Figure 2 shows the SEM images of the Al–TiO₂ powder mixture milled from the initial and milling times of up to 5 h. The results showed that the image of the nonmilling powder depicted that the raw aluminum powder has a relatively large particle size of approximately 70 ÷ 140 µm, and the size of titanium oxide powder is relatively smaller, i.e., approximately 250 ÷ 350 nm, as shown in Figure 2A. Before milling, the titanium dioxide particles were unevenly dispersed and adhered to the surface of the aluminum foil. After 1 h of milling, an improved distribution of TiO₂ in the mixture was achieved, and the aluminum foils were broken down to approximately 20–30 µm in size, as shown Figure 2B. The surface morphologies of the Al–TiO₂ powder mixture with increasing grinding time from 2 to 4 h are shown in Figures 2C–E. The results clearly showed that the aluminum particles no longer existed in the foiled form but changed to a rounded shape with fine TiO₂ particles distributed evenly on the surface. Until ball milling, the raw Al–TiO₂ powder for 5 h, the raw materials were well blended, and the TiO₂ particles were almost indistinguishable from aluminum, as shown in Figure 2F. The results indicated that the milling time enhanced the distribution of TiO₂ powder in the host Al matrix. In other words, the mechanical milling process has shown its effectiveness in producing deformations and defects in the raw powders and increasing the temperature in the jar during milling might facilitate the diffusion process. As the milling time increased, the aluminum particles were broken, crushed, and mixed with the titanium dioxide particles. The increase in temperature in the jar during milling also increased.

The phase forms in Al–TiO₂ powder before and after thermal treatment as a function of milting time are shown in Figure 3. The XRD patterns of samples with 5 h of milled powder showed that only Al and TiO₂ phases were obtained in the XRD patterns, indicating that the milling process only promoted the homogeneous distribution of TiO₂ (PDF Card No. 01-075-1537) into the host Al matrix (PDF Card No. 00-001-1176), whereas no reaction to form the new phase occurred. The observation results further confirmed that the milling process only enhanced the mixing between raw Al and TiO₂ powders. In other words, the milling process simply refines the sizes of the aluminum particles and creates an even distribution of the TiO₂ particles in the powder mixture. Evidence of the influence of milling time on the phase form was clearly obtained in the XRD pattern, as shown for various milling times. The main diffraction peaks of the Al₂O₃ (PDF Card No. 00-005-0712) and Al₃Ti (PDF card No. 03-065-4202) phases were obtained in the XRD spectra for 1 h milling, indicating that the reaction between TiO₂ and Al already occurred after thermal treatment at 750°C for 30 min. Our results were consistent with recently reported characterization of the reaction of TiO₂ and Al (Zhang et al., 2011). The observation results showed that the Al₂O₃ and Al₃Ti phases already formed even with a low milling time of 1 h. However, the strong main diffraction peaks of Al and TiO₂ still appeared in the XRD patterns for samples fabricated with milling times from 1 h to 3 h. The main intensity of the Al and TiO₂ peaks tended to decrease as the milling time increased and completely disappeared when the milling time increased over 4 h. Thus, we suggested that a short milling time could not evenly distribute the TiO₂ into the aluminum, resulting in low energy for the full formation of Al₃Ti and Al₂O₃. The observation results indicated that controlling the milling time was one of the key factors for enhancing the phase form during thermal treatment.

The formation of Al-Ti intermetallic in the system was the results of the following reactions (Khoshhal et al., 2010; Zhang et al., 2011; Binh et al., 2021): 13 A l + 3 T i O 2   →   3 A l 3 T i + 2 A l 2 O 3 (3.1) 7 A l + 3 T i O 2   →   3 A l T i + 2 A l 2 O 3 (3.2) 5 A l + 3 T i O 2   →   A l T i 3 + 2 A l 2 O 3 (3.3)

For further information, the Gibbs free energy of the reactions were calculated based on thermodynamic data from literatures which were shown in Figure 4 (Kattner et al., 1992; Gui-rong et al., 2010; Khoshhal et al., 2010). Thermodynamic calculations indicated that the Gibbs free energy of the (3.1) reaction was the lowest in a wide range of temperature, followed by the (3.2) and finally the (3.3) reaction. Therefore, the (3.1) reaction was in favor and the Al₃Ti was the favorable intermetallic to be formed. Other thermodynamic assessment of the Al-Ti intermetallic formation also concluded this order of intermetallic formation, the Al₃Ti has the lowest Gibbs free energy of formation, followed by the AlTi, and finally the AlTi₃ (Kattner et al., 1992; Khoshhal et al., 2010).

To investigate further the distribution of the phases in the samples, samples were fabricated with a ball milling time of 4 h. The surface morphology of the samples is shown in Figure 5A. The SEM results exhibited three regions consisting of bright, gray, and dark regions. The observation results suggested that the microstructure of the samples consisted of three components, where small dark particles were distributed on a gray matrix and a small white grain was occasionally present. The size of the dark particle was approximately 1–2 µm, and its distribution was not uniform throughout the matrix. Furthermore, EDS mapping for Al, O, and Ti, as shown in Figures 5B–D, respectively, suggested that the dark particles were aluminum–oxygen compounds, the gray matrix was a compound of aluminum and titanium, and the bright grain was a Ti-rich phase. On the basis of the crystal structural analysis, we suggested that the gray matrix was Al₃Ti, whereas the dark particles were Al₂O₃, and the white grain composition remained unidentified.

The formation of the in situ Al₃Ti–Al₂O₃ composite could be divided into two stages. Initially, the reaction between the two raw materials, i.e., aluminum and titanium dioxide, leads to the formation of alumina and titanium metal by the following reaction equation: 4 A l + 3 T i O 2   →   2 A l 2 O 3 + 3 T i

Thus, the titanium metal then reacts with aluminum and forms Al₃Ti by the following reaction equation: 3 A l + T i   →   A l 3 T i

Thermodynamically, the formation of Al₃Ti is favored because of the subzero Gibbs free energy of the reaction. The mechanism of the reaction is believed to be that the titanium atom diffuses into the crystal lattice of aluminum; thus, Al₃Ti has the lattice structure of aluminum (FCC). This reaction continues until the titanium metal is exhausted.

If the raw material had excess aluminum, then the final product includes Al₂O₃, Al₃Ti, and Al. If there was excess titanium, AlTi and AlTi₃ could be formed in accordance with the following reaction processing steps: A l 3 T i + 2 T i   →   3 A l T i A l T i + 2 T i   →   A l T i 3

On the basis of the observation of crystal structural analysis combined with surface morphology, a Ti-rich phase was found in the composite specimen, which indicated the presence of excess titanium in the raw material. In this case, the excess titanium might have caused the formation of the unidentified Al–Ti compound in the composite. Furthermore, the surface morphologies of samples fabricated at a milling time of 5 h are shown in Figure 6. The results showed that the Al₂O₃ particles were finer in size and had a better distribution in the matrix than those of samples fabricated with a milling time of 4 h, as shown in Figure 5A. The surface morphologies of the samples showed that the Ti-rich grains broke into thin streaks and remained in the composite. The observation results were consistent with the XRD characterization observations. In addition, the archived results further confirmed that the milling time was a key factor for enhancing the reaction of raw materials to form the phase during thermal treatment.

An increase in milling time provides extra energy for the reactions. Given that a change in the Ti-rich phase structure occurred, the Ti-rich phase can possibly react with other components to form a new phase. As suggested above, AlTi and AlTi₃ intermetallics could be formed in the matrix. The energy provided after 5 h of milling and sintering at 750°C for 30 min was not enough to initiate the formation of large-scale AlTi/AlTi₃ intermetallics, and the phases might appear after longer milling and/or higher temperature sintering.

The influence of milling time on the porosity and the microhardness of the Al₃Ti–Al₂O₃ composites is shown in Figures 7, 8, respectively.

As depicted in Figure 7, the porosity of the material ranges from approx. 14 ÷ 28%, which is relatively large for composite material. Low compaction pressure during compaction could be one of the causes while the others might be short milling time and relatively low sintering temperature. Longer milling time might help refine the particles and give better distribution of the components, as the porosity decreased from approx. 20% down to 14% when milling time increased from 4 to 5 h.

The observation results showed that the microhardness of the composite dramatically increased as the milling time increased from 1 h to 5 h, as shown in Figure 8. At a short milling time of 1 h, the microhardness was approximately 50 HV but dramatically increased to approximately 130 HV when the milling time was raised to 5 h. The microhardness obtained in this work was comparable to those reported for the Al–Al₃Ti–Al₂O₃ composite, also fabricated by the powder metallurgy route, which is approximately 140 HV (Lakra et al., 2020). Similar composites prepared by other methods, such as thermal decomposition, have considerably lower microhardness, i.e., approximately 50 HV (Azarniya and Hosseini, 2015). Increasing the milling time was expected to enhance the Al₃Ti distribution in the Al₂O₃ matrix, resulting in an enhancement in the microhardness. Thus, we suggested that an increase in milling time resulted in a smaller grain size, improved material diffusion, and thus created a more uniform structure, resulting in an improved microhardness of the composite.