of our work, which can be exhibited as dashed Line 1. This discrepancy is suggested to be caused by the influence of the volume fraction of eutectic phase. In this experiment, the volume fraction of eutectic phase (∼10vol%) is less than 30vol% mentioned in literature [36]. Kim et al. [38] found the volume fraction of eutectic phase should have the impact on the K value in the following equation:


  • Φ is the volume fraction of eutectic phase;

  • f(Φ) can be approximate to 0.5 when the volume fraction is from 10vol% to 30vol% according to this literature [39].

Take logarithm of both sides, Eq. (6) can be simplified into:

Thus, the tendency between lnK and T should be observed to offset down along y axis compared to literature [35]. This tendency is in good accordance with the experimental values at the fully coarsening stage in our work. Conclusively, the experimental relationship in the fully coarsening stage can be described as: transient coarsening stage

In this stage, with the increasing temperature, the lnK values have experienced the processes of the slight rise and accelerated rise. To be specific, the transition rate of lnK versus T in each stage can be approximate to dlnK/dT, and this value should experience an increment. In this process, the initial rising rate of stable stage, which can be described as dlnKss/dT, is slower than the theory value (dlnKfcs/dT) of fully coarsening stage, as can be realized from dashed Lines 1 and 2 drawn in Fig. 8a. However, the accelerated rise of lnK value has made this variation tendency approaching to the theoretical tendency, which exhibited in Fig. 8b. According to these features in different temperature ranges, we may affirm that the shape change must be the essential reason to cause this alteration.

This shape change process should be ascribed to the spheroidization [40]. In this process, the long needle-like phase has ruptured into several spherical or near spherical particles by the decrease of strain energy. During the spheroidization, the distribution of eutectic phase should be changed, and the diffusion of element should also be affected. The constants c and d can be used to describe these changes, and the relationship between lnK and T can be expressed as:

Through matching the variation tendency before the occurrence of the spheroidization, the empirical values of c and d in this experiment can be obtained as −12.21 and 0.257, respectively.

When the shape has changed from the needle-like to spherical, the diffusion path, mainly the interface between phase and matrix, should be cut short. This has caused the diffusion to become harder after shape change. Thus, the constant d is smaller than 1, indicating the easier diffusion when the eutectic phase has kept needle-like morphology at low temperatures. Moreover, the shape change can also affect interfacial energy, equilibrium solubility at interface and distribution of phase, making the constant c have a negative value. Eventually, these effects have resulted in the apparent distinction after morphological changes, and these results are in good accordance with reference [41].

After further investigating the coarsening process as the increase of temperature, it is conscious that the impact of time is non-negligible. In other words, the spheroidization is a time-dependent behavior. The spheroidization can be initiated if the holding time is long enough in the slightly lower temperature exposure, and it has required the sufficient time to accomplish at the higher temperature exposure. That is to say, if the exposure duration can either be long enough or short enough, there are two boundaries reflected in the variation of lnK values, as shown in Fig. 10.

Fig. 10.

The transform of lnK when investigate time change.


The curve I is the theoretical prediction of K, if the Al9FeNi phase shows a near-spherical morphology [35]. The curve II is the actually preliminary variation of K, if the Al9FeNi phase shows a needle-like morphology. The spheroidization can make the variation of K change from curve II to curve I. When there is the sufficient time for coarsening process, the K change is able to begin in a very low temperature, and finish at a medium temperature.

The temperature range that allows for the inadequate spheroidization in whole duration should be very tiny. Thus, the change of K may be just a mutation (i.e., the transient coarsening stage), which has accorded with the curve III. If the coarsening has occurred in a short time interval, the spheroidization is not likely to be accomplished in this procedure. Therefore, the switch of K should become more difficult, and the curve I cannot be reached, which can be described as curve IV instead.

From another point of view, there are two temperature points: (1) the start of spheroidization during the whole thermal exposure process; (2) the finish of spheroidization before the first interval (24h in this experiment).

The first point corresponding to the start of spheroidization is usually vague and hard to estimate, but the second point is the T2 as mentioned above. T2(t) means total time spent for spheroidization, which can be described as [42]:


  • k is Boltzmann's constant (1.38×10−23J/K);

  • Ds is diffusion coefficient of rate-controlling solute (Fe) in the interface;

  • γ is the interfacial energy between eutectic phase and matrix (0.2J/m2[36]);

  • ρ is the radius of the original cylinder (160nm);

  • Φ denotes the atomic diameter of rate-controlling solute (Fe, 0.254nm).

From this formula, it can be seen that the diffusion coefficient (Ds) and the radius of the original cylinder (ρ) can exert highly influence on the rupture time of spheroidization. The calculation of specific diffusion at the interface of eutectic phase can be referred to literature [42]:


  • D0 is 91m2/s;

  • EA,V is 74.3kJ/mol [42].

  • EA,inter is 258.1kJ/mol referred to [43,44].

When the temperature is 873K, the time of spheroidization can be calculated as approximately 0.7h, which means the eutectic phase should fully accomplish the shape change during the first interval. When temperature is 773K, the time of spheroidization can be calculated as approximately 18.7h. While in this experiment, even after sample exposed 72h at 773K, the eutectic phase should mainly keep the needle-like morphology, as shown from length direction in Fig. 5b. However, some other factors can affect the spheroidization as the temperature decreases. Most importantly, the participation of vacancy in diffusion should be reduced at lower temperatures on account of the reduced vacancy concentration [45]. Thus, the diffusion coefficient Ds should be diminished, and the required spheroidization time should be further lengthened as well. In consideration of the absence of detailed theoretical models, two hypotheses are proposed here:

  • (1)

    The vacancy should fully participate in diffusion only when the vacancy concentration is high enough. This is to say, an adequate temperature is needed. In this experiment, 873K is deemed to be the critical temperature to produce enough vacancy, as the eutectic phase can complete spheroidization rapidly as discussed above.

  • (2)

    The reduction of the participation of vacancy in diffusion appears as the decrease of EA,V. The coefficient ‘n’ is used to describe this reduction, and formula (9) should be modified as:

    The coefficient ‘n’ can be estimated relative to the vacancy concentration at certain temperature versus it at 873K, which can be described as:

In this experiment, the first interval is 24h, which is the time allowed for fully spheroidization. Accordingly, the temperature required can be calculated as about 847K, which is the critical temperature between the transient coarsening stage and the fully coarsening stage. This calculation value is in good accordance with experiment results, as the eutectic phase is still in spheroidization process even after 72h at 823K.

3.4The hardness reduction caused by thermal exposure

Fig. 11a exhibits the hardness values of the Al–Fe–Ni alloy exposed to temperatures from 573K to 898K holding for 24h, 48h and 72h, accordingly. Under a certain temperature, the reductions of the HV values can be seen with the prolonging holding duration. In general, the HV values of the Al–Fe–Ni alloy have the similar time-dependent tendency at the temperatures from 573K to 898K. Notably, a sharp decline of HV values can be seen at the first 24h. Meanwhile, it is found that this sharp decline is much severer at a higher exposing temperature. Furthermore, the reduction rates of the HV values from 24h to 72h are enhanced with the growing temperatures.

Fig. 11.

(a) The variation of hardness versus time; (b) fitting of hardness change over time.


Analogized to the Hall–Petch type correlations, it is to generate a relationship between the hardness and particle interval [46,47]:

Since the total volume fraction of the eutectic phase is constant, we can have following approximation relations:


  • Z is amount of particles;

  • C is total volume fraction of Al9FeNi phase.

Finally, the empirical relationship of the hardness and time can be formulated as:

where g is constant independent of time.

A series of curves can be acquired by using this relationship to fit the steady reduction process showed in Fig. 11a. The rationalized results are exhibits in Fig. 11b.

HV0 has been influenced by sharp decline in the first 24h, while g is affected by the coarsening behavior more importantly. The deterioration of alloy can be promoted by the gradually coarsening, which has reflected by the enlargement of g. In Fig. 11b and Table 4, it is found that the g values are stable in lower temperature, and increased remarkably at higher temperatures. This is a good correspondence to the microstructure evolution as discussed above.

Table 4.

The g values at the corresponding temperatures.

Temperature (K)  g 
573  14.587 
673  15.655 
773  15.333 
823  18.22 
873  20.46 
898  20.893 

This work has studied the thermal stability of Al–Fe–Ni eutectic alloy. Based on the experimental results, microstructure characterizations and analytical discussions, the following conclusions can be drawn hereby:

  • (1)

    During the whole temperature range from 573K to near melting temperature, three stages of coarsening process can be divided according to the microstructure evolution of eutectic phase:

    • Stable stage: the eutectic phase has mainly kept stable.

    • Transient coarsening stage: the eutectic phase has experienced the local ablation, coarsening and spheroidization simultaneously.

    • Fully coarsening stage: the eutectic phase should be fully ruptured, spheroidized and then coarsened rapidly.

  • (2)

    The relationship between coarsening coefficient K and temperature is affected by the shape change. This influence is time-dependent, and the choices of experiment time have different effects. Meanwhile, the diffusion caused by morphologic changes of the eutectic phase has played a main role in the whole process, not only affected the coarsening rate but also the spheroidization process.

  • (3)

    The hardness values of the alloys at different temperatures have presented the similar decreasing tendencies. In addition, the shape change and rapid coarsening of eutectic phase have resulted in the accelerated softening of the alloy.

Conflicts of interest

The authors declare no conflicts of interest.


This research was funded by the National Key Research and Development Program of China (Grant No. 2018YFB1106302), the Research Fund (Project No. 15X100040018) at Shanghai Jiao Tong University (China), and the project (Grant No. 2017WAMC002) sponsored by Anhui Province Engineering Research Center of Aluminium Matrix Composites (China).

T. Koutsoukis, M.M. Makhlouf.
Rendering wrought aluminium alloys castable by means of minimum composition adjustments.
Int J Cast Met Res, 30 (2017), pp. 1-13
L.F. Mondolfo.
Aluminum alloys: structure and properties.
Butterworths, (1976),
J.E. Gruzleski, B.M. Closset.
The treatment of liquid aluminum–silicon alloys.
Am Foundry Soc, (1990),
S.Z. Lu, A. Hellawell.
The mechanism of silicon modification in aluminum–silicon alloys: impurity induced twinning.
Metall Trans A, 18 (1987), pp. 1721-1733
F. Stadler, H. Antrekowitsch, W. Fragner, H. Kaufmann, P.J. Uggowitzer.
Effect of main alloying elements on strength of Al–Si foundry alloys at elevated temperatures.
Int J Cast Met Res, 25 (2012), pp. 215-224
M. Zhu, Z. Jian, G. Yang, Y. Zhou.
Effects of T6 heat treatment on the microstructure, tensile properties, and fracture behavior of the modified A356 alloys.
Mater Des, 36 (2012), pp. 243-249
Y. Du, Y.A. Chang, B. Huang, W. Gong, Z. Jin, H. Xu, et al.
Diffusion coefficients of some solutes in fcc and liquid Al: critical evaluation and correlation.
Mater Sci Eng A, 363 (2003), pp. 140-151
E. Sjölander, S. Seifeddine.
The heat treatment of Al–Si–Cu–Mg casting alloys.
J Mater Process Technol, 210 (2010), pp. 1249-1259
K.I. Hirano, R.P. Agarwala, M. Cohen.
Diffusion of iron, nickel and cobalt in aluminum.
Acta Metall, 10 (1962), pp. 857-863
Z.J. Cao, G. Kong, C.S. Che, Y.Q. Wang, H.C. Peng.
Experimental investigation of eutectic point in Al-rich Al–La, Al–Ce, Al–Pr and Al–Nd systems.
J Rare Earths, 35 (2017), pp. 1022-1028
A. Niikura, A. Kawahara, G. Kimura, T. Doko.
Recrystallization behaviour in twin-roll continuous cast Al–Fe–Ni alloy.
Mater Sci Forum, 519–521 (2006), pp. 1635-1640
C. Suwanpreecha, P. Pandee, U. Patakham, C. Limmaneevichitr.
New generation of eutectic Al–Ni casting alloys for elevated temperature services.
Mater Sci Eng A, 709 (2017), pp. 46-54
Z.C. Sims, O.R. Rios, D. Weiss, P.E.A. Turchi, A. Perron.
High performance aluminum–cerium alloys for high-temperature applications.
Mater Horiz, 4 (2017), pp. 1070-1078
R.X. Li, Z. Wang, Z.C. Guo, P.K. Liaw, T. Zhang, L.G. Li, et al.
Graded microstructures of Al–Li–Mg–Zn–Cu entropic alloys under supergravity.
Sci China Mater, (2018), pp. 1-9
H. Bitterlich, W. Löser, L. Schultz.
Reassessment of Ni–Al and Ni–Fe–Al solidus temperatures.
J Phase Equilib, 23 (2002), pp. 301-304
S.M. Hao, K. Ishida, T. Nishizawa.
Role of alloying elements in phase decomposition in AlNiCo magnet alloys.
Metall Trans A, 16 (1985), pp. 179-185
N. Ono, A. Tsukahara, R. Kainuma, K. Ishida.
The properties of two-phase Ni–Al–Fe shape memory alloys in the virgin and shape-memory-cycled states.
Mater Sci Eng A, 273–275 (1999), pp. 420-424
L. Zhang, Y. Du, H. Xu, C.Y. Tang, H.L. Chen, W.Q. Zhang.
Phase equilibria of the Al–Fe–Ni system at 850°C and 627°C.
J Alloys Compd, 454 (2008), pp. 129-135
J.M. Kim, H.S. Yun, J.S. Shin, K.T. Kim, S.H. Ko.
Mold filling ability and hot cracking susceptibility of Al–Fe–Ni alloys for high conductivity applications.
J Teknol, 75 (2015), pp. 71-77
A. Leenaers, S.V.D. Berghe, E. Koonen, S. Dubois, M. Ripert.
Post-irradiation examination of AlFeNi cladded U3Si2 fuel plates irradiated under severe conditions.
J Nucl Mater, 375 (2008), pp. 243-251
M. Wintergerst, N. Dacheux, F. Datcharry, E. Herms, B. Kapusta.
Corrosion of the AlFeNi alloy used for the fuel cladding in the Jules Horowitz research reactor.
J Nucl Mater, 393 (2009), pp. 369-380
M.V. Canté, C. Brito, J.E. Spinelli, A. Garcia.
Interrelation of cell spacing, intermetallic compounds and hardness on a directionally solidified Al–1.0Fe–1.0Ni alloy.
Mater Des, 51 (2013), pp. 342-346
W.J. Boettinger, L.A. Bendersky, R.J. Schaefer, F.S. Biancaniello.
On the formation of dispersoids during rapid solidification of an Al–Fe–Ni alloy.
Metall Trans A, 19 (1988), pp. 1101-1107
B.L. Silva, J.G. Dessi, L.F. Gomes, M.M. Peres, M.V. Cante, J.E. Spinelli.
Assessing microstructures and mechanical resistances of as-atomized and as-extruded samples of Al–1wt.%Fe–1wt.%Ni alloy.
J Alloys Compd, 691 (2017), pp. 952-960
F. Průša, D. Vojtěch, A. Michalcová, I. Marek.
Mechanical properties and thermal stability of Al–Fe–Ni alloys prepared by centrifugal atomisation and hot extrusion.
Mater Sci Eng A, 603 (2014), pp. 141-149
T. Koutsoukis, M.M. Makhlouf.
An alternative eutectic system for casting aluminum alloys I. Casting ability and tensile properties.
Light Met, (2015), pp. 277-281
T. Koutsoukis, M.M. Makhlouf.
Alternatives to the Al–Si eutectic system in aluminum casting alloys.
Int J Metalcast, 10 (2016), pp. 342-347
C.L. Chen, G.D. West, R.C. Thomson.
Characterisation of intermetallic phases in multicomponent Al–Si casting alloys for engineering applications.
Mater Sci Forum, 519 (2006), pp. 359-364
M. Mrotzek, E. Nembach.
Ostwald ripening of precipitates during two successive heat treatments performed at different temperatures.
Acta Metall, 56 (2008), pp. 150-154
I.M. Lifshitz, V.V. Slyozov.
The kinetics of precipitation from supersaturated solid solutions.
J Phys Chem Solids, 19 (1961), pp. 35-50
C.Z. Wanger.
Theory of precipitate change by redissolution.
Z Electrochem, 65 (1961), pp. 581-591
F. Gruy, M. Cournil.
Theoretical aspects of Ostwald ripening in systems with high solubility and high solid volume fraction.
Acta Metall, 44 (1996), pp. 2537-2545
J. Li, C. Guo, Y. Ma, Z.J. Wang, J.C. Wang.
Effect of initial particle size distribution on the dynamics of transient Ostwald ripening: a phase field study.
Acta Metall, 90 (2015), pp. 10-26
V.A. Snyder, J. Alkemper, P.W. Voorhees.
Transient Ostwald ripening and the disagreement between steady-state coarsening theory and experiment.
Acta Metall, 49 (2001), pp. 699-709
G.R. Speich, R.A. Oriani.
The rate of coarsening of copper precipitates in an alpha-iron matrix.
Trans Metall Soc AIME, 233 (1965), pp. 623-630
M.K. Premkumar, A. Lawley, M.J. Koczak.
Processing and microstructure of powder metallurgy Al–Fe–Ni alloys.
Metall Trans A, 23 (1992), pp. 3219-3230
A.D. Setyawan, D.V. Louzguine, K. Sasamori, H.M. Kimura, S. Ranganathan, A. Inoue.
Phase composition and transformation behavior of readily solidified Al–Ni–Fe alloys in α-Al-decagonal phase region.
J Alloys Compd, 399 (2005), pp. 132-138
K.H. Kim, M.S. Kim.
High temperature tensile properties of hypereutectic Al–25Si based alloy.
Met Mater Int, 24 (2018), pp. 136-141
A.D. Brailsford, P. Wynblatt.
The dependence of Ostwald ripening kinetics on particle volume fraction.
Acta Metall, 27 (1979), pp. 489-497
R.W. Heckel, R.D. Gregorio.
The growth and shrinkage rates of second-phase particles of various size distributions – II. Spheroidization of a eutectoid steel.
Trans Metall Soc AIME, 233 (1965), pp. 2001-2011
K. Kim, P.W. Voorhees.
Ostwald ripening of spheroidal particles in multicomponent alloys.
Acta Metall, 152 (2018), pp. 327-337
E. Ogris, A. Wahlen, H. Lüchinger, P.J. Uggowitzer.
On the silicon spheroidization in Al–Si alloys.
J Light Met, 2 (2002), pp. 263-269
G.M. Hood.
The diffusion of iron in aluminium.
Philos Mag, 21 (1970), pp. 305-328
G.F. William Jr..
Correlation between frequency factor and activation energy for diffusion in aluminum.
Scr Metall, 6 (1972), pp. 1139-1144
S.H. Zhang, X.L. He, Y.Y. Chu, J. Ke.
Behaviour of vacancy influence during diffusion in metals.
Acta Metall Sin (Engl Lett), 5 (1992), pp. 395-401
B.L. Silva, I.J.C. Araujo, W.S. Silva, P.R. Goulart, A. Garcia, J.E. Spinelli.
Correlation between dendrite arm spacing and microhardness during unsteady-state directional solidification of Al–Ni alloys.
Philos Mag Lett, 91 (2011), pp. 337-343
B.L. Silva, A. Garcia, J.E. Spinelli.
The effects of microstructure and intermetallic phases of directionally solidified Al–Fe alloys on microhardness.
Mater Lett, 89 (2012), pp. 291-295
Copyright © 2019. Brazilian Metallurgical, Materials and Mining Association
Journal of Materials Research and Technology

Subscribe to our newsletter

Article options
Cookies policy
To improve our services and products, we use cookies (own or third parties authorized) to show advertising related to client preferences through the analyses of navigation customer behavior. Continuing navigation will be considered as acceptance of this use. You can change the settings or obtain more information by clicking here.