Characterization of Mechanically Milled and Spark Plasma Sintered Al 2124-CNT Nanocomposites

In the present work, ball milling and spark plasma sintering were used to develop Al2124-CNT nanocomposites. The effect of milling time on the grain size and lattice strain of the ball milled Al2124 alloy powder and the effect of sintering time and temperature on the grain size of the matrix in spark plasma sintered Al2124 alloy and CNT-reinforced Al2124 nanocomposites were investigated. The density and hardness of the developed materials were evaluated as functions of the sintering parameters. It was found that ball milling not only reduced the particle size of the Al2124 powder but also decreased the grain size of the αaluminum phase to 50 nm and increased its lattice strain. A milling time of 6 hours was found to be the optimum time to reach a nanostructured α-aluminum matrix. The grain size of the αaluminum phase in the sintered samples increased with increasing sintering temperature and time to reach maximum values at a sintering temperature of 500°C and a sintering time of 20 minutes. Although sintering led to grain growth, the grain size of the α-aluminium matrix remained in the nanometer range and did not exceed 150 nm. The relative density and hardness of the sintered samples increased with increasing sintering temperature and time to reach maximum values at a sintering temperature of 500°C and a sintering time of 20 minutes.


Introduction
The inadequacy of metals and alloys in providing both strength and stiffness to a structure has led to the development of metal matrix composites (MMCs) [1], where rigid ceramic reinforcements are embedded in a ductile metal or alloy matrix.MMCs combine metallic properties, i.e. ductility and toughness, with ceramic characteristics, i.e. high strength and stiffness.Attractive physical and mechanical properties such as high specific modulus, strength-to-weight-ratio, fatigue strength, thermal stability and wear resistance can be obtained with MMCs [2,3].These composites are extensively used in automobile and aerospace applications.Further improvement of the properties of MMCs was possible through the use of nanoreinforcements and/or nanostructured matrices having particle size and grain size, respectively, less than 100 nm.This led to the development of Metal Matrix Nanocomposites (MMNCs) [4].However, there are challenges associated with processing nanocomposites with the desired properties [5].Achieving a uniform distribution/dispersion of the nano-size reinforcement phase is not easy using liquid-processing methods.Fortunately, the use of solid-state processing methods such as mechanical milling or mechanical alloying [6][7][8][9][10][11][12] not only permitted the development of homogenous nanocomposite powders [4, 9,] with uniform distribution of the nano-size reinforcement but also enabled the synthesis of nanostructured matrices such as copper [13][14][15], nickel [14], tungsten [16], cobalt [17], magnesium [18], Al-Mg [19], and aluminum [20,21].The use of conventional sintering methods such as hot pressing, high temperature extrusion, and hot isostatic pressing to consolidate nanostructured powders often results in grain growth which affects the properties of the end product.Preventing or at least minimizing grain growth to maintain the nanostructure features of the matrix is possible through careful control of the consolidation parameters, particularly heating rate, sintering temperature and time.In this regard, spark plasma sintering (SPS), also known as field assisted sintering (FAST), has been shown to be an effective non-conventional sintering method for obtaining fully dense materials [22][23][24].The SPS is not only a binder-less process but also does not require a precompaction step.The powder is directly filled into a graphite die, through which the current is passed and to which the pressure is applied.It was used to consolidate deferent types of metals, alloys, and metal matrix nanocomposites [25][26][27][28][29][30][31][32].Aluminum alloys [33,34] have low weight and good properties; as a result they are used in many engineering applications including automotive and aerospace.Their properties can be improved with addition of either micron-sized reinforcements such as SiC [35] and Al 2 O 3 [36] or nano-sized reinforcements such as SiC [37][38][39][40] and carbon nanotubes (CNTs) [41][42][43].Although, a few published works reported the development of CNT-reinforced aluminum nanocomposites using mechanical milling and spark plasma sintering [27,29,32,44], the developed composites were not fully characterized.The objective of the present work is to investigate the effect of milling time on the grain size and lattice strain of the mechanically milled Al2124 matrix alloy and the effect of sintering time and temperature on the grain size of the matrix in the spark plasma sintered CNT reinforced Al2124 nanocomposite.Also, the density and hardness of the developed materials will be evaluated as functions of the sintering parameters.

Materials and experimental procedures
Al2124 prealloyed powder, supplied by the Aluminium Powder Co. Ltd., and carbon nanotubes, supplied by Nanostructured & Amorphous Materials Inc. USA, were used in this investigation.The prealloyed powder was first wet milled using a planetary ball mill (Fritsch Pulverisette equipment P5, Germany) for deferent milling times to reduce the crystallite size of the aluminum matrix.The powder was charged into cylindrical stainless steel vials (250 ml in volume) together with stainless steels balls (10 mm in diameter).A ball-to-powder weight ratio of 10:1 was used.The wet milling experiments were carried out at room temperature for different milling times at a speed of 200 rpm in the argon atmosphere to prevent oxidation of the powders.The nanostructured powder milled for 5 hours was used to prepare a homogenous Al2124-based composite powder containing 1 wt.%CNTs, the suitable amount of CNTs, i.e. 1 wt.%, was obtained in previous studies [27,32].The mixture was first sonicated in a probe sonicator for 30 minutes.The sonicated mixture was charged into cylindrical stainless steel vials (250 ml in volume) together with stainless steels balls (10 mm in diameter) and milled for 1 hour.A ball-to-powder weight ratio of 10:1 was used.The dried powder was further dry-milled for 10 minutes to break down the agglomerates.For SPS, the prepared powders were directly charged into a graphite die through which the current was passed.A 20 mm graphite die was used and a pressure of 35 MPa was applied during the whole sintering process, i.e. during heating and holding time.The heating rate was 100 K/min.The samples were vacuum sintered at 400, 450, and 500°C for 5, 10, 15, and 20 minutes using a fully automated spark plasma sintering equipment (FCT system, Germany), model HP D 5. The sintering temperature was measured using a thermocouple inserted in the graphite die through a drilled hole.A graphite sheet was used to minimize friction between the die walls and the powder as well as to ease the ejection of the sample after the sintering has been completed.A field Emission Scanning Electron Microscope (Tescan Lyra-3, Czech Republic) was used for characterization of the milled and sintered samples.A high resolution X-ray diffractometer (Bruker D8, USA) was used for recording X-ray diffraction (XRD) patterns of the milled powders and sintered samples.The grain size and lattice strain evolution of the matrix phase during milling and sintering was evaluated using the following equation [45] BB r cosθ = k λ / L + ηsinθ where B r is the full width at half-maximum (FWHM) of the diffraction peak after correction for the instrument broadening, k is constant (with a value of 0.9); λ is wavelength of the X-ray radiation (λ = 0.15405 nm); L and η are the grain size and internal strain, respectively; and θ is the Bragg angle.
The B r is related to the measured width of the peak (B obs ) and peak broadening caused by factors except the particle size effect (B i ), frequently called instrumental broadening factor and calculated using a fully annealed sample at 400°C for 2 hours, through the following formula The density of the sintered samples was measured using an Alfa Mirage electronic densimeter, model MD-300s (accuracy of 0.001 g/cm 3 ) and quantified according to Archimedes principle.Vickers microhardness of the spark plasma sintered samples was measured using a digital microhardness tester (Buehler, USA).All measurements were conducted by applying the same conditions: a load of 100 gf, a time of 12 s.The data reported were the average of 10 values.

Results and discussion
The morphology of particles of the as-received Al2124 powder is shown in fig.1(a).As can be seen, the particles have deferent shapes ranging from almost spherical to irregular, more information on the particle size distribution and the structure and composition of the Al2124 alloy was reported elsewhere [28].Ball milling of the powder reduced the particle size; and particles were fractured to smaller fragments and became flattened with an increase in milling time.This is due to the fact that ball milling of ductile materials, such as metals involves fracturing, cold welding, and rewelding of the particles.The formation of welded flakes was more pronounced at a relatively short milling time of 10 hours; these flakes were found to break into smaller debris after a relatively long milling time (20 hours).Fig. 1(b) shows typical particle' morphology of the powder milled for 20 hours.
The ball milling not only reduced the particle size but also decreased the grain size of the α-aluminum phase and increased its lattice strain as presented in fig.s 2(a) and 2(b), respectively.It can be seen that ball milling of the powder for 20 hours decreased the grain size of the α-aluminum phase from around 250 to 50 nm, which led to the formation of a nanostructured matrix.This reduction in the grain size is comparable to that obtained by Kubota et al. [20], who found that the grain size of the pure Al powder decreased to 38 nm for after 8 hours of dry milling using steel balls and vials at a milling ratio of 7:1.A slightly larger value for grain size was reported by Khan et al. [21], who found it to be 75 nm after 20 hours of dry milling using steel balls and vials at a milling ratio of 16:1 and milling intensity of 120 rpm for pure Al powder.It is evident from fig. 2(a) that the major reduction in the grain size of the αaluminum phase occurred during the first 6 hours of milling after which the decrease was not significant.This behavior is possible in mechanical milling because after a certain milling time, the system will reach an equilibrium state where the rate of grain refinement and particle fracturing becomes equal to the rate of grain growth and recovery.Also, the smaller grains are already saturated with defects and dislocation pile-ups and, hence, the lattice structure cannot continue to develop the same way as is known for coarse-grained metals [21].It is generally understood that once the nanocrystalline structure is fully developed, further reduction in grain size is nearly impossible due to the excessive stress required for deforming the nano-sized grains.The creation and movement of dislocations under these circumstances is difficult and hence grain boundary sliding becomes the dominant deformation mechanism.The rate of recovery is another parameter, which limits the grain size reduction during mechanical milling.Partial recovery may occur during mechanical milling, especially for metals with a low melting point.For such metals, the formation of high-dislocation density regions is hindered by the increased recovery [18].
The analysis of the evolution of the α-aluminium grain size as function of milling time showed that a milling time of 6 hours is the optimum time to reach a nanostructured matrix.Therefore, Al2124 and Al2124 + 1 wt.%CNTs samples milled for 6 hours were spark plasma sintered at different sintering temperatures for different sintering times to investigate the densification and evaluate the hardness of samples.The density of the sintered samples as a function of sintering temperature and sintering time is presented in fig.s 3. The relative density of the Al2124 alloy sintered at 400°C increased from 93.50 to 95.14 % with the increase in sintering time from 5 to 20 minutes.Also, it increased from 96.24 to 97.43 % and from 99.3 to 99.77 % when the alloy was sintered at 450 and 500°C, respectively, for similar sintering times.A similar trend was observed for the Al2124+1wt.%CNTs nanocomposite.Its relative density at 400°C increased from 94.44 to 94.96 % with the increase in sintering time from 5 to 20 minutes.Also, it increased from 96.79 to 97.94 % and from 98.81 to 99.98 %, as a result of increasing sintering time from 5 to 20 minutes, at 450 and 500°C respectively.
Overall, the relative density of the sintered samples increased with increasing sintering temperature and sintering time to reach maximum values at a sintering temperature of 500°C and a sintering time of 20 minutes.This is due to the fact that sintering is a thermally activated process controlled mainly by diffusion [46,47].Therefore, the higher the sintering temperature and time, the higher the diffusion rate and the lower the remaining porosity.It can be concluded that full densification was achieved using the spark plasma sintering process.The evolution of grain size of the α-aluminium matrix in the sintered samples as a function of the sintering temperature and sintering time is presented in fig. 4. For he Al2124 alloy sintered at 400°C, the increase in sintering time from 5 to 20 minutes led to the increase in the grain size of the α-aluminium matrix from 79.2 to 97.2 nm.The further increase in the sintering temperature to 450 and 500°C resulted in a further increase in the grain size; where it increased from 110.3 to 136.6 nm and from 100.2 to 150.6 nm with the increase in the sintering time from 5 to 20 minutes.The nanocomposite displayed a similar behavior; the grain size of the α-aluminium matrix in the nanocomposite increased from 50.2 to 83.6 nm with the increase in the sintering time from 5 to 20 minutes at 400°C.On the other hand, the increase in the sintering time from 5 to 20 minutes at sintering temperatures to 450 and 500°C led to the increase in the grain size of the α-aluminium matrix from 71.5 to 125.4 nm and from 100.4 to 150.5 nm, respectively.
It is worth mentioning here that the grain size increased with an increase in the sintering temperature and sintering time to reach maximum values at a sintering temperature of 500°C and a sintering time of 20 minutes and although sintering led to grain growth in all samples, the grain size of the α-aluminium matrix remained in the nanometer range and did not exceed 150 nm.This is in agreement with the fact that nanostructured materials often exhibit a remarkable resistance to grain growth [48].
Isothermal grain growth can be described by the equation: Where G 0 and G are the grain sizes at initial time t 0 and isothermal holding time t, respectively.K is the material's constant that depends on the temperature: where Q is the activation energy for grain growth.Fig. 5 shows the hardness of the sintered samples as a function of sintering temperature and sintering time.For the Al2124 alloy sintered at 400°C, the increase in the sintering time from 5 to 20 minutes led to the increase in hardness from 99.31 to 100.37.The further increase in the sintering temperature to 450 and 500°C resulted in a further increase in hardness: it increased from 104.25 to 106.63 and from 105.87 to 110.24 with the increase in the sintering time from 5 to 20 minutes.The nanocomposite displayed a similar behavior: its hardness increased from 81.31 to 95.51 with the increase in the sintering time from 5 to 20 minutes at 400°C.On the other hand, the increase in the sintering time from 5 to 20 minutes at sintering temperatures 450 and 500°C led to the increase in hardness from 104.85 to 107.6 and from 109.48 to 118.19, respectively.
As can be clearly seen from fig. 5, the hardness of the sintered samples increased with increasing sintering temperature and sintering time to reach maximum values at a sintering temperature of 500°C and a sintering time of 20 minutes.
The increase in microhardness with an increase in the sintering temperature is similar to an increase in the density because of elimination of pores during sintering.It is worth mentioning here that besides pores elimination, precipitation of very fine particles such as Mg 2 Si, CuAl 2 and CuMgAl 2 could take place and contribute to the improvement of microhardness because the elements present in the α-aluminium solid solution in excess of the solubility limit may form precipitates during sintering [28].
Generally, the grain size d dependence of the yield stress σ ys is described by a general expression (Hall-Petch relationship) where σ 0 is the lattice friction stress, k is a Hall-Petch slope.Vickers hardness of a polycrystalline material can be related to its yield strength through a simple relationship H v /σ ys ≈ 3. Therefore, the hardness H v can be related to the grain size by where H 0 and k are constants.
It is clear from the above relationship that the increase in the grain size reduces the hardness of a material.However, an increase in hardness of the Al2124 alloy and the Al2124 + 1wt.%CNTs nanocomposite despite an increase in the grain size is due to the fact that during sintering pores are eliminated and the density of the material increases.Therefore, the hardness of the material strongly depends on its density and the effect of grain growth will be small.This is more meaningful, specifically with a process such as spark plasma sintering where the heating rate is high, the sintering temperature is low, and the sintering time is short compared to other conventional sintering processes [22,[49][50][51].The hardness reported in this study for the Al2124 alloy was higher than the hardness of other alloys, processed under similar conditions, such as Al6061 [27], Al-7Si-0.3Mg[52], and Al-12Si-0.3Mg[31].Also, the hardness of the Al2124+1wt.%CNTsnanocomposites was higher than the hardness of other nanocomposites processed under similar conditions and reinforced either with CNTs or SiC.These include Al6061+1wt.%CNTs [27], Al-7Si-0.3Mg+0.5CNTs[52], Al-12Si-0.3Mg+0.5CNTs[52], and Al-7Si-0.3Mg+SiC[31].

Conclusion
The fully densified Al2124 nanostructured alloy and Al2124+1wt.%CNTnanocomposites were produced from nanostructured powders prepared through mechanical milling and through spark plasma sintering process.Mechanical milling of the Al2124 powder reduced the particle size and decreased the grain size of the α-aluminum phase to 50 nm and increased its lattice strain.A milling time of 6 hours was found to be the optimum time to reach an α-aluminum nanostructured matrix.The grain size of the α-aluminum phase in sintered samples increased with increasing sintering temperature and time to reach maximum values at a sintering temperature of 500°C and a sintering time of 20 minutes.Although sintering led to grain growth, the grain size of the α-aluminium matrix remained in the nanometer range and did not exceed 150 nm.The relative density and hardness of the sintered samples increased with increasing sintering temperature and time to reach maximum values at a sintering temperature of 500°C and a sintering time of 20 minutes.

Fig. 4 .
Fig. 4. Grain size of the α-aluminium matrix in the sintered samples.

Table 1 summarizes
Vickers hardness values of selected spark plasma sintered aluminum based nanocomposites.Vickers Hardness of selected spark plasma sintered Al-based nanocomposites.