The Effects of Double-Step Homogenization on Precipitation Behavior of Al₃Zr Dispersoids and Microstructural Evolution in 2196 Aluminum Alloy

Abstract

The effects of double-step homogenization processes on the precipitation of Al₃Zr dispersoids and the dissolution of the primary phases of 2196 aluminum alloy were studied by optical microscopy, scanning electron microscopy (SEM), and transmission electron microscopy (TEM). It was revealed that first-step homogenization facilitates the nucleation of Al₃Zr, while second-step homogenization results in the dissolution of the primary phases and the growth of Al₃Zr dispersoids. The nanosized θ’ precipitates formed in the first-step homogenization are dissolved after the second-step homogenization. The optimum homogenization process was selected as 400 °C/10 h + 520 °C/24 h, which effectively dissolves the primary phases and promotes the formation of refined distribution of Al3Zr dispersoids. This phenomenon is mainly caused by the highest nucleation rate of the Al₃Zr phase at 400 °C. While reducing the heating rate of the homogenization process can increase the number density of the Al₃Zr dispersoids and reduce the precipitate free zone (PFZ), it does not alleviate the inhomogeneity of the Al₃Zr precipitation. These results are expected to be meaningful for tailoring the industrial homogenization processing of as-cast Al-Cu-Li alloy.

1. Introduction

Al-Li alloys have emerged as the most promising lightweight materials for aerospace and space applications due to their excellent comprehensive properties, such as low density, high strength, high stiffness, and corrosion resistance [1,2,3,4]. As a typical third-generation Al-Li alloy, 2196 Al-Cu-Li alloy has been widely used in manufacturing aircraft floor beams, fuselage beams, trusses, and pillars, exhibiting obvious advantages compared to conventional 2xxx and 7xxx aluminum alloys [4,5,6].

Homogenization is a crucial step in the processing of 2196 alloy and achieving the desired properties. This process is effective in dissolving non-equilibrium coarse constituents, thereby reducing microsegregation and dendrites and improving the processability of the alloy. Moreover, homogenization plays an important role in controlling the recrystallization behavior of the alloys by forming nanosized coherent Al₃Zr dispersoids [7]. The presence of these dispersoids produces dispersion strengthening through the Orowan mechanism and inhibits recrystallization through grain boundary pinning, which ultimately, improves the mechanical properties of the alloy [8,9,10]. However, the homogenization process faces two major challenges: (1) Conventional single-step homogenization cannot fully dissolve all the coarse constituents with different melting points, resulting in an inadequate strengthening potential. (2) Due to the low diffusion rate and solubility of Zr in the Al matrix, Zr is prone to segregation at the dendrite center during solidification, resulting in low Zr concentration between dendrite arms and grain boundaries [11,12,13,14]. The uneven distribution of the Al₃Zr dispersoids is unfavorable for the recrystallization resistance of the alloy [13,14,15,16]. Additionally, the presence of coarse primary Al₃Zr constituent can even promote recrystallization [8,17,18].

Intensive efforts have been dedicated to optimizing the homogenization process of Al-Cu-Li alloys [10,11,12,13,19,20,21]. Based on homogenization kinetic analysis, Li et al. [19] proposed that the optimal homogenization process of Al-2.8Cu-1.4Li-0.3Mn-0.12Zr alloy was 510 °C/18 h and that the dendritic network as well as the microsegregation of the alloying elements were effectively eliminated. Guo et al. [16] proposed that the double-step homogenization exhibited great advantages compared to the conventional single-step homogenization for the 2195 alloy. Double-step homogenization not only prevents the occurrence of overburn but also provides more favorable conditions for the uniform precipitation of Al₃Zr dispersoids. The optimal homogenization parameter was selected as 460 °C/16 h + 515 °C/20 h. Deng et al. [20] further demonstrated that the double-step homogenization promoted the uniform distribution of Al₃Zr dispersoids and the precipitation of the T1 phase during the following aging treatment. Conversely, Liu et al. [10] proposed a double-step homogenization process of 350 °C/12 h + 470 °C/24 h for 2195 alloy, which resulted in fine and dense Al₃Zr distribution and, thus, a significant improvement in its properties. Therefore, inconsistent homogenization processes were reported for the Al-Cu-Li alloy, especially the first stage of the double-step homogenization, which limited the properties improvement in this type of alloy. In order to optimize the homogenization process for the Al-Cu-Li alloy system, a systematical investigation into the effect of double-step homogenization temperature on the precipitation of Al₃Zr dispersoids and dissolution of constituent phases is essential, especially for the first step of this homogenization.

In the present work, a commercial 2196 alloy with a low Cu/Li ratio of 1.67 (much lower than the 2195 alloy with a Cu/Li ratio of 4) was selected due to its extensive application of extrusion profiles. A series of double-step homogenization processes with different first-stage temperatures were designed to explore its effect on the microstructural evolution of the 2196 alloy. The dissolution of the constituent and precipitation of Al₃Zr dispersoids were systematically studied. Finally, the diffusion behavior of the solute atoms was studied through microstructural characterization and kinetic analysis, and the classical nucleation theory related to thermodynamics and kinetics was used to verify the model.

2. Materials and Methods

The experimental material was the as-cast 2196 Al-Cu-Li alloy with Φ125 mm in diameter and 250 mm in length, received from CHINALCO Research Institute of Material Application and Technology in Beijing, China. The alloy ingot was prepared by a casting metallurgy method under an argon atmosphere, and the composition of the studied alloy (wt.%) was examined by an inductively coupled plasma optical emission spectrometer (ICP-OES). The chemical composition of the alloy is shown in

Table 1. The chemical compositions (Wt.%) of 2196 aluminum alloy.

Alloy	Cu	Li	Mg	Ag	Mn	Zn	Zr	Fe	Al
2196	2.80	1.67	0.39	0.38	0.13	0.13	0.09	0.021	Balance

. Various homogenization treatments were carried out on the alloy. Firstly, first-step homogenization with 4 different temperatures (350 °C/10 h, 375 °C/10 h, 400 °C/10 h, and 425 °C/10 h) with a heating rate of 30 °C/h was performed. A heating rate of 30 °C/hour relates to heating from room temperature to the specified temperature at a rate of 30 °C/hour, after which the sample was held at this temperature for the specified time. After then, all the ingots were second-step homogenized at 520 °C/24 h and air-cooled to room temperature. Another sample with a heating rate of 300 °C/h at the first stage of homogenization was selected for comparison. The schematic diagram of the entire experimental route is shown in Figure 1.

The grain size of the samples was observed by polarized light optical microscopy (PLOM) performed using a ZEISS Axio Observer 3 material microscope (Carl Zeiss Microscopy Gmb, Jena, Germany). The second phases were characterized by scanning electron microscopy performed using a TESCAN S8000 GMH SEM (Carl Zeiss Microscopy Gmb, Jena, Germany) with an accelerating voltage of 20 kV. The morphology and structure of the Al₃Zr dispersoid were characterized using a FEI Tecnai G2 F20 transmission electron microscope (TEM, FEI, Portland, OR, USA) operated at 200 KV. The TEM samples were prepared by twin-jet polishing with an electrolyte solution of 30% HNO₃ and 70% CH₃OH. The operating temperature and voltage were set to −25 °C and 12 V, respectively. For the quantification of the dispersoids, more than 500 particles were counted based on 10 TEM images. The average thicknesses of samples were measured by convergent beam electron diffraction (CBED).

3. Results

3.1. As-Cast Microstructure

Figure 2 shows the representative SEM image and EDS quantification results of the as-deposited alloy. As shown in Figure 2a, the as-deposited alloy consisted of fine equiaxed grains with grain sizes of about 30–50 µm. Some pores were visible in the alloy, which is the main drawback of a spray-formed alloy. Some discontinuous or semi-continuously network primary phases were distributed along the grain boundaries. Additionally, abundant plate-like and irregular-morphology phases were observed in the grain interior. Figure 2b shows the SEM image of primary phases under higher magnification. The compositions of these second phases were analyzed by energy dispersive spectroscopy (EDS), and the results are shown in Figure 2c. Various types of primary phases, including Al-Cu, Al-Cu-Fe, Al-Cu-Mg, and Al-Cu-Zr phases can be identified. By comparing the results here with the results reported in the literature, it can be judged that the Al-Cu phase is θ(Al₂Cu), the Mg-containing phase is S(Al₂CuMg), and the Fe-containing phase is Al₇Cu₂Fe [18,19,21]. The plate-like particles in the interior of the grains can be identified as the T_B phase, as shown in Figure 2h.

The PLOM image of the as-cast alloy is shown in Figure 2a. The alloy exhibited a typical as-cast eutectic structure with a clear dendritic network within the grains. The average grain size of the as-cast alloy was measured as 142.5 ± 16.6 µm. From the backscatter electron (BSE) image shown in Figure 2b, large numbers of coarse continuous non-equilibrium eutectic phases were distributed along the grain boundaries, and some spherical or elongated particles were formed in the grain interior. The enlarged SEM image and corresponding EDS mappings are shown in Figure 2c–g. The EDS analysis results of points A, B, and C are shown in

Table 2. Chemical compositions (at.%) of intermetallic phases in Figure 2.

Point	Al	Cu	Mg	Cr	Fe	Mn
A	76.16	13.28	10.17	0.22	0.15	0.11
B	70.31	17.66	3.01	0.03	6.72	2.27
C	77.09	17.73	4.80	0.03	0.23	0.13

. It is clear that the dendrite network was mainly enriched with Cu, and some Mg signals can be detected in the regions with lower contrast. Based on the point analysis shown in Table 2, two primary phases can be identified: θ (Al₂Cu) and S (Al₂CuMg) phases. Moreover, some particles with high contrast can be identified as the Al₇Cu₂Fe-phase based on the EDS point results and the mapping shown in Figure 2d,f. The Mn element was partially detected in the Al₇Cu₂Fe phase, which was caused by the mutual replacement between Fe and Mn. The primary Al₃Zr phase was not detected in this alloy due to its low Zr concentration.

3.2. Microstructural Evolution during First-Step Homogenization

Figure 3 shows the PLOM and BSE images of the alloy after first-stage homogenization at different temperatures. The dendrite structure was generally reduced as the homogenization temperatures increased. The BSE images of Figure 3b,d,f,h show that the continuous primary phases distributed along the grain boundaries were gradually broken up and some spherical or elongated particles in the grain interior dissolved in the matrix. However, most of the dendritic network still existed, and the area fractions of the primary phases remained almost unchanged (the area fractions decreased from 3.45% to 3.10% as the first-stage homogenization temperatures increased from 350 °C to 425 °C), indicating that the low temperature in the first-step homogenization could not eliminate the macrosegregation and dendrites of the as-cast alloy. Except for the dissolution of the primary phases, some precipitations can be observed from the enlarged SEM images shown in Figure 4. Most of the precipitates exhibit a plate-like morphology. Based on previous literature [22,23,24], these precipitates can be identified as the θ’ phase. The distributions of the θ’ phase were inhomogeneous for all the homogenization temperatures, and a precipitate-free zone was observed in the core of the grains. Additionally, the number of densities in the θ’ phase generally decreased as the homogenization temperatures increased, which is mainly caused by the low drive force for precipitation during high-temperature homogenization.

3.3. Microstructural Evolution during Second-Step Homogenization

After second-step homogenization at 520 °C/24 h, regardless of the different first-step homogenization temperatures, most of the microsegregation was eliminated (shown in the PLOM images in Figure 5). BSE-SEM images of the four samples are shown in Figure 6, combined with the EDS results shown in

Table 3. Chemical compositions (at.%) of intermetallic phases in Figure 6.

Point	Al	Cu	Mg	Ag	Fe	Mn	Zn
A	82.66	7.48	1.70	0.39	2.25	5.41	0.09
B	90.25	7.91	1.00	0.56	0.03	0.06	0.16
C	87.33	9.07	2.32	0.21	0.03	0.93	0.10
D	80.57	11.69	4.39	0.38	4.39	0.57	0.36
E	73.97	16.33	1.49	0.18	6.41	1.05	0.53
F	94.63	1.79	2.20	0.12	1.05	0.11	0.09

, the low melting point (Cu, Mg, and Ag)-enriching phases were dissolved in the matrix, which left merely the insoluble Al₇Cu₂Fe phases. The Al₇Cu₂Fe phase was almost unchanged after the homogenization, which was mainly caused by the low solubility of Fe in the Al matrix. The S-Al₂CuMg phases were fully dissolved into the matrix, with only a limited number of Al₂Cu phases remaining in the Al matrix. This occurrence was likely due to insufficient homogenization time during production. The relatively low content of the residual Al₂Cu phase suggests that its impact on the precipitation process of the alloy may be negligible. Additionally, the nanosized θ’ precipitates formed in the first-step homogenization dissolved after the second-step homogenization at higher temperatures. At this stage, no clear difference was observed in the four samples with different first-step homogenization temperatures. The area fractions of the residual second phase in the four samples were basically the same (0.50%, 0.18%, 0.21%, and 0.26% for first-step homogenization at 350 °C, 375 °C, 400 °C, and 425 °C, respectively).

3.4. Precipitation Behavior of Al₃Zr Dispersoids during Different Homogenization Processes

Figure 7 shows the high-angle angular dark-field scanning transmission electron microscopy image and corresponding selected area diffraction (SAD) pattern of the Al₃Zr dispersoids formed after homogenization at 400 °C/10 h + 520 °C/24 h. The additional diffraction spots at the center of the Al spots confirmed the L1₂ structure of the Al₃Zr dispersoids.

To reveal the precipitation of the Al₃Zr dispersoids after the different homogenization processes, SEM characterization was performed on the four double-step homogenized samples. Figure 8 shows the distribution of the Al₃Zr dispersoids at the different regions of the grains. For the four samples, three different regions were observed. In the region closest to grain boundaries, designated as area 1, a few of the Al₃Zr dispersoids can be observed. This phenomenon can be attributed primarily to the diminished concentration of Zr within the region adjacent to grain boundaries. In area 2, a low density of Al₃Zr dispersoids with relatively larger sizes can be observed. Conversely, in the vicinity of the grain center, referred to as region 3, highly dense Al₃Zr dispersoids were generated. Therefore, the precipitation of the Al₃Zr dispersoids was inhomogeneous for all the homogenization processes. Figure 9 shows the HAADF-STEM images of the Al₃Zr dispersoids formed in the grain center (region three) of the four samples. The average particle size (r), number density (N_v), and volume fraction (f_v) of the Al₃Zr dispersoids formed in the four different double-step heat-treated samples are shown in

Table 4. Average radius, number density, and volume fraction of Al₃Zr dispersoids in the four different homogenized samples.

Homogenization Parameters	r (nm)	Nv (μm−3)	fv (%)	fv/r (μm−1)
350 °C/10 h + 520 °C/24 h	18	9.3	0.48	0.267
375 °C/10 h + 520 °C/24 h	14	9.7	0.32	0.229
400 °C/10 h + 520 °C/24 h	11	12.1	0.61	0.554
425 °C/10 h + 520 °C/24 h	13	7.7	0.39	0.3

. The sample homogenized at 400 °C/10 h + 520 °C/24 h exhibited the highest number density, volume fraction, and smallest average size. The number density of Al₃Zr in this process was more than 25% higher than in other processes, meaning it was selected as the optimum homogenization process in this work. For all four samples, the Al₃Zr dispersoids exhibited two different configurations: linear clusters and blocky clusters, which was consistent with previous reports [16,25].

3.5. Precipitation of Al₃Zr Dispersoids during Different Heating Rates

To study the effect of heating rate on the precipitation of Al₃Zr dispersoids, different heating rates (30 °C/h and 300 °C/h) were adopted for the optimized homogenization process of 400 °C/10 h + 520 °C/24 h. SEM images of the two samples with different heating rates are shown in Figure 10. The sample with a slow heating rate of 30 °C/h demonstrates a much higher number density of Al₃Zr dispersoids compared to the sample with a rapid heating rate of 300 °C/h, while the width of PFZ was smaller, indicating that the slow heating rate was beneficial in the precipitation of the Al₃Zr dispersoids. However, both kinds of samples exhibited the inhomogeneous distribution of the Al₃Zr phase in grains, implying that regulation of the heating rate could not efficiently improve the uniformity of the Al₃Zr phase.

4. Discussions

Based on the above results, it was demonstrated that the regulation of the first-step homogenization temperature can significantly promote the precipitation of Al₃Zr dispersoids. Specifically, a low-temperature treatment during the first stage of the double-step homogenization promotes the nucleation of Al₃Zr due to the large driving force, while a subsequent high-temperature treatment facilitates the diffusion of Zr atoms and promotes the formation of numerous Al₃Zr dispersoids. The homogenization treatment mainly consisted of the nucleation of Al₃Zr and the dissolution of primary phases, as discussed below.

4.1. Nucleation Model of Al₃Zr Dispersoids

The precipitation kinetics model of Al₃Zr dispersoids is based on the Kampmann–Wagner numerical (KWN) model [26], which has been widely used by previous researchers [15,26,27,28,29]. In the KWN model, the evaluation of the precipitate nucleation rate is based on the classical nucleation theory [30]. This theory applies to binary alloys and is generally considered as not strictly applicable to multicomponent alloys [31].

In the present work, an approximate method to calculate the nucleation rate was used, which ignored the incubation time before nucleation. The nucleation rate can be expressed by [18]:

$${\mathrm{J}=\mathrm{N}}_0\frac{\mathrm{kT}}{\mathrm{h}}\exp \left[\frac{-\left(\mathrm{Q}+{4\mathsf{\pi }\mathsf{\sigma }\mathrm{r}}^{{*}^2}\right)}{3\mathrm{kT}}\right]$$

(1)

where J is the nucleation rate per unit volume. N₀ is the number of nucleation sites per unit volume, which is calculated by the number of Zr atoms per unit volume in this paper. The values of average radius and numerical density are shown in Table 4. Q is the activation energy for the transfer of Zr across the nucleus interface (assumed to be approximately 2.07 eV atom⁻¹ [32]). The k and h are the Boltzmann constant and Planck constant, respectively. T is the thermodynamic temperature and σ is the interface energy, which is 0.1 J/m² [10,16]. The r^* is the critical radius.

The critical radius is calculated using the Gibbs–Thomson equation [18,28]. The critical radius r^* can be written as:

$${\mathrm{r}}^{*}=\frac{{2\mathsf{\sigma }\mathrm{V}}_{\mathrm{a}}}{\mathrm{kTln}\frac{\mathrm{c}}{{\mathrm{c}}_{\infty }^{\mathsf{\alpha }}} }$$

(2)

where V_a is the atomic volume, c is the instantaneous concentration of Zr in the matrix, and ${\mathrm{c}}_{\infty }^{\mathsf{\alpha }}$ is the concentration of Zr in the matrix when it is in equilibrium with Al₃Zr. The ${\mathrm{c}}_{\infty }^{\mathsf{\alpha }}$ value is obtained from the solvent line of the metastable Al₃Zr phase in the binary Al-Zr system calculation [18]. The radius of each newly formed particle is set to be slightly larger than the critical radius so that these particles can grow. According to the KWN procedure, newly nucleated particles are arbitrarily taken as 10% greater than r^*.

After normalization, the nucleation rate of Al₃Zr dispersoids alongside temperature is shown in Figure 11. The peak nucleation rate occurs at 400 °C, and then, the nucleation rate decreased rapidly alongside the increase in temperature. Therefore, the nucleation rate of Al₃Zr dispersoids at 400 °C was the largest, which was more consistent with the experimental results, in which the highest number density of Al₃Zr dispersoids was formed using this condition.

4.2. Homogenization Kinetic Analysis

The experimental results show that the nonequilibrium eutectic phase and dendrite segregation gradually disappear in the homogenization process. The distributions of the major elements (Cu, Mg, and Ag) in the grain and at the grain boundaries are uneven. Therefore, it is of great significance to study the diffusion law in the dendritic region for studying the distribution of elements in the homogenization process. According to References [33,34], the initial concentration of the elements in the interdendritic region can be expressed by the Fourier series components of the cosine function:

$$\mathrm{w}(\mathrm{x})=\overline{\mathrm{w}}+0.5\Delta {\mathrm{w}}_0\cos (2\mathsf{\pi }\mathrm{x}/\mathrm{L})$$

(3)

where L is the interdendritic spacing, $\Delta {\mathrm{w}}_0$ is the initial amplitude of the composition segregation, and $\overline{\mathrm{w}}$ is the average concentration of the element.

According to Fick’s second law [35] and boundary conditions, w (x,t) is:

$$\mathrm{w}(\mathrm{x},\mathrm{t})=\overline{\mathrm{w}}+0.5\Delta {\mathrm{w}}_0\cos (2\mathsf{\pi }\mathrm{x}/\mathrm{L})\exp \left(-4\mathrm{Dt}{\mathsf{\pi }}^2/{\mathrm{L}}^2\right)$$

(4)

where D is the diffusion coefficient of the elements in the matrix, and t is the diffusion time.

The attenuation law of the cosine distribution in Equation (4) can be described by the attenuation function [36]:

$$\mathrm{w}(\mathrm{x},\mathrm{t})=0.5\Delta {\mathrm{w}}_0\exp (-{4\mathrm{Dt}\mathsf{\pi }}^2{/\mathrm{L}}^2)$$

(5)

whereby it is generally believed that when the component segregation amplitude decreases to 1%, the element distribution is uniform, then:

$$1\%=\exp (-4\mathrm{Dt}{\mathsf{\pi }}^2/{\mathrm{L}}^2)$$

(6)

the diffusion coefficient D is:

$${\mathrm{D}=\mathrm{D}}_0\exp (-\mathrm{Q}/\mathrm{RT})$$

(7)

where D₀ is the independent coefficient, Q is the diffusion activation energy, R is the gas constant, and T is the absolute temperature.

Replace Equation (7) and substitute Equation (6), the equation can be written as:

$$\frac{1}{\mathrm{T}}=\frac{\mathrm{R}}{\mathrm{Q}}\ln (4{\mathrm{tD}}_0{\mathsf{\pi }}^2/4.6{\mathrm{L}}^2)$$

(8)

assuming A = R/Q, B = 4.6/4π²D₀, the homogeneous dynamic equation can be obtained:

$$\frac{1}{\mathrm{T}}=\mathrm{A}\ln (\mathrm{t}/\mathrm{B}{\mathrm{L}}^2)$$

(9)

previous studies have shown that at the same temperature, the diffusion coefficient of Cu was much lower than for Mg, Zn, and Mn. Therefore, it can be considered that the homogenization process is mainly affected by Cu diffusion [21,34].

Set D₀ (Cu) = 0.084 cm²·s⁻¹, Q (Cu) = 136.8 kJ·mol⁻¹, and R = 8.314 J/(mol·K), as introduced in Equation (9), and homogenization kinetics curves of 2196 Al-Cu-Li alloy at different dendrite spacings can be obtained, as shown in Figure 12. Here, as the temperature increased, the diffusion time of elements shortened, and the microstructure greatly improved.

Quantitative metallographic analysis shows that the average dendrite spacing in the as-cast alloy was about 100 ± 15 μm, calculated from Figure 3. According to the homogenization kinetic curves, at a homogenization temperature of 520 °C, the corresponding diffusion time was about 24.8 h. This is in good agreement with the experimental results. Therefore, second-step homogenization at 520 °C/24 h can effectively dissolve the primary phases formed during casting, even though a low content of Al₂Cu phase can still be observed. Compared to the previous literature, and specifically reference [10], the optimal homogenization process for 2195 alloy was determined as 350 °C/12 h + 470 °C/24 h, with quantitative statistics indicating an average particle size of Al₃Zr at 22.6 ± 4.1 nm. Meanwhile, in reference [16], the optimum two-step homogenization parameter for as-cast 2195 Al-Li alloy was found to be 460 °C/16 h + 515 °C/20 h, which facilitates uniform precipitation of Al₃Zr particles with a diameter of 25.23 nm. After undergoing two-step homogenization at 460 °C/12 h + 510 °C/24 h, the Al₃Zr dispersoids in Al-Cu-Li-Zr alloy exhibited a finer and more uniform microstructure with an average size of 24.9 ± 1.1 nm, while the width of PFZ narrowed, as reported in reference [20]. In this paper, the optimized homogenization process of 400 °C/10 h + 520 °C/24 h can effectively dissolve the primary phases and refine the distribution of Al₃Zr dispersoids. As a result, the average particle size of the Al₃Zr dispersoids was reduced to 17.7 nm, which is expected to suppress the recrystallization and improve the mechanical properties of the alloys during the subsequent processing.

5. Conclusions

In this work, the influences of two-step homogenization processes on the dissolution of primary phases and precipitation of Al₃Zr dispersoids of an as-cast 2196 alloy were investigated. The main conclusions can be drawn as follows:

• During the second-step homogenization, regardless of the different first-step homogenization temperatures, most of the microsegregation is eliminated, and the θ’ precipitates formed during first-step homogenization are dissolved. The different first-step homogenization temperatures can result in a different distribution of Al₃Zr dispersoids after a second-step homogenization at 520 °C. The optimized homogenization process of 400 °C/10 h + 520 °C/24 h can effectively dissolve the primary phases and refine the distribution of Al₃Zr dispersoids.
• Decreasing the heating rate of the homogenization process can greatly increase the number density of the Al₃Zr dispersoids. However, regulation of the heating rate could not improve the uniformity of the Al₃Zr dispersoids.
• The corresponding precipitation of the Al₃Zr dispersoids model and dissolution kinetics of the primary phases have been developed for the homogenization process of the 2196 alloy, which corresponds well with these experimental results.