Mechanical and Thermal Properties of Aluminum Matrix Composites Reinforced by In Situ Al 2 O 3 Nanoparticles Fabricated via Direct Chemical Reaction in Molten Salts

: The development of novel methods for industrial production of metal-matrix composites with improved properties is extremely important. An aluminum matrix reinforced by “in situ” α -Al 2 O 3 nanoparticles was fabricated via direct chemical reaction between molten aluminum and rutile TiO 2 nanopowder under the layer of molten salts at 700–800 ◦ C in air atmosphere. Morphology, size, and distribution of the in situ particles, as well as the microstructure and mechanical properties of the composites were investigated by XRD, SEM, Raman spectra, and hardness and tensile tests. Synthesized aluminum–alumina composites with Al 2 O 3 concentration up to 19 wt.% had a characteristic metallic luster, their surfaces were smooth without any cracks and porosity. The obtained results indicate that the “in situ” particles were mainly cube-shaped on the nanometer scale and uniform matrix distribution. The concentration of Al 2 O 3 nanoparticles depended on the exposure time and initial precursor concentration, rather than on the synthesis temperature. The inﬂuence of the structure of the studied materials on their ultimate strength, yield strength, and plasticity under static loads was established. It is shown that under static uniaxial tension, the cast aluminum composites containing aluminum oxide nanoparticles demonstrated signiﬁcantly increased tensile strength, yield strength, and ductility. The microhardness and tensile strength of the composite material were by 20–30% higher than those of the metallic aluminum. The related elongation increased three times after the addition of nano-α Al 2 O 3 into the aluminum matrix. Composite materials of the Al-Al 2 O 3 system could be easily rolled into thin and ductile foils and wires. They could be re-melted for the repeated application.


Introduction
Metal-matrix composites based on aluminum and its alloy have been the subject of great scientific and practical interest over the last decades [1,2].The reason is the possibility to obtain lightweight metallic materials with unique properties (enhanced hardness, ductility, elastics modulus, and many others) by combining the advantages of constructional metals with the merits of filler materials.Metal-matrix composites based on aluminum and aluminum-metal-matrix composite alloys (AMMC) are promising for aerospace, military, automobile, and electronic areas because of their low density, high specific stiffness, strong wear resistance, reduced thermal expansion coefficient and high thermal conductivity [1,2].These materials sustain their stability characteristics over a wide temperature range, have high electrical and thermal conductivities, and low sensitivity to surface defects [3][4][5].
AMMCs are fabricated both by ex situ synthesis (e.g., liquid ingot casting and powder metallurgy), where reinforcement powder particles are added to the metal melt, and by in situ synthesis (e.g., exothermic dispersion, reactive hot pressing, reactive infiltration, and direct melt reaction), where particles are directly synthesized in the metal melt [6][7][8][9][10][11][12].The reinforcement in AMMCs can have a form of continuous/discontinuous fibers, whiskers, or particulates.The most frequently applied fillers are ceramic ones such as alumina, silicon carbides or nitride, carbon fiber materials, or graphene [13,14].
At present, researchers focus on the possibility of introducing the above-mentioned ceramic additives in the form of nanoparticles into the aluminum matrix.When moving to nanoscale hardeners, we have the opportunity to obtain an entirely new class of lightweight materials with increased specific strength, plasticity, and crack resistance.The mechanisms of deformation and fracture in aluminum-alumina composite materials reinforced by nanosized alumina particles can be fundamentally different from those in classical aluminum alloys.
There are some problems arising in the synthesis of Al-Al 2 O 3 composites: inhomogeneity of the distribution of oxide nanocrystals in the volume of the metal matrix; the tendency of nano-objects to agglomeration inside the matrix (1-3 µm); low wettability of oxide particles by molten aluminum (lattice consistency of the reinforcing phase and aluminum), and degradation of oxide materials (chemical, mechanical, etc.).
The mechanical addition of alumina particles into the metal melt by means of powder metallurgy [15][16][17] results in the interface pollution, poor wetting ability, and mechanical properties.The composites synthesized via the ex situ methods have high porosity and a lot of cracks.They are characterized by the appearance of "soft" and "hard" phases [17] because of the difference in aluminum particles' concentration and agglomeration [15].Hence, the mechanical properties of composites produced by powder metallurgy methods can vary greatly.
Selective laser melting can be used for the creation of advanced ball-milled Al-Al 2 O 3 nanocomposites with a two-fold increase in Vickers micro-hardness compared to the initial Al [18].The mechanical properties of Al-nano Al 2 O 3 produced by ball-milling and hot-pressing under uniaxial pressure of 50 MPa were improved due to the utilization of nano-sized reinforcement particles [19].To date, the compaction of several layers of aluminum foil and aluminum oxide powder, followed by rolling from 2 to 20 times, is another promising method of composite materials' fabrication [20][21][22].As a result, layered laminated Al-Al 2 O 3 composite structures are formed.The uneven distribution of aluminum oxide particles in the aluminum matrix, as well as the significant agglomeration of oxide particles [21], are the main problems of the laminated composite production that cause the uneven properties of the composite.Such defects are usually eliminated by increasing the number of rolling cycles.In addition, the non-isomorphism of the composites' properties is also a problem.
Sometimes, both powder metallurgy and rolling are combined for the synthesis of composites, which makes it possible to obtain Al-Al 2 O 3 composites with a better distribution of dispersing particles [23].
Direct reaction synthesis (DRS) refers to the process in which reactant powders or compacts of reactant powders are directly added to molten metal, and ceramic reinforcing particles are formed in situ through the exothermic reaction between reactants or between the reactant and the component of the melt.During the in situ process, the reinforcement phases formed in the matrix metal by direct chemical reactions are thermodynamically stable, free of surface contamination, and disperse more uniformly.These factors lead to the stronger particle-matrix bonding at the "in-situ" process.
DRS in situ synthesis can be applied as a commercial method because of its simplicity, low cost, and near-net-shape-forming capability.The biggest problem of the traditional DRS method is the necessity of a long mixing time at a high temperature, which causes the agglomeration of small initial particles [24,25].The introduction of alumina to the molten aluminum matrix by the "in situ" process was proved by different independent physical-chemical methods such as XRD, Raman spectroscopy, and others [26].
Nowadays, synthesizing Al-Al 2 O 3 nanocomposites is an important practical and fundamental task in the field of inorganic chemistry.There is a very high need to create a high-tech method for the synthesis of non-porous aluminum-alumina nanocomposites with improved mechanical properties.
The novelty of the present work is that the aluminum matrix was reinforced by "in situ" perfectly wetted and well-distributed Al 2 O 3 nanoparticles, formed via a direct chemical reaction between molten aluminum and rutile TiO 2 nanopowder under a molten salt layer at 700-750 • C in air atmosphere.The microstructure and mechanical properties of the composites were investigated by XRD, Raman spectroscopy, SEM, and tensile testing.
This research aims to develop a method for the industrial large-scale "in situ" synthesis of aluminum-based matrix composites, reinforced with ceramic nanoparticle inclusions.These composites are formed via the reaction between molten aluminum and the salt electrolyte flux, which contains rutile TiO 2 nanopowder as the precursor.The effect of nanoalumina inclusions content on mechanical and thermal properties of the Al-nano-Al 2 O 3 composite is analyzed.

Experimental
Titanium dioxide nanopowders were produced by electrochemical oxidation of titanium foil of 99.9 wt.% purity ("VSMPO-Avisma", Ekaterinburg, Russia) in the chloridenitrate melt according to the method described in detail elsewhere [27,28].The interaction between molten aluminum and nano-and micro-oxide powders of titanium dioxide was performed in molten alkali halide media in an alumina crucible containing the mixture of alkali chlorides with a weighed portion of oxide additive.Salt mixtures of alkali chlorides with melting points lower than 700 • C, with small additions of some fluorides, were used as a base electrolyte with the concentration of titanium dioxide ranging from 0.2 to 2 wt.%.We used commercial "chemically pure" salts provided by "Lanhit" Ltd., Moscow, Russia.
The aluminum sample A95 of 99.5 wt.% purity ("Rusal", Russia) consisted of a 1 cm high disc with a diameter strictly equal to the crucible diameter.The disc was loaded into a layer of the dry oxide-salt mixture at the bottom of the crucible.The primary admixture of technical aluminum was iron-about 0.5 wt.%.The aluminum disc was covered by a layer of dry alkali chlorides.The alumina crucible was exposed inside the vertical heating furnace at the temperatures range of 700-800 • C for 3 to 5 h.After the exposure, the liquid aluminum globule was poured into the new cold alumina crucible.After the solidification, the salt mixture was dissolved in distilled water.The aluminum composite ingot was separated and dried.Cooled to the ambient temperature, aluminum-alumina composite ingots were divided into four parts.One part was used for the determination of Al 2 O 3 content in the composite, the second part was used for XRD analysis, the third one was poured with current-carrying gum, polished by six different silicon solutions using a Struers disc-finishing machine (Copenhagen, Denmark), and examined by means of scanning electron microscopy (SEM, Raman spectroscopy (Wotton-under-Edge, England, UK), microhardness testing).The fourth biggest part was remelted to form cylindrical samples for a servo-hydraulic testing machine.
The processing of the studied composites is schematically shown in Figure 1.
The structures of titanium dioxide and aluminum-alumina composite were studied by means of SEM, XRD, and Raman spectroscopy.In order to determine the material composition and the composite structure, the purity of the oxide phase was analyzed using Raman spectra by a Renishaw U 1000 micro-Raman spectrometer (UK) connected to a Leica DML microscope equipped with 50× and 100× magnification lenses.An Ar + laser (Cobalt model) with a wavelength of 532 nm and power output of 20 mW was used as the excitation source.The diameter of the laser spot was about 1 µm; the acquisition time was set to 20 s.The XRD analysis was performed on a «RIGAKU» DNAX 2200PC diffractometer at ambient temperature with Cu Kα radiation.Subsequent analysis of the vibrational spectra was carried out using the embedded Wire 3.0 software (Wotton-under-Edge, England, UK); the spectral lines were approximated using the Fityk 0.9.8 curve fitting and data analysis application.The typical average spectra obtained are presented for each sample.The volumetric method determined the quantitative content of aluminum oxide in the produced AMMC.Volumetric analysis was carried out by dissolving a composite sample in 30% hydrochloric acid, followed by recalculating the released hydrogen per mass of pure aluminum in the sample according to reaction (1): The sample was then examined as follows: scanning electron microscopy (SEM) was performed using JEOL 5900LV and TESCAN MIRA3 microscopes; the sample surface was characterized by a Veeko Wyko NT 1100 optic profilometer-profilograph in the Vertical Scanning Interferometry (VSI) regime; the fatigue strength was examined by a servohydraulic testing machine INSTRON 8801, the sample microhardness was tested a WIN-HCU FISCHERSCOPE HM 2000 XYm system, which is designed for Martens hardness testing with the WIN-HCU software at a maximum load P from 1000 mN to 2000 mN, a loading time of 20 s, a time of exposure under loading of 15 s, and an unloading time of 20 s according to the ISO 14577 standard [29].
The characteristics measured by indentation were used for the calculation of the following parameters: the ratio of the indentation hardness to the contact elasticity modulus HIT/E* [30], the elastic recovery R e = h max −h p h max × 100% [31], the power ratio [32], and the plasticity index [29] that characterizes the ability of a material to resist elastoplastic strain.Ten measurements of indentation characteristics at each load were performed.The error of measurements was determined with a confidence probability p = 0.99.

Formation of Nano-Titanium Dioxide as a Precursor
The formation of titanium nano-oxide powders as precursors of composite synthesis was carried out by electrochemical oxidation of pure titanium in a chloride-nitrate melt in an argon atmosphere [27].The main function of the nitrate melt is corrosion protection against the aggressive environment due to the formation of dense protective oxide layers on metals.Chloride-nitrate melts allow synthesizing both dense protective coatings and nanocrystalline powders of metal oxides such as Al 2 O 3 , TiO 2 , ZrO 2 , and PbO [33,34] varying the nitrate concentration in the flux.Due to the alkali nitrates' thermal instability, the temperature range of their application is narrow, while amorphous oxides are usually formed at low temperatures.It is necessary to synthesize oxides using salt electrolytes with higher melting points to increase the temperature of the crystalline metal oxide synthesis.
The metal oxidation in the chloride-nitrate salt melt bulk at the current densities ranging from 3 to 5 mA/cm 2 lead to the accumulation of significant amounts of oxide nanopowder in the bulk salt electrolyte.The chemical composition and modification of the formed oxides depended mainly on the oxidation temperature.
Despite the nanometer size of titanium dioxide crystallites (Figure 2a), the X-ray diffraction pattern contained main rutile titanium dioxide lines of 2θ angle at 27.5  101), ( 200), ( 111), ( 210), ( 211), ( 220), (002), and (301), respectively (JCPDS no.#88-1175) [35] (Figure 2c).The broadening of lines or a halo was not observed, which may be explained by the fact that the formed oxide particles were well-faceted cubic crystals, and not spheroids, as is often the case, which happens in the synthesis of nano-oxides by the sol-gel method.Titanium dioxide nanopowders synthesized in a molten chloride-nitrate electrolyte at the current density of 3.5 mA/cm 2 were characterized by a high degree of crystallinity and took the form of titanium dioxide cubes with a size up to 30 nm (Figure 2a).According to the results of the Raman spectroscopy of the sample obtained at 700 • C, peaks characteristic for the modification of rutile TiO 2 were determined in the Raman bands: 144 cm −1 , 448 cm −1 , 612 cm −1 (Figure 2b).Further, rutile titanium dioxide nanopowder formed under the above conditions was used as a precursor.
Various industrially sol-gel methods for nano-and micro-powders production in the form of rutile and anatase were previously studied as precursors for the synthesis of the nanocomposite.However, all these nanopowders were sedimented to the bottom of the crucible and reaction (3) (Table 1) did not proceed.The titanium dioxide nanopowder in the rutile modification synthesized by the electrochemical oxidation of titanium in a chloride-nitrate melt was the only precursor with which the above reaction was successful.The AMMC material dispersion reinforcement is typically achieved by casting or powder metallurgy, with the ceramic particles being introduced ex situ into a solid or liquid matrix.A new AMMC production method has been developed on the basis of the controlled "in-situ" chemical reactions; reinforcing fillers are formed during the chemical interaction between the matrix components and reactive additives [9,12].Such AMMCs demonstrate excellent mechanical and physical properties, because coherent (i.e., having a standard atomic layer at the interface) or partially coherent interfaces are formed between the matrix and new phases that arise in "in-situ" reactions.
The uniform distribution of fine particles of alumina in the aluminum matrix was achieved by the modified method of molten salt synthesis.Molten halides (alkali or alkali-earth chlorides and/or fluorides) are the optimal reactive media for the formation of aluminum-alumina composites, as was shown in our previous papers [26,36].
The synthesis of alumina particles inside aluminum is carried out by a one-step chemical reaction of titanium dioxide with molten aluminum under the layer of molten halides in air atmosphere by reaction (3); therefore, newly formed alumina nanoparticles are perfectly wetted by the liquid metal.
Obviously, these reactions, except for the direct interaction between TiO 2 and molten sodium chloride, are thermodynamically feasible at sufficiently low temperatures of 700-800 • C. The presented process is a special case of aluminothermy.
The interaction between liquid aluminum and fine titanium dioxide particles by reactions ( 5) and ( 6) can proceed up to the formation of the intermetallic AlTi and Al 3 Ti titanium aluminide compounds.The Gibbs energies of reactions ( 5) and ( 6) are more negative than that of reaction (3).It makes reactions ( 5) and (6) more favorable.Titanium aluminides were not detected in the reaction product by X-ray phase analysis, Raman spectroscopy, EDS spectra of composites, and mapping sample surfaces in X-rays.Intermetallic compounds of aluminum and titanium or titanium species were not detected by means of an inductively coupled plasma optical emission spectrometer OPTIMA 4300 DV (Perkin Elmer, USA) after dissolving the resulting composite in hydrochloric acid.Thus, it can be argued that the interaction occurred according to the mechanism described above by reaction (3).
The reason for such a result is the appearance of a side reaction (4) in the system.The interaction between titanium atoms, sodium chloride, and gaseous oxygen according to reaction (4) resulted in the formation of Na 2 Ti 6 O 13 nanowires and nanorods, which was experimentally proved [37].The value of Gibbs energy of this reaction is 15 times greater that of all other reactions.
The composites obtained in this way were non-porous with a characteristic metallic luster (Figure 3).They exhibited high electrical conductivity and are excellently suited for mechanical processing.It was experimentally proved that the aluminum oxide content did not change from the surface of the aluminum drop to its center.The density of the obtained composite averaged 2.6977 g/cm 3 , which is very close to that of pure aluminum (2.6989 g/cm 3 ), and indicates the absence of porosity in the composite, which is typical for such samples obtained by traditional methods.According to the X-ray diffraction data, the thus obtained material was composed of Al and α-Al 2 O 3 [36].
Raman spectra of the oxide inclusions were collected from several crystals inside the aluminum matrix.The average spectrum of the oxide inclusion is presented in Figure 4.It is obvious that this spectrum was a standard spectrum of α-Al 2 O 3 with the characteristic bonds at 378 cm −1 , 417 cm −1 , 430 cm −1 , 448 cm −1 , 576 cm −1 , 644 cm −1 , and 749.5 cm −1 [38].It confirmed the XRD data about the synthesis of α-Al 2 O 3 inside the aluminum matrix [36].Micro-and nanosized inclusions of aluminum oxide were also clearly visible in the micrographs in the optic microscope (Figure 5a) and in backscattered electrons in a scanning electron microscope (Figure 5b,c).The EDS spectrum of inclusion showed the content of aluminum and oxygen in the stoichiometric ratio of Al 2 O 3 ; inclusions of aluminum oxide had sizes close to 100 nm.A very uniform distribution of oxide inclusions inside aluminum grains of the crosssection of Al-Al 2 O 3 composite is presented in Figure 6.
Inclusions in the form of long rods were crystals of iron aluminide Al 3 Fe (Figure 6), which were present in the starting aluminum as an admixture.According to the phase measurements of the cross-section of this sample, the Al area is 97.5%, Al 3 Fe-1.4%,and Al 2 O 3 -1.2%.The 3D image of the cross-section of AMMC (Figure 5d) also shows a uniform distribution of alumina nanoparticles inside the grain of the aluminum matrix.
To prove the possibility of finding nanoparticles mainly inside the grain rather than at the grain boundary, as would be expected from classical considerations, we calculated the value of the critical nucleus (r c ) during the crystallization of aluminum using well-known Formula (7) [39]: where σ is the surface tension of the metal melt; T 12 is the melting point of the metal; M is the molecular weight; Q 12 is the heat of crystallization (heat of fusion); ∆T is the supercooling of the melt, at which the nucleation of the crystallization center occurred; ρ is the density of the metal.As a result, the value of the radius of the critical nucleus of Al 2 O 3 was r c = 108.33nm.Thus, the size of the critical nucleus was three times larger than the initial size of the nanoparticles (30 nm), and the calculated value of synthesized αalumina in the aluminum matrix was very close to the experimental one, which varied from 100 to 150 nm.
Thus, it can be assumed that nano-α aluminum oxide formed as a result of the chemical interaction between fused aluminum and nanopowder of titanium dioxide and was not grouped along the grain boundaries but was uniformly distributed in the volume of the aluminum grain (Figures 5 and 6), which is a favorable factor in assessing the mechanical properties of the composite.
The composite samples were dissolved in hydrochloric acid HCl solution to determine the quantitative content of aluminum oxide in the produced AMMCs using the volumetric method.The concentration of Al 2 O 3 in the composite was determined by the difference in hydrogen evolution.
The data in Table 2 elucidates that an increase in the exposure time and concentration of titanium dioxide in the melt leads to an increase in the mass fraction of aluminum oxide in the composite.However, when the titanium dioxide concentration increased significantly above 1% and exposure time was up to 9 h, it was not possible to obtain samples with a reinforcing phase concentration of more than 19%.A slight decrease in the mass fraction of aluminum oxide in the composite caused by an increase in temperature to 750 • C could be associated with the interaction between atomic titanium and a salt flux, which resulted in the formation of sodium titanate.As can be seen from Table 2, we varied the concentration of aluminum oxide of AMMC not by introducing a certain amount of aluminum oxide into the aluminum matrix, but by changing the synthesis time and temperature.The content of aluminum oxide formed by the reaction (3) did not depend on the titanium dioxide weight.Titanium dioxide was a necessary seed for the formation of aluminum oxide in the composite; when interacting in a salt melt, it formed sodium hexatitanate nanofibers in stoichiometric amounts [37].

Thermal Analysis of Composite Al-Nano-Al 2 O 3
Thermal analysis of Al-Al 2 O 3 composites was measured by a thermal analyzer STA 449C Jupiter, NETZSCH at a heating rate of 10 K/min, the temperature determination error was <1.5 K. Synchronous analysis of samples of aluminum-nano-alumina composites was carried out by differential scanning calorimetry (DSC) and thermogravimetry as a function of temperature in air atmosphere.In the first case, the sample was heated to temperatures up to 723 • C, cooled and, without removing the sample from the crucible, heated repeatedly to the same temperatures, and then the sample was heated to the temperature of 800 • C. The samples of the initial aluminum and studied composites exhibited congruent melting, since the melting peaks were reproduced during repeated cooling and heating cycles, while the lines were expressed in one DSC anomaly with the same areas during each repeated cooling/heating cycle.The onset points changed no more than 0.3 • C, hence the maximum temperatures, indicating the termination of the melting process of the samples, differed slightly and fluctuated in the range of 669.9-671.4• C (Table 3).This phenomenon illustrates the property of peak reproducibility (Figure 7), which is a reflection of the reversibility of the phase transition during cooling and subsequent heating.The closeness of the value for the obtained composites to the parameters of the original metal can be primarily due to the indicators of the wettability of the dispersed oxide nanophase by aluminum.The data obtained by the thermogravimetry simultaneously with the differential scanning calorimetry also revealed that the thermal stability of the obtained composite materials was close to pure aluminum.The recorded change in the mass of the aluminum was equal to 0.74% and that in the composite samples during three heating and cooling cycles did not exceed 0.37%.Such an insignificant change in the mass of the sample during measurements, accompanied by the appearance of lines attributed to endothermic anomalies during DSC, directly indicated the presence of the process, which is a first-order phase transition with a high reproducibility index.

Measurement of Mechanical Properties of Composite Al-Nano-Al 2 O 3
Experimental methods to study mechanics of deformable "Al-nano-α-alumina" composites included hardness studies and tensile testing of the material with data registration in the form of "stress-strain" curves.For the in-depth mechanical studies, samples with an aluminum oxide concentration of 10 and 14 wt.% were selected.
The instrumented indentation data on the specific features of Al and Al-Al 2 O 3 behavior under elastoplastic strain are given in Table 4.It is evident that the "in situ" implementation of uniformly distributed nano-α Al 2 O 3 (10 and 14 wt.%) into the aluminum matrix decreased in the maximum and residual indentation depth h max and h p and increased in the Martens hardness H M and the indentation hardness at all loads HIT.Such simultaneous change in the characteristics h max , h p , H M , H IT , W e , and W t measured in the indentation process is commonly caused by the material hardening and mainly is observed in aluminum composites with uniformly distributed alumina nanopowders [40][41][42].In particular, for the samples "Al-nano-alumina ", an average increase in H IT was nearly 20-25%.Contact elasticity modulus E * decreased for Al-14% Al 2 O 3 composite at the load 1000 mN and did not change for all other composites.
The AMMC was characterized by higher resistance to elastoplastic strain as evidenced by an increase in the calculated parameters R e by factor 1.22-1.52,H IT/ E* by factor 1.2-1.5, and H 3  IT /E* 2 by the factor 2.1-2.7 (Table 5).The plasticity index δ A did not change at the increased hardness and remained relatively high in the range of 0.972-0.982.Mechanical tensile tests of aluminum-aluminum oxide composite samples were carried out in order to determine the strength and plastic properties.The tests were carried out in accordance with RF GOST 1497-84 on IV type samples (Figure 8).Typical room temperature stress-strain curves of the aluminum-alumina composites reinforced with varying alumina content are shown in Figure 9.It is evident that the yield strength did not increase with the increase of nano-αalumina content up to 14 wt.% in the composite (Figure 10).This is in opposition to conclusions drawn in [43] that the increase in tensile strength was observed only up to α-Al 2 O 3 concentrations of about 5 wt.% and the tensile strength decreased at other concentrations.The composites with an addition of 7-14 wt.% of α-Al 2 O 3 possessed tensile strengths of 67.56 and 78.16 MPa, which were 1.10 and 1.27 times greater than that of the aluminum sample (61.38 MPa).As can be seen from Figure 8, a further increase in the content of aluminum oxide in the composite to 19 wt.% neither improved, nor worsened, the mechanical properties of the composite.Thus, it can be considered that the aluminum oxide concentration of 14 wt.% was optimal from the point of view of the combination of all properties.A great simultaneous increase in the sample elongation accompanied this strong improvement of tensile strength.At the same time, significant hardening was accompanied by the elongation of the samples with the addition of 7-14 wt.% of α-Al 2 O 3 , which increased from 1.64 to 3.06 times with the greater addition of alumina and reached values of 21.32% and 39.77%, compared with 13% for pure aluminum.This means that the hardness, elasticity, and ductility of the aluminum-nano-α-Al 2 O 3 composite improved simultaneously according to the oxide concentration with uniformly distributed α-Al 2 O 3 nanoparticles inside the aluminum grains.Commonly, improving the hardness of materials leads to brittle deformation, however, the addition of up to 14 wt.% of nano-α-alumina into the aluminum matrix allows increasing simultaneously the strength, hardness, ductility, and elasticity of the resultant aluminum-nano-α-Al 2 O 3 composite.
Figure 11 shows SEM images of the tensile fracture surfaces of Al-Al 2 O 3 nanocomposites.The fracture mode of the initial matrix Al is predominantly not ductile, with several dimples all over the surface.As shown in Figure 11b When the Al 2 O 3 content in the composite was increased, the number and depth of the dimples decreases significantly, and the dimples' size increased at least three times, leading to a very ductile type of failure.This three-fold increase in the length of the cracks was in excellent agreement with the three-times-increased elongation of AMMC under room temperature stress-strain tests.As seen in Figure 11b, nanoparticles of α-alumina did not agglomerate in the aluminum matrix and were uniformly distributed even after tensile strength tests.Cracks or porosity were not observed because of the very uniform distribution of α-Al 2 O 3 nanoparticles in the aluminum matrix.

Strengthening Mechanism of Al-Al 2 O 3 Nanocomposites
The strengthening mechanism of Al-Al 2 O 3 nanocomposites can be described by the strengthening of fine nanoparticles.Fine Al 2 O 3 particles can act throughout the matrix as barriers to the dislocations and thus the elevated mechanical properties may significantly improve the hardness and ductile properties of the composites compared to the base alloy.The contributions of different strengthening mechanisms were estimated according to well-known formulas [43] (Table 6).
The ultra-hard α-Al 2 O 3 nanoparticles in the Al matrix acted as barriers to the motion of dislocations generated in the matrix.The higher particle density caused the Orowan strengthening, which is the main influence according to Table 6.It is obvious that the experimental values were much closer to the sum of the load transfer (∆σ load ) and Orowan strengthening (∆σ Orowan ).The yield strength of the composites increased mostly because of the Orowan strengthening.However, an increased addition of α-Al 2 O 3 nanoparticles up to 14 wt.% did not lead to the agglomeration of the reinforcement, and we did not observe any subsequent degradation of the hardness values as opposed to Srivastana et al. and Su et al. [1,7].Further increase in the concentration of highly dispersed α-Al

Figure 3 .
Figure 3. Optic photo of the ingot (a) and cross-section (b) of Al-Al 2 O 3 .

Figure 4 .
Figure 4. Raman spectrum of the oxide inclusion inside the Al matrix.

Figure 9 .
Figure 9.Typical stress-strain curves of the aluminum-alumina composites reinforced with varying alumina content.

Figure 10 .
Figure 10.Dependence of tensile and yield strengths of AMMCs on the alumina concentration.
, 14 wt.% nano Al 2 O 3 -reinforced composite also displayed dimples similar to the Al matrix, except that the length and depth scale of the dimples was very different.No cracks on the fracture image of Al-nano-α Al 2 O 3 nanocomposite were observed at tensile strength tests.

Figure 11 .
Figure 11.morphologies of Al (a) and AMMC with 14 wt.% of Al 2 O 3 (b) after room temperature stress-strain tests.
2 O 3 nanoparticles up to 19 wt.% did not result in any hardness improvements as compared to the composite containing 14 wt.% of α-Al 2 O 3 , but we did not observe any deterioration in the hardness.The concentrations above 19 wt.% Al 2 O 3 in the composite were not produced by our method.The strengthening from fine nanoparticles can describe the strengthening mechanism of Al-Al 2 O 3 nanocomposites.Fine Al 2 O 3 particles spread throughout the matrix can act as barriers to the dislocations, and thus mechanical properties of composites (hardness and ductile properties) are improved as compared to the base alloy.The increase in hardness and strength (approximately 1.2-1.3times) of the new composite material compared to the original aluminum is explained by the inclusion of α-Al 2 O 3 particles into the aluminum matrix.The α-Al 2 O 3 particles are in their strongest allotropic corundum modification, which has a hardness close to that of a diamond.An unusual increase in the relative elongation of the composite by a factor of three makes it possible to roll up to 15 µm thick foil on conventional rolling equipment with a deformation of about 500%, as well as to draw 0.2 mm thick wire from Al-Al 2 O 3 nanocomposites (Figure 12).

Table 1 .
Gibbs energies of some possible reactions in the Al-TiO 2 -NaCl-O 2 system.Interaction between Molten Aluminum and Titanium Nanodioxide under the Layer of Molten Halides and Analysis of Interaction Products

Table 2 .
The influence of the concentration of TiO 2 precursor, temperature, and exposure time on the concentration of Al 2 O 3 inside the obtained Al-Al 2 O 3 composite.

Table 3 .
Onset points, melting temperatures, and Gibbs energies of melting of Al-Al 2 O 3 samples.

Table 4 .
Results of microindentation to the surface of Al and Al-Al 2 O 3 samples at different maximum indentor loads P.

Table 5 .
Elastic recovery R e , ratios H IT /E* and H IT 3 /E* 2 , and plasticity δ A at different maximum indentor loads P for the surface of Al and Al-Al 2 O 3 samples.

Table 6 .
Contributions of different strengthening mechanisms to the yield stress of Al-nano-Al 2 O 3 composites.