Experimental Characterization and First-Principles Calculations of Zn Segregation at the β″-Mg₅Al₂Si₄/Al Interfaces in Al-Mg-Si Alloys

Abstract

In 6000-series Al-Mg-Si alloys, Zn is commonly added to enhance the aging response for improved properties. However, the segregation of Zn to interfaces and its interaction mechanisms with strengthening phases remain unclear. Combining experiments and theoretical calculations, we systematically investigated Zn segregation behavior at the β″/Al interfaces in Al-Mg-Si alloys. Using a modified β″-Mg₅Al₂Si₄ model, we determined that the interface (100)_β″//(130)_Al has a slightly smaller formation energy of 1.06 kJ/mol, while (001)_β″//($\overline{3}$20)_Al has a relatively smaller interfacial energy of 116 mJ/m². HAADF-STEM analysis revealed these interfacial morphologies and atomic distributions, showing that Zn atoms not only enter the β″ phase but are also inclined to segregate at the interfaces by occupying the Si3/Al sites. Furthermore, the stability of the β″/Al interfaces and the Zn segregation behavior are well explained at the atomic scale, with calculations showing that stronger hybridization between Zn-3d and Si-3s orbitals facilitates Zn segregation at the interfaces.

1. Introduction

6000-series alloys are widely used in automobiles due to their excellent comprehensive mechanical properties, high specific strength, high quality of welded surface and good corrosion resistance. The alloys can be strengthened by heat treatment and therefore have higher strength in the peak-aged state. The age-hardening effect of Al-Mg-Si alloys is attributed to the precipitation of a high number density of nano-sized and (semi-)coherent metastable phases. Early investigations have shown that the precipitation sequence of Al-Mg-Si alloys can be defined as supersaturated solid solution (SSSS) → Mg-Si GP zone → intermediate phase β″ → intermediate phase β′(U1, U2, B′) → equilibrium phase β [1,2,3]. The needle-like β″-Mg₅Si₆ precipitates are generally considered to be the main contributors to the hardening. They have a monoclinic structure with lattice parameters of a = 1.516 nm, b = 0.405 nm, c = 0.674 nm, β = 105.3° and α = γ = 90°, as determined by Zandbergen et al. [4] through high-resolution electron microscopy (HREM) observations. Combining atom probe tomography (ATP) and first-principles calculations, Hasting et al. [5] indicated that the β″ phase contains Al and Mg/Si ≈ 1.1, and its most stable structural stoichiometry ratio is Mg₅Al₂Si₄. In addition, Saito et al. [6] further theoretically studied the lattice mismatch between the β″ phase of different compositions and the Al matrix, where the lattice mismatch between Mg₅Al₂Si₄ and the Al matrix was closest to the experiment results. The result of this study reconfirmed the composition of the β″ phase as Mg₅Al₂Si₄.

Recently, experimentalists have demonstrated that the addition of Zn can effectively increase the strength of the Al-Mg-Si alloys. According to Zhang et al. [7], the yield strength (YS) and ultimate tensile strength (UTS) of Al-2.40Mg-0.67Si (wt.%) alloys were significantly improved by up to 52.2% and 8.5% after the addition of Zn compared to the alloy without Zn. After proper heat treatment, they were further increased by up to 35% and 28%, respectively, compared to the as-cast state. This is mainly due to the prominent role of the alloying element Zn in controlling the precipitation process of strengthening phases. Most studies also reported that the addition of Zn could promote the precipitation of Mg-Si phases and significantly improve the aging response of Al-Mg-Si alloys [8,9]. For example, Zhu et al. [10,11] showed that the addition of Zn to 6000 series alloys accelerated the formation of aging precipitated phases and increased the number density of precipitated phases, thus enhancing the artificial age-hardening response of the alloy. Saito et al. [12] increased the number of needle-like precipitated phases during the aging process by adding 1 wt.% Zn to Al-Mg-Si alloys, which correspondingly improved the hardness of the alloy. Guo et al. [13] found that in Al-Mg-Si-Cu-Zn alloys with 2.0–4.0 wt.% Zn, both Mg-Zn and Zn-containing Mg-Si precipitated phases could appear during the aging process, whereas only the latter were observed in alloys with 1.0 wt% Zn. Bartawi et al. [14] pointed out that small additions of Zn (<0.06 wt.%) to the AA6082 alloy do not change the aging precipitation sequence but affect the precipitate structure of the β″ phase by introducing sub-unit structures, leading to an increase in disordered structures. Xu et al. [15] also stated that the improved strength of the Al-Mg-Si alloys is attributed to the increase in number density of β″ precipitates when Zn content does not exceed 1.0 wt%. Yang et al. [16] combined 3DAP and TEM to investigate the age-hardening response in Al-0.5Mg-0.5Si (wt.%) alloys with 5.0 wt.% Zn addition and found that Zn addition promoted the precipitation of a higher number density of Mg-Si precipitates and contributed a higher Mg/Si ratio because of the formation of Mg-Zn precipitates. Therefore, it is recognized that the precipitation of Mg-Si phases can be effectively promoted by Zn addition and a proper amount of Zn can significantly improve the age-hardening response of Al-Mg-Si alloys. The age-hardening effect of Zn addition is mainly derived from the enhancement of related clusters, GP zones and β″ precipitates [7]. However, the underlying mechanisms of Zn’s role are not well understood, and even the interaction between Zn and the main hardening phase β″ during the ageing process of the Al-Mg-Si alloys is not clear.

A thorough understanding of the atomic formation mechanisms of precipitates in aluminum alloys is essential for alloy design and modification. By integrating experimental observations with theoretical simulations, particularly using Density Functional Theory (DFT), researchers have already studied the electronic structures and orbital hybridization characteristics of some precipitates in Al-Mg-Zn and Al-Mg-Si systems [17,18,19]. This approach provides a solid theoretical foundation for optimizing the performance of aluminum alloys. For instance, by employing HAADF-STEM and DFT calculations, Saito et al. [20] studied the structure of Zn-containing β″ phases and suggested that these Si3/Al sites were preferential for Zn incorporation. Wang et al. [21] systematically investigated the interfacial properties between the β″-Mg₅Si₆ precipitate and the Al matrix, and obtained low-energy interfaces. Ehlers [22] studied the β″/Al interface configuration stability over the full precipitate cross-section in an Ag-Mg-Si alloy by a first-principles-based hierarchical multi-scale model scheme. In addition, several studies on the interaction of alloying elements with the β″/Al interface have been reported. For example, it has been reported that the preferential occupancy of Cu atoms at the interface could increase the interfacial energy and provide strain relief for the misfit dislocations due to their small atomic volume [23]. The micro-alloying element Sn [24] could significantly refine the precipitate microstructures in Al-Mg-Si alloys aged at 250 °C and lead to the formation of composite β′/β″ precipitates. However, there is still a lack of the research on Zn segregation at the β″/Al interfaces in Al-Mg-Si alloys.

Here, HAADF-STEM characterization methods combined with first-principles calculations will be applied to systematically investigate Zn segregation behavior at the β″/Al interfaces in Al-Mg-Si alloys. The interface structures and interaction mechanisms are analyzed at the atomic level based on a modified β″-Mg₅Al₂Si₄ model. Therefore, a theoretical basis will be provided for the study of the aging precipitation behavior of Al-Mg-Si alloys for automotive plates.

2. Materials and Methods

2.1. Experiments

The experimental material for this study was based on the Al-0.9Si-0.6Mg-0.1Mn and Al-0.9Si-0.6Mg-0.6Zn-0.1Mn alloys (wt.%), and alloys were cold-rolled into 0.95 mm thick sheets. The cold-rolled sheets were solution heat-treated at 550 °C for 5 min and then quenched to room temperature. Subsequently, the alloy was isothermally aged at 170 °C for 8 h to obtain the microstructures of the β′′ phases, as shown in Figure 1.

Tensile tests were conducted using a universal testing machine (CMT4303, Meister Industrial Systems, Shenzhen, China) at a tensile rate of 2 mm/min. The tensile specimens were rod-shaped, with a gauge length of 30 mm and a diameter of 5 mm. Hardness tests were performed using a Vickers hardness tester (WOLPERT 430SVD, Norwood, MA, USA) under a test load of 5 kg and a dwell time of 15 s.

Bright-field TEM images were captured using a transmission electron microscope (Tecnai G2 F20 S-Twin, FEI Company, Hillsboro, OR, USA) operating at 200 kV, with the imaging direction along the [100]_α axis. HAADF-STEM experiments were performed using a spherical aberration probe-corrected transmission electron microscope (Titan Cubed Themis G2 300, FEI Company, Hillsboro, OR, USA) operating at 300 kV. The STEM resolution was ≤ 0.06 nm and the collection angle of the HAADF detector was 27–165 mrad. All HAADF-STEM images were taken along the [100]_α zone axis and were Fourier filtered with a bandpass mask covering all the spots, followed by an inverse Fourier-filtered transform. TEM specimens were prepared by electropolishing with a solution of 1/3 HNO₃ and 2/3 methanol at a temperature of −25 °C.

2.2. DFT Calculations

All calculations were performed by using the Vienna ab initio simulation package (VASP) in the framework of density of functional theory. For Al, Mg, Si and Zn atoms, the projector augmented wave (PAW) method was employed for describing the electron-iron interaction, [25,26,27] and the generalized gradient approximation (GGA) [28] in the Perdew–Burke–Ernzerhof (PBE) form for the exchange and correlation between electrons [29]. The energy relaxation for each system was continued until the forces on all the atoms were converged to within 0.001 eV/Å. The cut-off energy for the basis set was 500 eV. The k-point convergence was tested for all cell optimizations, and the density of the Monkhorst–Pack k-point grid was increased until the total energy of the system was converged to an accuracy of 1 meV.

3. Results and Discussions

3.1. Comparative Analysis of Mechanical Properties in Zn-Containing and Zn-Free Alloys

Figure 2a shows the mechanical properties of the two alloys aged at 170 °C for 8 h. It is evident that the alloy containing 0.6 wt.% Zn exhibits significantly superior performance compared to the Al-Mg-Si alloy. Remarkably, the strengths of Zn-added alloy increase by nearly 3.6%, resulting in a yield strength of 340 MPa and an ultimate tensile strength of 373 MPa. Furthermore, the addition of Zn also leads to an increase in the peak-aged hardness of the Al-Mg-Si alloy, as shown in Figure 2b. In conclusion, the incorporation of 0.6 wt.% Zn into the Al-Mg-Si alloy results in an improvement in the mechanical properties of the alloy. The alloy’s superior performance, significantly improved by the addition of Zn, correlates with its influence on microstructural features during the aging process, as discussed later in the following sections.

3.2. The Experimental Characterization of Zn Segregation at β″/Al Interfaces

As shown in Figure 3, both alloys aged at 170 °C for 8 h exhibit the precipitation of needle-like β″ phases. The addition of a small amount of Zn significantly increases the quantity of β″-phase precipitates, which is the primary reason for the enhanced mechanical properties in the Zn-containing alloy.

Furthermore, as reported in previous work [20,21], three typical β″/Al interfaces were observed based on the structural characteristics of the precipitates: (100)_β″//(130)_Al, (001)_β″//($\overline{3}$20)_Al and (010)_β″//(010)_Al. It was found that the β″ phase grows in the [100]_Al direction perpendicular to the (010)_β″//(010)_Al interface, adopting a needle-like morphology. This is because the interfacial orientation (010)_β″//(010)_Al has a relatively higher energy than (100)_β″//(130)_Al and (001)_β″//($\overline{3}$20)_Al, as revealed by DFT theoretical calculations [21] based on the β″-Mg₅Si₆ model.

Figure 4 displays the atomic structures of the β′′ phase and its relatively stable interfaces with the Al matrix by HAADF-STEM. The crystal cell of the β′′ phase is highlighted with yellow lines, and its low-density cylinder (LDC) characteristic structure is marked with white dashed lines. According to the imaging principle of the HAADF-STEM mode, the brightness of the atomic columns in the image is determined by the atomic number Z, which indicates that Zn atomic columns are the brightest in the HAADF-STEM images. In previous work, Saito et al. [6] indicated that the lattice mismatch between Mg₅Al₂Si₄ and the Al matrix is closest to the experiment by considering different compositions of the β″ phase, confirming that the composition of the β″ phase is Mg₅Al₂Si₄. Therefore, with reference to the β″-Mg₅Al₂Si₄ model, we marked the atomic distribution near the interfaces (100)_β″//(130)_Al and (001)_β″//($\overline{3}$20)_Al separately. It can be expected from Figure 4 that Zn atoms can not only enter the bulk of the β′′ phase, but also segregate at the β″/Al interfaces. Additionally, it can be observed that both interfaces pass through the inversion center of the β″ phase. The Si3/Al sites near the two interfaces are occupied by the segregated Zn atoms. In the subsequent sections, DFT calculations will be employed to thoroughly investigate the energetics of Zn substitution at various lattice sites, as well as the atomic- and electronic-scale interactions involved.

3.3. The Modelling of Zn Segregation at β″/Al Interfaces

3.3.1. The Stability of β″/Al Interfaces

To accurately model Zn segregation behavior at the β″/Al interfaces in Al-Mg-Si alloys, we firstly developed two atomic structures of the stable interfaces based on the β″-Mg₅Al₂Si₄ models. Figure 5 gives the structures of two stable interfaces (100)_β″//(130)_Al and (001)_β″//($\overline{3}$20)_Al, which correspond to interfaces A and B, respectively.

The energy of formation of the interfaces can be evaluated by the following equation:

$$E_{\mathrm{f}}=E_{\mathrm{AB}}-{xNE}_{\mathrm{A}}-\left(1-x\right){NE}_{\mathrm{B}},$$

(1)

where E_AB is the total energy of the supercell of the interfaces with N atoms, which is composed of the phases A and B. x denotes the phase fraction of the phase A, and (1 − x) represents that of the phase B. E_A and E_B represent the energy per atom of the fully relaxed phases A and B. Then, the formation energy can be divided into two parts: the interfacial energy and the elastic strain energy from the mismatch between A and B. The corresponding relationship can be expressed as

$${\displaystyle \frac{E_{\mathrm{f}}}{N}}={\displaystyle \frac{2\mathrm{S}\mathsf{\sigma }}{N}}+\mathsf{\xi },$$

(2)

where S refers to the area of the interface, $\mathsf{\sigma }$ is the interfacial energy per unit area, and $\mathsf{\xi }$ is the strain energy per atom. The factor is equal to two as there are two interfaces in the periodic supercell. The interfacial energy σ can be derived as follows:

$$\mathsf{\sigma }={\displaystyle \frac{1}{2\mathrm{S}}}(E_{\mathrm{AB}}-E_{\mathrm{A}}^0-E_{\mathrm{B}}^0),$$

(3)

$E_{\mathrm{A}}^0$ and $E_{\mathrm{B}}^0$ are the total energies of two slabs in the interface supercell, which are constrained to match the lattice constants at the interfaces.

The lattice parameter, energy of formation and interfacial energy of β″-Mg₅Al₂Si₄//Al interfaces are listed in

Table 1. The lattice parameter, energy of formation and interfacial energy of stable β″-Mg₅Al₂Si₄//Al interfaces. The data from β″-Mg₅Si₆//Al interfaces in reference [21] are listed as a comparison.

Interface	Lattice Parameter				Energy of Formation (kJ/mol)	Interfacial Energy (mJ/m2)
	a (Å)	b (Å)	c (Å)	γ (°)
A	14.92	6.606	4.054	105.6	1.06	162
Ref. [21]	14.94	6.591	4.046	106.5	1.06	124
B	15.01	6.651	4.007	105.4	1.29	116
Ref. [21]	14.83	6.651	4.052	106.8	1.27	100

. It can be seen that the results of our simulations are consistent with those in the literature, even though the taken atomic models of β″ are different [21]. Also, we can safely deduce that the entry of Al into the Mg₅Si₆ phase does not affect the preference of the β″//Al interface orientation. By comparing the stability of β″-Mg₅Al₂Si₄//Al interfaces, it can be obtained that the interface A has a slightly smaller formation energy of 1.06 kJ/mol while the interface B has a relatively smaller interfacial energy of 116 mJ/m². Compared to other interfacial orientations, these two interfaces are easily formed and observed in experiments.

3.3.2. The Occupation Sites of Zn at β″/Al Interfaces

Based on the above interface models, the preferred occupation sites of Zn at the β″/Al interfaces are also investigated. The substitution energy ($∆E_{\mathrm{sub}}$) of Zn at the interfaces is defined as follows:

$$∆E_{\mathrm{sub}}=(E_{\mathrm{tot}}^{\mathrm{sub}}-E_{\mathrm{tot}})-({\mu }_{\mathrm{Zn}}-{\mu }_i),$$

(4)

where $E_{\mathrm{tot}}^{\mathrm{sub}}$ and $E_{\mathrm{tot}}$ denote the total energy of the interface supercell after and before the i atom is substituted by the Zn atom. μ denotes the chemical potential of the corresponding element. For Al, it is referenced to the chemical potential in its bulk state, while for other elements, they are referenced to the substitution sites in the Al matrix. A negative value of $∆E_{\mathrm{sub}}$ means that Zn occupation of these sites is energetically favorable, with more negative values suggesting a stronger driving force for Zn incorporation at these positions.

The results of the calculated Zn substitution energies are summarized in Figure 6. Near both interfaces, Mg atoms are hardly replaced by Zn atoms with high substitution energies of 0.64–0.67 eV. Compared to Mg atoms, the substitution energies of Si are relatively small, being 0.36–0.42 eV. Notably, Zn substitution for Al exhibits the most favorable energetics, with the lowest energy range of −0.05 to −0.01 eV, and negative substitution energies for the majority of positions. The negative values underscore the thermodynamic preference for Zn to occupy Al sites, particularly at the β″/Al interfaces, where it can significantly reduce the formation energies, thereby stabilizing these structures. With negative substitution energies, it is expected that Zn can easily segregate at β″/Al interfaces. Furthermore, the penetration of Zn into the bulk β″ phase through substitution at Si3/Al sites (Al4 and Al5 in Figure 6a) has also been investigated. These sites present the most favorable scenario for Zn incorporation within the bulk, with a notably low substitution energy of −0.11 eV. This finding is in alignment with the results from HAADF-STEM experiments, which confirm the susceptibility of Si3/Al sites to Zn substitution in both the bulk phase and at the interfaces.

The underlying mechanism of Zn segregating at the interfaces was also revealed by analyses of the density of states (DOSs) and the bond orders. The bond orders were evaluated by using the density-derived electrostatic and chemical (DDEC6) charge partitioning method, as implemented in the CHARGEMOL code [30,31,32]. As shown in Figure 7, Zn-3d orbitals show a narrow and intense peak near −8 eV energy. In Al bulk, they hybridize more strongly with Al-3s orbitals. The average bond order of the Zn-Al bond is 0.24 and the sum of the bond orders for Zn is 2.91. When segregated to the Al1 site at the interface, Si-3s and Mg-3s orbitals together are involved in hybridization with Zn-3d orbitals. In contrast, the interaction with Mg-3s orbitals is significantly weaker. The calculated bond orders of the Zn-Mg, Zn-Si and Zn-Al bonds are 0.05~0.1, 0.5~0.64 and 0.21~0.33, respectively. The bond strength of Zn-Si is even larger than that of Zn-Al. It is also found that the sum of the bond orders for Zn is increased to 3.13. It can be safely concluded that the stronger hybridization between Zn-3d and Si-3s orbitals promotes Zn segregation at the interfaces.

4. Conclusions

By combining experiments and theoretical calculations, we systematically investigated Zn segregation behavior at the β″/Al interfaces in Al-Mg-Si alloys. The β″ phase prefers to grow along the <100>_Al direction, adopting a needle-like morphology as observed in the HAADF-STEM experiments. Based on the modified β″-Mg₅Al₂Si₄ model, the interface (100)_β″//(130)_A has a slightly smaller formation energy of 1.06 kJ/mol while the interface (001)_β′′//($\overline{3}$20)_Al has a relatively smaller interfacial energy of 116 mJ/m². Zn atoms not only enter the β″ phase, but also are inclined to segregate at the interfaces by occupying the Si3/Al sites, facilitated by strong Zn-3d and Si-3s orbital hybridization. The addition of Zn contributes to enhanced mechanical properties, with the alloy exhibiting a yield strength of 340 MPa, an ultimate tensile strength of 373 MPa, and increased peak-aged hardness. These findings elucidate the role of Zn in reinforcing the β″ phase within 6000 series alloys.