In this work, the effect of a slight change in the composition of the Al-Fe amorphous alloys (from Al84Fe16 to Al82Fe18) and the substitution of Y for Al (2 at.%) on their crystallization kinetics was studied. According to the X-ray diffraction analysis, powders of the Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloys with a fully amorphous structure were formed after 100 h of mechanical milling of the mixtures of the elemental powders. The crystallization behavior of the alloys was also studied by transmission electron microscopy. Upon heating up to a temperature of the first exothermic peak, α-Al crystals precipitated from the amorphous Al84Fe16 matrix. During crystallization of the Al82Fe18 alloy, crystals of the Al6Fe intermetallic compound formed along with α-Al crystals. Substitution of Y for 2 at.% of Al in the Al82Fe16Y2 alloy made crystallization of the alloy more complicated: α-Al, Al6Fe, and Fe4Y crystals coexisted with an amorphous phase. The activation energies corresponding to the first crystallization event of the alloys were calculated using the Kissinger and Ozawa methods. The values obtained by these two methods were in good agreement with each other and the same trends of changing with the alloy composition were observed. The Avrami exponent was determined from the Johnson-Mehl-Avrami equation and showed that crystallization at the first stage is interface-controlled growth for all the three powder alloys studied.

1. Introduction

Amorphous aluminum-based alloys containing ≥80 at.% Al constitute a group of materials promising for structural applications. The main advantages of Al-based metallic glasses are high strength, low density, high elastic strain, and good corrosion resistance [1, 2]. At the same time, irons aluminides have attracted interest due to high specific strength and excellent corrosion resistance at elevated temperatures under oxidizing, carburizing, and sulfurizing environments [3]. Although metallic glasses possess unique properties, they are metastable and tend to crystallize during continuous heating. In order to produce bulk materials retaining an amorphous structure or bulk nanostructured composites containing nanometer-sized crystals dispersed in amorphous matrices, it is necessary to determine the stability of metallic glasses during thermal treatment and evolution of their structure upon crystallization [46]. The driving force for crystallization of amorphous materials is the Gibbs free energy difference between the amorphous and the crystalline states. The crystallization mechanisms and the structure and composition of the crystallization products depend on the initial chemical composition of the amorphous phase and its preparation method. Studies of the crystallization behavior of amorphous alloys are necessary for evaluating their thermal stability against crystallization, determining the structure evolution paths of the amorphous alloy powders upon consolidation and for elucidating the fundamentals of the processes of the glass formation, nucleation, and growth.

In the present work, the kinetics of crystallization of the mechanically alloyed amorphous Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloy powders was studied using differential scanning calorimetry (DSC), X-ray diffraction (XRD), and transmission electron microscopy (TEM).

2. Experimental

Amorphous alloy powders with nominal compositions of Al84Fe16, Al82Fe18, and Al82Fe16Y2 were synthesized via mechanical alloying [7]. Mixtures of Al, Fe, and Y were milled in a planetary ball mill (P100, South Korea) using hardened steel vials and hardened steel balls of 5 mm diameter. The rotation speed of the disc was 350 rpm and the ball to powder weight ratio was 20 : 1. Hexane was used as a process control agent to prevent extensive agglomeration and contamination during mechanical milling. DSC was conducted using NETZSCH STA 409C at heating rates of 5, 10, 20, and 40 K/min to study the crystallization kinetics of the amorphous alloys. XRD was performed using RIGAKU RINT-2000 diffractometer with CuKα radiation (λ = 0.15405 nm) to study the phase transformations during mechanical milling and upon subsequent heating. The microstructure of the powders was studied by TEM carried out using a JEOL JEM-2100 microscope.

3. Results and Discussion

The XRD analysis of the powders shows that the amorphous Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloys formed after 100 h of milling. The XRD patterns of the powders (Figure 1(a)) exhibit only broad maxima between 35° and 50° (2θ) indicating that amorphization took place during milling [7]. In order to study the structural evolution during heating, the amorphous alloy powders were annealed in the DSC using continuous heating at 20 K/min up to different temperatures to reach different crystallization events revealed by exothermic peaks and then cooled down to room temperature at 50 K/min. The DSC scans in Figure 2 show the first crystallization peaks of the three alloys recorded at 5, 10, 20, and 40 K/min. The onset temperatures, , of the first crystallization peak and the peak temperatures, , of the Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloys are presented in Table 1. It can be seen that a slight change in the composition of the Al-Fe amorphous alloys, from Al84Fe16 to Al82Fe18, and substitution of Y for Al (2 at.%) result in an increase of both and .

Figure 1(b) shows the XRD patterns of the Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloys heated in DSC up to a temperature corresponding to completion of the first crystallization event. Figures 35 show TEM images of the Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloys that experienced the first crystallization event along with selected-area diffraction (SAD) patterns. The contrast observed in the TEM images is due to separation of the alloys into solute-enriched and solute-depleted amorphous regions. The segregation of Al was clearly observed in the TEM images of all three alloys. The α-Al nanocrystals then preferentially nucleated at the interface between the separated phases. In the Al84Fe16 alloy, only α-Al crystals separated from the amorphous phase and had a size ranging from 10 to 30 nm (Figures 3(a) and 3(c)). The SAD pattern in Figure 3(b) corresponds to an amorphous phase and that in Figure 3(d) to crystalline α-Al. The crystallization behavior the Al82Fe18 was slightly different from that of the Al84Fe16. The Al82Fe18 alloy remained partially amorphous, as can be seen from Figures 4(a)-4(b). In the areas that underwent crystallization, crystals of the Al6Fe intermetallic were found together with crystals of α-Al (Figures 4(c)-4(d)). Due to a low concentration of the intermetallic compound, the Al6Fe phase could not be detected by the XRD [7]. The concentration of the crystalline phases separated from the amorphous phase (Figure 4(a)) in the Al82Fe18 is higher than in the Al84Fe16 alloy. During heating of the Al82Fe16Y2 alloy, crystals of the Fe4Y intermetallic formed, as can be concluded from the TEM data (Figure 5). The nucleation and growth of the α-Al phase occurred in a manner similar to that observed during crystallization of Al88RE8Ni4 amorphous alloys [8].

It is known that the more complex the composition of the crystallites precipitating from an amorphous phase, the higher the thermal stability of the alloys [9]. From the DSC analysis of the three alloys, it can be concluded that their thermal stability increases in the following row: Al84Fe16-Al82Fe18-Al82Fe16Y2. The thermal stability of amorphous alloys is often interpreted using the heat of mixing . The more negative the , the larger the atomic constraint force and the higher the thermal stability. between Fe and Al (solvent) is −11 kJ/mol, while between Y and Al (solvent) is −38 kJ/mol [10]. As between Y and Al as a solvent is more negative, should increase with increasing Y content in the alloy.

One of the most important kinetic parameters of the crystallization process is the activation energy. The thermal stability of glassy materials is related to the activation energy values. The Kissinger [11] and Ozawa [12] methods are used to calculate the activation energy of the amorphous to crystalline phase transformation in nonisothermal crystallization processes. Kissinger proposed a method for determining the activation energy of a simple decomposition reaction using differential thermal analysis scans recorded at different heating rates [9]. The heating rate of a reaction is related to the peak temperature recorded during the thermal analysis. By considering the variation of each peak temperature with the heating rate β, the activation energy can be calculated by the Kissinger method using the following equation:where is the activation energy of crystallization, is the temperature of the exothermic peak, β is the heating rate, and is the gas constant.

Figure 6 presents the plots of versus () and straight lines are obtained with slopes for the first exothermic peak. From these data, the activation energy values for the first crystallization event of the Al84Fe16, Al82Fe18, and Al82Fe16Y2 amorphous alloy alloys were calculated to be 496.0, 340.8, and 565.0 kJ/mol, respectively.

In order to confirm the results obtained using the Kissinger method, another nonisothermal method was applied to calculate the activation energy, the Ozawa method, which uses the following equation:where β, , , and are the heating rate, peak temperature, activation energy, and gas constant, respectively.

Figure 7 shows the plots of against , and was calculated from the slope of the straight lines. The values of obtained using the Ozawa equation are 506.6, 351.9, and 576.1 kJ/mol for the Al84Fe16, Al82Fe16, and Al82Fe16Y2 amorphous alloy powders, respectively. The activation energies for crystallization of the Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloys calculated by the Kissinger and Ozawa methods using the onset crystallization temperature and peak temperature of the first exothermic peak are summarized in Table 2. The values of the activation energies calculated from the Ozawa equation are only slightly higher than those obtained from the Kissinger equation.

It is known that the onset crystallization and peak temperatures are associated with the nucleation and growth processes, respectively [13]. In the case of the Al84Fe16 alloy, the obtained activation energy for nucleation (626.8 kJ/mol) is higher than the activation energy for growth (496.0 kJ/mol), indicating that the nucleation process for the Al84Fe16 alloy is more difficult than the growth process. This can be seen from the TEM image in Figure 3(a): only a few crystals of Al nucleated and their crystallite size was from 10 to 30 nm. However, in the case of the Al82Fe18 and Al82Fe16Y2 alloys, the activation energies for nucleation (284.4 and 464.9 kJ/mol) are lower than the activation energies for growth (340.8 and 565.0 kJ/mol), indicating that the nucleation process occurs easier than the growth process. As was confirmed by TEM (Figures 4(a) and 4(c)), many crystals precipitated from the amorphous phase in the Al82Fe18 alloy, and the crystallite size of Al was less than 10 nm. The crystallite size of Al in both crystallized Al82Fe18 and Al82Fe16Y2 alloys was less than 10 nm, which is smaller than the size of the Al crystallites formed in the Al84Fe16 alloy. This can be explained by the influence of the intermetallic phases, Al6Fe and Fe4Y, which restrain growth of the Al crystals. An increase in the Fe content and partial substitution of Y for Al make the nucleation process easier (compare the values of 284.4, 464.9, and 626.8 kJ/mol for Al82Fe18, Al82Fe16Y2, and Al84Fe16 alloys, resp.).

An increase in the Fe content in the alloy makes the growth process easier, while partial substitution of Y for Al alloy hinders the growth process. Yttrium has a larger atomic radius (1.8 Å) than Al (1.18 Å) [10]. Larger Y atoms increase the potential barrier and hinder diffusion of atoms during the crystallization process of the amorphous alloys. Consequently, the activation energy increases when the substituting atoms have larger radii, implying that, in the Al82Fe16Y2 alloy, a higher energy barrier must be overcome for atomic rearrangement and diffusion of atoms in the crystallization process. The values of the apparent activation energies strongly depend on the type of the primary phase. In the Al84Fe16 alloy, the primary phase is α-Al; in the Al82Fe18 and Al82Fe16Y2 alloys, the primary phases are the α-Al and the intermetallic phases.

Generally, the crystallization kinetics of amorphous alloys is studied using the John-Mehl-Avrami equation [14]:where is the crystallization volume fraction at time , is the Avrami exponent, and is the reaction rate constant related to the absolute temperature described by the Arrhenius equation:where is a constant, is the activation energy, is the gas constant, and is the absolute temperature.

In nonisothermal DSC, the heating rate is controlled, so the temperature can be expressed aswhere is the starting temperature.

Combining (3) and (4), the crystallized volume fraction can be plotted against the temperature using the following equation:where and are the initial and final crystallization temperatures of the crystallization peak, respectively. Figure 8 presents the plots of the crystallized volume fraction, , as a function of the temperature at different heating rates. All these curves show S-shape indicating crystallization of the amorphous alloy.

A method to determine the Avrami exponent was proposed by Ozawa [12]:where the crystallized volume fraction, , at any selected temperature is calculated from (6).

Using equation (7) and values of at the peak temperature of the first crystallization event from Figure 8, the Avrami exponent can be determined. For the Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloys, the Avrami exponent was calculated to be 0.88, 0.85, and 0.83, respectively.

An alternative method [14] to obtain the Avrami exponent is through the activation energy calculated by the Kissinger method [11]:

Plots of versus for ranging from 15% to 85% at different heating rates are shown in Figure 9 (the Johnson-Mehl-Avrami (JMA) plots). The Avrami exponent was obtained from the slopes of these plots. The values for the Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloys calculated by using (8) are 0.94, 0.90, and 0.82, respectively. The values of the Avrami exponent obtained by these two methods are listed in Table 3.

The Avrami exponent provides information on the nucleation and growth mechanism of new crystalline grains during the phase transition. The values of the Avrami exponent for the Al84Fe16, Al82Fe18, and Al82Fe16Y2 alloys are below 1 and close to 1. The Avrami exponent (originally expected to be , 2, 3, 4 [14, 15]) can also have noninteger fractional values corresponding to physically complex microstructural evolution. The obtained values of suggest that the transformation at the first stage includes structural changes and homogeneous reactions with the interface-controlled growth [15, 16]. These crystallization mechanisms are consistent with the conclusions drawn from the data obtained by TEM.

4. Conclusions

In this study, the effect of the Fe/Al ratio and Y substitution for Al in the Al-Fe and Al-Fe-Y amorphous alloys (Al84Fe16, Al82Fe18, and Al82Fe16Y2) on their crystallization kinetics was investigated. Higher crystallization onset temperatures of the first crystallization peak with an increase in the Fe content and substitution of Y for Al indicated that the thermal stability of the alloys was improved. TEM studies showed that only α-Al crystals precipitated from the amorphous phase in the Al84Fe16 alloy. An increase in the Fe content resulted in the formation of the crystals of the Al6Fe intermetallic, as was concluded from the TEM analysis of the Al82Fe18 alloy. There were three crystalline phases precipitating from the amorphous phase in the Al82Fe16Y2 alloy: α-Al, Al6Fe, and Fe4Y. The values of the activation energies for the first crystallization stage of the amorphous alloys were calculated using the Kissinger and Ozawa methods. These two methods gave values that were in good agreement with each other and revealed the same trends of the changes of the activation energy with the alloy composition. The activation energy for nucleation increases in the Al82Fe18-Al82Fe16Y2-Al84Fe16 order. The activation energy for growth increases when two Y atoms substitute for two Al atoms and decreases when two Fe atoms substitute for two Al atoms in the Al84Fe16 amorphous alloy. The Avrami parameter for the first crystallization peak of all three amorphous alloys was lower than 1, suggesting that the transformation was interface-controlled growth, which was consistent with the results of the TEM analysis. It was found that a small change in the radius of the substituting element influences the structure of the crystallization products.

Competing Interests

The authors declare that they have no competing interests.


This research is funded by Vietnam National Foundation for Science and Technology Development (NAFOSTED) under Grant no. 103.02-2012.19.