Fabrication of Ni, Si and Mg Supersaturated Solid Solutions in Al by Mechanical Alloying

Corresponding Author: A. Molina Centro de Investigación en Ingeniería y Ciencias Aplicadas, Universidad Atuonoma del Estado de Morelos, Mexico Email: arturo_molina@uaem.mx Abstract: In this work, a comparison between the synthesis process of nanostructured Al base alloys modified with Si, Ni and Mg and the formation of their respective supersaturated solid solution is presented. The samples were obtained via Mechanical Alloying (MA) at a speed of 300 RPM for 2, 5, 10, 20 and 30 h under an inert Ar atmosphere. Lattice parameter evolution, semiquantitative X-Ray Diffraction Analysis and SEM micrographs were correlated to determine the attained solubility in the Al matrix. The results showed that powder refinement and solubility rates are directly affected by the hardness of the solute element and the limiting solubility in equilibrium state respectively. Therefore the Al-Mg alloy achieved the highest solubility of the samples after 10 h of milling, meanwhile, the Al-Si alloy showed the lowest solubility even after 20 h of milling. Microstructural analyses by X-Ray Diffraction revealed that the lattice parameter showed different behavior depending on the alloy system, in some cases showing a deviation from the behavior predicted by Vegard's law. Finally, this work demonstrates that the successful synthesis of supersaturated solid solutions via MA is dependent on the alloy system as well as milling conditions.


Introduction
New technological developments require advanced materials with properties such as lightweight, high strength, fatigue resistance and corrosion resistance (Atul et al., 1993;Heinz et al., 2000;Yvon and Carré, 2009), but on the other hand, it's desirable that they are low cost to produce with raw material availability. Good candidates include microstructurally designed Al alloys, capable of being tailored to withstand harsh environments while exhibiting good mechanical properties essential for advance engineering applications (Reed et al., 2009;Nie and Muddle, 1998;Vajpai et al., 2015). Some of the techniques employed include Solidsolution strengthening, High-pressure torsion, large strain extrusion and spark plasma sintering, these techniques have shown to be capable of producing materials such as Ni-based super alloys (Takizawa et al., 2015), as well as Al, Mg and steel alloys with ultrafinegrained structures and superior strength (Figueiredo and Langdon, 2012) and Li-ion batteries with grains conserved in the nano-scale range that enhances conductivity (Kali and Mukhopadhyay, 2014). Therefore, the ability to control the microstructure of the materials paves the way to future technological advances capable of revolutionizing aerospace, automotive, medical and production industries (Schinhammer et al., 2010;Brenne et al., 2016;Burger et al., 1996).
Supersaturated Solid Solutions (SSS) consist in alloys with solute contents that exceed the solid solubility limit under equilibrium conditions (Suryanarayana and Froes, 1992), this higher content of solute elements in the crystal lattice have shown to improve the mechanical properties of the base metal trough mechanisms such as solid solution strengthening, precipitation strengthening, grain boundary strengthening and high dislocation densities (Suryanarayana, 1996). The main advantage of materials with such strengthening mechanisms is that they can be designed microstructurally for their intended application (Eskin, 2003), this is done via different heat treatment regimens that produces metal matrix composites (MMC) with a refined microstructure and a controlled precipitation of second phases capable of withstanding harsh environments while exhibiting high strength, corrosion resistance and good wear resistance. Therefore, the development of new low weight Al alloys with improved mechanical properties compasses the investigation of SSS, their heat treatment regimens and the mechanisms involved in their synthesis via different techniques. Different techniques to produces SSS alloys have been widely studied, but the Mechanical Alloying (MA) process has been continuously proved to be a reliable synthesis method with some added benefits. Contrary to other methods such as the Rapid Solidification Process (RSP), the MA process is capable of providing greater solubility extension and highly refined homogenous powders (Patra et al., 2016); this is obtained through the constant collisions exerted on the milled powder that creates nonequilibrium conditions (Shingu and Ishihara, 1996). Overall, the MA process consists of a more flexible method to produce SSS but one that requires the optimization of the milling parameters to produce the required system conditions to obtain SSS alloys.
Previous works have shown that annealing treatment of powders after milling or high temperatures during milling inhibit the formation of SSS as it produces the precipitation of second more stable phases (Schwarz, 1995). Other studies indicate that varying other milling parameters such as the use of Process Control Agents (PCA) and controlled atmospheres impact greatly the synthesis of SSS (Li et al., 2010). There is also contradicting evidence concerning the use of Hume-Rothery rules to predict the formation of SSS for the MA process (Moumeni et al., 2006;Bansal et al., 1994). Comparative studies of RSP (Suryanarayana, 2018;Mukhopadhyay et al., 2008) with MA still have not concluded the difference in mechanisms that produce contradictory behaviors for the same alloying systems. Thus, there is still a knowledge gap concerning the driving forces behind the formation of SSS via MA and a reliable method that helps us to predict the successful synthesis of SSS.
This work aims to compare the mechanisms behind the synthesis of SSS for different alloying systems (Al-Si, Al-Ni and Al-Mg) under the same milling parameters. The rate of solubility, particle refinement, lattice strain evolution and microstructural properties are examined and compared. This is intended for determining the alloy system characteristics that affect the synthesis process via MA and differentiate the behavior demonstrated to that as expected by other techniques. In future works, this may help in the selection of milling parameter (such as milling time and speed regime) to successfully achieve the synthesis of SSS.

Materials and Methods
Aluminum, nickel, silicon and magnesium powders with purity above 99.5% and an average particle size of 50 microns were milled in tool-grade steel vials with stainless steel balls and with a Ball to Powder Ratio (BPR) of 10:1. The milled powder mixtures consisted of Al-3Ni (wt.%), Al-3Si (wt.%) and Al-3Mg (wt.%) with 0.3 wt. % of stearic acid added as Process Control Agent (PCA) (Fogagnolo et al., 2006). A low-energy planetary ball mill (FRITSCH, Pulverisette model) was used for milling the powders at 300 rpm for 2, 5, 10, 50 and 30 h under an argon atmosphere to protect the obtained powders from reacting.
The resulting powder samples were characterized by Scanning Electron Microscopy (SEM) with a LEO 1450 equipment to determine the microstructural evolution and powder properties during the milling process with the aid of an image processing program (ImageJ) for measuring the perimeter of the observed particles ( Fig.  1) to calculate the equivalent circular perimeter diameter according to Equation 1: Where: P d = The equivalent circular perimeter diameter [µm] P p = The particle perimeter [µm] X-Ray Diffraction (XRD) patterns were obtained with a BRUKER D2 PHASER diffractometer using CuK α radiation over a 2-theta range of 20 to 80 degrees. Calculation of lattice parameter from Bragg's Law as described in Equation 2 where performed for solubility analysis according to Vegard's Law (Equation 3) and compared to those obtained by semiquantitative analysis.

Scanning Electron Microscopy Imaging
Initial powders SEM micrographs (Fig. 2) show different morphological characteristics for the powders, consisting in spherical for those of Al (Fig. 2a), amorphous for those of Mg (Fig. 2d) and irregular for those of Ni (Fig. 2b) and Si (Fig. 2c) powders. Their difference in morphology is due to the use of distinct production methods, Al powders where atomized meanwhile, machining was used for those of Ni and Si and mechanical milling processing was used for the Mg samples. Morphological properties such as particle shape, particle size, particle size distribution, agglomerates, etc. have been shown to affect principally the initial stages of milling (Parvin et al., 2008) since, after continuous impacts, the powders converge into a predominantly amorphous morphology (Fogagnolo et al., 2003).
After 30 h of milling, the morphology of the powders became predominantly amorphous for the Al-Si (Fig. 3a) and Al-Ni system (Fig. 3b) and predominantly quasispherical for the Al-Mg system (Fig. 3c). The hardness of initial powders affect the evolution of the system, therefore a ductile-brittle system has been shown to result in higher cold work and fracture than that of a ductile-ductile system (Gheisari et al., 2009). Hence in the Al-Si and Al-Ni samples, more amorphous final powders are observed; meanwhile, the Al-3% Mg sample shows that the magnesium high cold-work hardening rate and ductility gave resulted in an enhanced particle refinement while maintaining a very similar morphology to that of the Al initial powders. This higher refinement in the ductile-ductile system is congruent with previous studies that have compared the morphological evolution of different alloying systems by MA (Zhou and Rao, 2004). Figure 4 shows that the more efficient system for homogeneous particle refinement was that of the Al-3% Mg, since it displays a narrower particle size distribution represented by the upper and lower quartiles. In contrast to this, the less homogeneous powders was the Al-3%Ni sample, which displays higher upper and lower quartiles difference as well as higher maximum and minimum values. Nonetheless, the Al-Mg system obtained the highest average particle size; this is behavior is attributed to the ductility of Mg and Al, which provides a MA process dominated by cold-welding. It is worth mentioning, that the Al-Ni system displayed low homogenization in particle size, probably related to the irregular size of the initial Ni powders.

Phase Identification
XRD patterns revealed that during the MA process, the peaks corresponding to the solute elements disappeared at different times depending on the alloying system, the Al-Mg XRD pattern (Fig. 5a) shows that after 10 h of milling time, the Mg peaks are barely noticeable, semiquantitative analysis showd that the observable area under the curve after 30 h of milling corresponds to 0.23 wt.% of Ni. A small content of Al 2 O 3 as a secondary phase is observable after 20h. As it does not contain Mg, it should not affect notoriously the Mg content dissolved in the Al Matrix, therefore by semiquantitative analysis, the Al-Mg system achieved supersaturation (above 1.2 wt% of Mg solubilized) after 10 h of milling. The next system that presented the most noticeable reduction in solute element peaks was the Al-Ni system (Fig. 5b), in this case, the superposition of Al and Ni peaks (near 44.5 2-theta degrees) required the deconvolution of the XRD patterns, the semiquantitative analyses showed that as Ni has 0% solubility at room temperature, the supersaturation was successful even after the initial stages of milling furthermore, it was possible to obtain approximately 2.78% of solubility. Finally, the Al-Si patterns (Fig. 5c) showed during all the milling stages the presence of Si peaks, agreeing with this result, the semiquantitative analysis showed that it achieved the less amount of solubility in the Al matrix.

Crystallite Refinement
To fully study the phenomenon occurring during the MA process, it is necessary to compare crystallite size evolution and lattice strain. Grain size refinement increased with milling time in all the alloying systems studied in this work and can be observed in the broadening of the Al peaks in XRD patterns. A subtle variation of the phenomenon between the samples is shown in the Williamson-Hall plots (Fig. 6). This can be attributed to the low concentration of the alloying elements. Most noticeable, all the samples achieved almost maximal crystallite size refinement after 20 h of milling time, displaying a balance of the cold welding and fracture events after the early stages of milling (10 h). This resulted in the typical asymptotic behavior of the mechanical alloying process (Cocco et al., 2000) that can be seen most noticeable in the Al-3Si (wt.%) system in Fig. 6c probably due to the higher hardness value of Si and high melting point that hinders further dislocation creation and movement (Garroni et al., 2014). The minimum grain sizes obtained were 70 nm for the Al-3Ni (wt.%) system, 55 nm for the Al-3Mg (wt.%) system and 47 nm for the A-3Si (wt.%) system, these values are in agreement with those corresponding to the plot published by (Koch, 1993) (Fig. 7) in which the minimum grain size of various alloys obtained by the MA technique were shown to display an inverse relationship with their melting point, the minimum grain size of the obtained values in this work are shown with a positive deviation in the graph due to low milling energy conditions, this statement holds true also when taking into account the best precision of the XRD technique in crystallite size determination (±10%) as discussed by (Al-Aqeeli et al., 2008).
Lattice strain calculated by Williamson-Hall plots showed to increases as milling time elapsed (Fig. 6), with some noticeable relaxation phenomenon in the Al-3Mg (wt.%) system after 10 h of milling time, this behavior has been shown to be related to accumulation of defects, crystal lattice reordering and heating during the MA process as shown by Cryogenic Milling with higher lattice strain values and lower occurrence of lattice relaxation (Zhou et al., 2003). It is observable a maximum increase in lattice strain at initial milling stages for all the alloying systems studied in this work, this behavior is due to the increase in work hardening experienced during the MA process that translates into higher energy required to deformation to take place (Azimi and Akbari, 2011) and therefore it can be observed slower rates of lattice strain as milling time elapses.    (Koch, 1993) Lattice Parameter Evolution Lattice parameter calculated during milling time is observed in Fig. 8 and can be inferred from the displacement of XRD patterns peaks. Both the Al-3Ni system and Al-3Si system show oscillations in the lattice parameter that has been previously attributed to contamination (Blázquez et al., 2009), reordering and formation of second phases (Kim et al., 1996). Meanwhile, stabilization of the lattice parameter expected in Solid Solutions seems to be only obtained in the Al-Ni system, showing only a small fluctuation in the late stages of milling (30 h).
The usual method to study the solubility in SSS is through the use of the Vegard's law and lattice parameter calculations. Figure 9 shows the relationship given by Vegard's law between lattice parameter and solubility of the alloying element.
The calculated lattice parameter for the Al-Mg system showed a negative deviation from that expected by Vegards law, therefore the final lattice parameter value corresponds to that of a solubility of 3.2 wt.% of Mg, meanwhile the Al-Ni and Al-Si systems behaved with no deviations and the obtained solubility values were 2% and 1.5% respectively. Furthermore, the Al-Mg system does no presents an asymptotic behavior, rather a linear behavior. Meanwhile, the other samples present an asymptotic behavior characteristic of the MA process. This can be attributed to the enhanced mechanical alloying process in ductile-ductile systems once the fracture-welding process reaches equilibrium (Aizawa and Song, 2016).

General Discussion
XRD patterns and SEM micrographs show the effects of the collisions imparted by the balls onto the powders during the mechanical alloying process. This translates into cold welding of the particles during the initial stages of the process, with modest particle size and morphology changes. Posterior to this, the energy provided by the collisions fractures the powder and a balance between cold working and fracture is reached in the late stages (Murty and Ranganathan, 1998). The powder processing, therefore, results in particle size and grain size refinement, observable trough the peak broadening of the XRD patterns and corroborated by the Williamson Hall plots calculations, as well as homogenization of the characteristics of the powders during the late stages of milling (Rojas et al., 2006).
The solubility of the alloying elements was approximated by XRD semiquantitative analyses and showed to be congruent with those obtained by Vegard's Law calculations from the obtained lattice parameters. The negative deviation from Vegard´s Law in the Al-Mg system and the presence of a small Al 2 O 3 peak in the XRD pattern seems to indicate Oxygen contamination during the MA process even under an inert Ar atmosphere. This behavior is similar to those reported by (Botcharova et al., 2003) in which precautions were incapable of preventing contamination from the milling media and environment. Another common phenomenon associated with the contamination of powders can be seen in the oscillation of the lattice parameter and lattice strain during the late stages of milling.

Conclusion
In this work is demonstrated the viability of producing SSS of Al-3wt%Si, Al-3wt%Mg and Al-3wt%Ni by mechanical alloying under low energy milling conditions. Although difficult to obtain, semiquantitative analysis of XRD patterns can be used for guidance of the solubility phenomenon in the MA process, but further precision requires the use of lattice parameter calculations and Vegard's Law.
Vegard's Law is prone to small errors due to contamination from the milling environment, but it nonetheless, it's still plays an important and relevant role in the study of solid solutions obtained by the MA technique.
The solubility rate obtained during the processing of the powders is completely dependent on the alloy system even with low alloying element content.
The mechanisms behind the synthesis of SSS via the MA process are unique to this method and therefore it is incorrect to use results obtained from other processes to infer the expected behavior trough the MA process.
Finally, the results obtained showed that even low concentrations of different alloying elements are capable of affecting the evolution of particle and grain refinement.

Acknowledgement
This work was financed by CONACyT [Grant Number 402858]; the authors would like to thank sincerely Rene´ Guardian Tapia for the support provided in SEM images.

Author's Contributions
A. Sedano: Designed the research plan, all experimental development of the research, analysis of results and writing of the paper.
A. Molina: Designed the research plan, process of mechanical alloying and mechanical milling, analysis of results and writing of the paper. S.A. Serna: Participated in the microstructural characterization of the alloys and analysis of results and writing of the paper. R.A. Rodríguez-Díaz: Performed the X-ray diffraction analysis and analysis of results.
A. Torres-Islas: Participated in the microstructural characterization of the alloys and analysis of results.

Ethics
This article is original and contains unpublished material. The corresponding author confirms that all of the other authors have read and approved the manuscript and no ethical issues involved.