betaFeedback

WorksPhysics and Astronomy

Decoding Complex Compositions in Topologically Close-Packed Nano-plates of Magnesium Alloys: A High-Throughput Route to Stable Precipitates

Abstract

Coherent topologically close-packed nanoplates in magnesium alloys can effectively improve strength and creep resistance.However, the formation mechanisms of several metastable TCP nanoplates remain unclear, and the traditional trial-and-error methods hinder the rapid discovery of novel TCP precipitate strengthened Mg alloys.By integrating the thermodynamic and kinetic conditions for TCP precipitation, we developed a two-step high-throughput screening strategy, identifying 43 previously unreported TCP nanoplates in a series of Mg alloys. These findings highlight the critical need for precise compositional characterization of nanoprecipitates and establish a theoretical framework for designing creep-resistant Mg alloys containing TCP nanoplates.

Corresponding author.

Email: [email protected] (X.Y. Pang)

[email protected] (G. W. Qin)

Magnesium alloys, as the lightest structural metallic materials, have attracted considerable research attention for their applications in aerospace and automotive industries, wherein weight reduction is crucial for lowering energy consumption and CO2CO_{2} emissions. In Mg-alloy development, precipitation hardening is often recognized as an effective strengthening strategy, achieved by forming dense, coherent nanoprecipitates during aging treatments [1]. Among these nanoprecipitates, coherent topologically close-packed (TCP) nanoplates along {0001}hcp\{0001\}_{\mathrm{hcp}} basal planes are particularly significant, as they not only provide exceptional strength enhancement but also improve creep resistance, making them ideal for elevated-temperature applications [1, 2, 3, 4]. These TCP nanoplates have been identified across various alloy systems, including stable phases like Mg2Ca\mathrm{Mg}_{2}\mathrm{Ca} and Al2Ca\mathrm{Al}_{2}\mathrm{Ca} phases in Mg-Ca-(Al) alloys [5, 6, 7], metastable β2′\beta_{2}^{\prime}-MgZn2_2 phase in Mg-Zn alloys [8, 9], and single unit-cell height γ′′\gamma^{\prime\prime} metastable phases in Mg-RE (rare earth)-(Zn, Ag) series alloys [4, 10, 11]. Given the outstanding performance of TCP nanoplates in Mg alloys, the systematic identification of alloy systems capable of forming TCP nanoplates has become a crucial priority in Mg alloy development.

However, traditional trial-and-error methods are costly and inefficient for rapid alloy development. Although experimental phase diagrams and density functional theory (DFT)-calculated convex hull diagrams provide thermodynamic insights into equilibrium or metastable phases [13], they cannot determine whether a TCP phase can precipitate from the parent matrix (a process intricately associated with kinetics) and what its resulting morphology will be. Our recent study[12] reveals that the full structural evolution pathway of TCP nanoplates within the hcp-Mg matrix is governed by the formation of 3-layer {0001}hcp\{0001\}_{\mathrm{hcp}}-oriented unstable solute clusters (designated as MRM_R-type hcp-ordering, Figure 1), involving only shuffle-based displacements. During aging, under thermal fluctuations, once solute clusters evolve into the 3-layer MRM_R-type structures at a specific moment, they will spontaneously collapse into a TCP structure. Remarkably, this structural instability of specific solute clusters can be identified by structural relaxation in DFT calculations, establishing a kinetic criterion for TCP formation: if the 3-layer MRM_R-type hcp-ordering fails to induce hcp→\toTCP transitions upon structural relaxation, coherent TCP nanoplates cannot form kinetically.

Schematic illustration of the spontaneous transformation of unstable 3-layer M_R-type hcp-ordering into a 3-layer TCP structure upon structural relaxation

Figure 1. Schematic illustration of the spontaneous transformation of unstable 3-layer MRM_{\mathrm{R}}-type hcp-ordering into a 3-layer TCP structure upon structural relaxation [12].

Meanwhile, understanding the formation mechanism of TCP nanoplates from a thermodynamic perspective is also crucial. Here, we first use several common TCP nanoplates to elucidate the thermodynamic conditions for the precipitation of stable and metastable TCP nanoplates from typical dilute Mg alloys and explain the physical mechanism behind complex experimental phenomena. For stable/equilibrium phase precipitation, taking the Mg2Ca\mathrm{Mg}_2\mathrm{Ca} and Al2Ca\mathrm{Al}_2\mathrm{Ca} phases in Mg-Ca-(Al) alloys as examples, we reconstructed the 0 K Mg-Ca and Mg-Al-Ca convex hull diagrams (see Figure 2) using structural data from the Open Quantum Materials Database (OQMD)[14, 15]. Convex hull analysis is a powerful method for evaluating phase stability. A phase is thermodynamically stable if its formation energy at a given composition is lower than that of all other phases of the same composition and any linear combination of energies of other phases in the phase space. The collection of all stable phases in a given phase space constitutes the convex hull, while phases located above the convex hull are thermodynamically unstable. The parameter EhullE_{\mathrm{hull}} measures the distance from the convex hull line and quantifies thermodynamic stability, with stable phases residing exactly on the hull (Ehull=0.0E_{\mathrm{hull}}=0.0 eV/atom) and metastable/unstable phases located above it (Ehull>0.0E_{\mathrm{hull}}>0.0 eV/atom). According to convex hull theory[16], in binary or ternary systems, unstable phases with compositions between stable phases will decompose into their neighboring stable phases, while those above a stable phase with identical composition will transform into that stable phase.

DFT-calculated zero-temperature Mg-Ca and Mg-Al-Ca convex hull diagrams

Figure 2. DFT-calculated zero-temperature Mg-Ca and Mg-Al-Ca convex hull diagrams.

Figure 2 shows that both Mg2_2Ca with a C14C14 structure and Al2_2Ca with a C15C15 structure lie on the convex hull, indicating that these two Laves structures are stable phases. Additionally, they also form a tie-line with Mg. A tie-line represents a straight line connecting two stable phases on a convex hull diagram, meaning that two phases can coexist. Therefore, Mg-Ca-(Al) alloys of suitable compositions thermodynamically favor the precipitation of stable Mg2_2Ca and Al2_2Ca phases during aging. In convex hull diagrams, precipitation of stable TCP phase from the parent Mg matrix not only demands that phase displays Ehull=0.0E_{\mathrm{hull}}=0.0 eV/atom but also forms a tie-line with Mg. In contrast, the thermodynamic precipitation mechanism of metastable TCP nanoplates is more intricate. Here, we first aim to explain why the originally stable MgZn2_2 phase in the Mg-Zn system occurs in the form of metastable β2′\beta_2^{\prime}-MgZn2_2 nanoplates during the aging of Mg-Zn alloys.

The experimental Mg-Zn phase diagram [17] and the Mg-Zn convex hull diagram (see Figure 3b) both indicate that the MgZn2_2-C14 structure is a stable phase. According to the established thermodynamic conditions for stable-phase precipitation, this MgZn2_2-C14 phase should form a tie-line with Mg. Instead, there are Mg21_{21}Zn25_{25} and Mg4_4Zn7_7 stable phases situated between MgZn2_2-C14 and Mg, preventing the precipitation of a stable MgZn2_2-C14 phase from Mg-Zn alloys. However, recent experiments [9, 18] commonly suggest the precipitation sequence in Mg-Zn alloys as supersaturated solid solution (S.S.S) →\rightarrow GP zones →\rightarrow β1′→β2′\beta_1^{\prime} \rightarrow\beta_2^{\prime} (MgZn2_2 nanoplates) →β\rightarrow\beta-Mg21_{21}Zn25_{25}, with β2′\beta_2^{\prime}-MgZn2_2 nanoplates clearly identified as metastable intermediates. This naturally raises the question of why the originally stable MgZn2_2 phase forms during the aging of Mg-Zn alloys and exists in a metastable form.

Atomic configurations, convex hull diagrams, and distribution statistics for MgZn$_2$ and Mg$_2$Nd phases

Figure 3. a. Atomic configurations of C14, C15, and C36-structured MgZn2_2 phase. b. DFT-calculated Mg-Zn convex hull diagram. c. The distribution statistics of EhullE_{\mathrm{hull}} values for C14, C15, and C36-based Mg(Zn, Mg)2_2 structures. d. Atomic configurations of C14, C15, and C36-structured Mg2_2Nd phase. e. DFT-calculated Mg-Nd convex hull diagram. f. The distribution statistics of EhullE_{\mathrm{hull}} values for C14, C15, and C36-based Mg2_2(Mg, Nd) structures.

In reality, nanosized precipitates often have non-stoichiometric compositions with excess matrix elements, as evidenced in systems like the Al2_2Ca phase in Mg-Al-Ca alloys[19], Al3_3Sc in Al-Sc alloys[20], and η\eta-MgZn2_2 phases in Al-Mg-Zn alloys[21, 22]. Our recent research[12] demonstrates that TCP precipitation in the Mg matrix follows a nonclassical nucleation mechanism, where hcp→\toTCP transitions mainly rely on the distribution of larger atoms in the nucleus rather than smaller ones. For example, during the formation of the MgZn2_2 phase (see Figure S5), an inhomogeneous nucleus with Mg occupying anti-sites (Mg in Zn positions) can still trigger hcp→\toTCP transitions. Given the nonclassical nucleation characteristics of TCP formations and the constraints of precipitation dynamics, realistic TCP nanoplates readily retain excess matrix Mg. In this context, the metastable β2\beta_2-MgZn2_2 nanoplates may also incorporate excess Mg, resulting in Mg(Zn, Mg)2_2 stoichiometry. To evaluate the thermodynamic feasibility of Mg(Zn, Mg)2_2 precipitation, we mixed varying amounts of Mg into the Zn positions within the three basic C14, C15, and C36 Laves structures (Figure 3a). The formation energies of these Mg(Zn, Mg)2\mathrm{Mg})_2 structures were plotted on the Mg-Zn convex hull diagram to assess their phase stability. The Method section in Supplementary Information details the generation of various Mg(Zn,Mg)2\mathrm{Mg}(\mathrm{Zn},\mathrm{Mg})_2 structures, and the distribution statistics of their EhullE_{\mathrm{hull}} values are shown in Figure 3c.

Schematic illustration of nonclassical nucleation behavior of MgZn2 nucleation in the Mg matrix

Figure S5. Schematic illustration of nonclassical nucleation behavior of MgZn2_2 nucleation in the Mg matrix, depicting nuclei with (a) non antisite Mg atoms, (b) one antisite Mg atom, and (c) two antisite Mg atoms. An 882-atom model was employed for DFT calculations.

fig:3b shows that the formation energies of all Mg(Zn,Mg)2\mathrm{Mg}(\mathrm{Zn},\mathrm{Mg})_2 structures lie above the convex hull line, indicating their metastable nature. EhullE_{\mathrm{hull}} is typically utilized as an indicator of phase stability to predict the formability of a material, where lower EhullE_{\mathrm{hull}} values suggest higher thermodynamic feasibility. Although the exact boundary distinguishing formable and non-formable materials remains undefined, a general threshold of Ehull≤0.03E_{\mathrm{hull}} \leq0.03 eV/atom is often considered a rule of thumb for the possible formation of metastable phases[16, 23]. In Figure 3c, a large proportion of Mg(Zn,Mg)2\mathrm{Mg}(\mathrm{Zn},\mathrm{Mg})_2 structures (where excess Mg atoms occupy Kagomé layers), including C14C14, C15C15, and C36C36 types, fall below this threshold. This distribution proves the thermodynamic feasibility of metastable Mg(Zn,Mg)2\mathrm{Mg}(\mathrm{Zn},\mathrm{Mg})_2 precipitation and predicts that C14C14, C15C15, and C36C36 structures could all present in β2′\beta'_2 nanoplates. Experimental observations[24] have reported β2′\beta'_2 nanoplates with both C14C14 and C15C15 structures. Therefore, we rectify that the stoichiometry of metastable β2′\beta'_2 nanoplates should be Mg(Zn,Mg)2\mathrm{Mg}(\mathrm{Zn},\mathrm{Mg})_2 rather than pure MgZn2\mathrm{MgZn}_2. Moreover, careful characterization of compositions is crucial for both stable and metastable nanoprecipitates, particularly in identifying the incorporation of matrix elements, although this is always an experimental challenge.

This study further extends to elucidate the thermodynamic precipitation mechanism of metastable Mg2Nd\mathrm{Mg}_2\mathrm{Nd} Laves nanoplates, which exhibit unique system-dependent formation behavior. A striking observation is that these metastable Mg2Nd\mathrm{Mg}_2\mathrm{Nd} exclusively appear in the Mg-Nd-Ca ternary system[25], while remaining absent in binary Mg-Nd alloys. In Mg-Nd binary alloys, the reported final equilibrium precipitate is Mg41Nd5\mathrm{Mg}_{41}\mathrm{Nd}_5[18, 26], consistent with the Mg-Nd convex hull diagram (see Figure 3e), where Mg41Nd5\mathrm{Mg}_{41}\mathrm{Nd}_5 is the stable phase forming a tie-line with Mg. Since the metastable Mg2Nd\mathrm{Mg}_2\mathrm{Nd} lies outside the Mg41Nd5\mathrm{Mg}_{41}\mathrm{Nd}_5-Mg region, its precipitation from Mg-Nd alloys is thermodynamically impossible. When further considering whether nonclassical nucleation behavior (with Mg occupying Nd positions) could facilitate the precipitation of metastable Mg2(Nd,Mg)\mathrm{Mg}_2(\mathrm{Nd},\mathrm{Mg}) phases, our evaluation of phase stability for C14C14, C15C15, and C36C36-derived Mg2(Nd,Mg)\mathrm{Mg}_2(\mathrm{Nd}, \mathrm{Mg}) structures (Figure 3e, Figure 3f) shows that most have positive formation energies and EhullE_{\mathrm{hull}} values well above 0.03 eV/atom. Thus, the precipitation of metastable Mg2(Nd,Mg)\mathrm{Mg}_2(\mathrm{Nd}, \mathrm{Mg}) nanoplates from Mg-Nd alloys is thermodynamically unfeasible.

In Mg-Nd-Ca ternary alloys, we found that the formation of Mg2(Ca,Nd,Mg)\mathrm{Mg}_2(\mathrm{Ca}, \mathrm{Nd}, \mathrm{Mg}) metastable structures by incorporating Nd and excess Mg into Mg2Ca\mathrm{Mg}_2\mathrm{Ca} nanoplates is thermodynamically feasible. The Mg-Nd-Ca convex hull diagram (Figure 4a) shows an Mg-Mg2Ca\mathrm{Mg}_2\mathrm{Ca}-Mg41Nd5\mathrm{Mg}_{41}\mathrm{Nd}_5 triangle on the Mg-rich side. When introducing Nd and Mg into Ca positions and generating C14C14, C15C15, and C36C36-based Mg2(Nd,Mg)\mathrm{Mg}_2(\mathrm{Nd}, \mathrm{Mg}) structures, where their compositions require to lie within the Mg-Mg2Ca\mathrm{Mg}_2\mathrm{Ca}-Mg41Nd5\mathrm{Mg}_{41}\mathrm{Nd}_5 triangle, the EhullE_{\mathrm{hull}} values of generated Mg2(Nd,Mg)\mathrm{Mg}_2(\mathrm{Nd}, \mathrm{Mg}) structures have a large amount of portion fall below 0.03 eV/atom. This demonstrates the thermodynamic feasibility of metastable Mg2(Nd,Mg)\mathrm{Mg}_2(\mathrm{Nd}, \mathrm{Mg}) precipitation.

Mg-Nd-Ca convex hull diagram and distribution statistics of Ehull values

Figure 4. a. DFT-calculated zero temperature Mg-Nd-Ca convex hull diagram. b. The distribution statistics of EhullE_{\mathrm{hull}} values for C14C14, C15C15, and C36C36-based Mg2(Ca,Mg,Nd)\mathrm{Mg}_2(\mathrm{Ca}, \mathrm{Mg}, \mathrm{Nd}) structures.

Thus, under the structural templating of the stable Mg2Ca\mathrm{Mg}_2\mathrm{Ca} phase, the formation of metastable TCP nanoplates with complex compositions is accessible. Experimentally, when using high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) to characterize nanoprecipitates, the high Z-contrast of Nd easily causes researchers to focus solely on Nd, overlooking lower Z-contrast elements like Ca and Mg. We once again highlight the critical need for accurate compositional determination of nanoprecipitates.

Although metastable TCP nanoplates in Mg alloys may offer temporary creepresistance improvements, stable TCP nanoplates are more desirable for developing Mg alloys with superior long-term high-temperature creep resistance. This study devises a two-step screening strategy (Figure 5) to identify Mg alloy systems capable of forming stable TCP nanoplates. By combining thermodynamic and kinetic conditions for TCP precipitation, this strategy utilizes DFT-calculated convex hull diagrams from the OQMD[14, 15] to search for stable TCP phases forming a tie-line with Mg, followed by evaluating the structural stability of corresponding MRM_{\mathrm{R}}-type XX-ordering to determine the kinetic feasibility of TCP formation. TCP phases in Mg alloys, including Laves (AB2AB_{2}-type) and Laves-like (AB5AB_{5}, AB3AB_{3}, and A2B7A_{2}B_{7}-types) phases, can be categorized into three groups: Mg-containing binary phases (e.g., Mg2_{2}Ca) in binary alloys, Mg-free binary (e.g., Al2_{2}Ca) and Mg-containing ternary phases (e.g., LaMg2_{2}Cu9_{9}) in ternary alloys. A total of 64 Mg-XX binary and 2022 Mg-XX-YY ternary convex hull diagrams from the OQMD were employed in this screening process.

A two-step screening strategy to identify effective alloying elements in Mg-X binary and Mg-X-Y ternary alloys that can form coherent stable TCP nanoplates

Figure 5. A two-step screening strategy to identify effective alloying elements in Mg-XX binary and Mg-XX-YY ternary alloys that can form coherent stable TCP nanoplates.

During the second-step screening, since the nonclassical nucleation nature in hcp→\rightarrowTCP transformations dictates that Mg-free TCP phase nucleation predominantly depends on the distribution of large atoms instead of small ones[27], only MRM_{\mathrm{R}}-type XX-ordering needs to be examined in Mg-XX-YY alloys (where XX denotes large atoms and YY small atoms), obviating the need for extensive examinations of MRM_{\mathrm{R}}-type XX-YY-orderings across 124 candidates. Given that hcp→TCP\mathrm{hcp}\to\mathrm{TCP} transitions involve a lattice transformation from triangular to kagomé structures (which changes the θ\theta value from 120° to 180°[27]), as shown in Figure 1, a larger change in θ\theta value within our model correlates with a higher tendency for TCP formation. Hence, we adopted θ\theta as a structural indicator to examine the structural stability of MRM_{\mathrm{R}}-type XX-orderings, setting a threshold of θ>150∘\theta>150^\circ based on the observed clustering distribution of θ\theta values across various Mg-XX combinations (see Figure S4). For this screening step, we used 162-atom models (Figure S1-Figure S1)) with different MRM_{\mathrm{R}}-type XX-ordering sizes for structural relaxations. Further details on DFT calculations and models are provided in the Supplementary Information.

Atomic structural models in the situations of Mg$_{\mathrm{small}}$-X and Mg$_{\mathrm{large}}$-X

Figure S1. a-b. Atomic structural models in the situations of Mgsmall_{\mathrm{small}}-X and Mglarge_{\mathrm{large}}-X. The Mg, large, and small atoms are marked with grey, dark-pink, and light-blue balls, respectively.

Variation of θ value of relaxed $M_R$-type $X$-ordering with calculated radius ratios

Figure S4. a. The variation of θ\theta value of relaxed MRM_R-type XX-ordering with calculated radius ratios (Relement−X/R(Mg,Ti,Zr)R_{\mathrm{element-X}}/R_{\mathrm{(Mg, Ti, Zr)}}) for (Mg, Ti, Zr)Small_{\mathrm{Small}}-X scenarios. A unit-cell structure of MRM_R-type hcp-ordering is shown in the inset of the panel, and the pink and grey balls represent the large and small atoms, respectively. b. The variation of θ\theta value of relaxed MRM_R-type XX-ordering with calculated radius ratios (R(Mg,Ti,Zr)/Relement−XR_{\mathrm{(Mg, Ti, Zr)}}/R_{\mathrm{element-X}}) for (Mg, Ti, Zr)Large_{\mathrm{Large}}-X scenarios. A large-sized structure of MRM_R-type XX-ordering is shown in the inset of the panel, and the grey and blue balls represent the large and small atoms, respectively.

No.Mg-XX binary systemsStructural typeNo.XXMg-XX-YY systemsStructural type
1Mg2_2Ca†C14C1421HoHoFe2_2C15C15
2Mg2_2YbC14C1422HoFe3_3AB3AB_3
3MgCo2_2C36C3623HoOs2_2C14C14
24HoRe2_2C14C14
25HoTc2_2C14C14
No.XXMg-XX-YY systemsStructural type1CaCaAl2_2†C15C15
2CeCeAl2_2C15C152CeAl2_2C15C15
3CsK2_2C15C153CsCsK2_2C15C15
4CsNa2_2C14C144
5DyAl2_2C15C155DyAl2_2C15C15
6DyOs2_2C14C146DyDyOs2_2C14C14
7DyRe2_2C14C147DyRe2_2C14C14
8DyTc2_2C14C148DyTc2_2C14C14
9ErAl2_2C15C159ErAl2_2C15C15
10ErCo2_2C15C1510ErCo2_2C15C15
11ErFe2_2C15C1511ErFe2_2C15C15
12ErErFe3_3AB3AB_312ErErFe3_3AB3AB_3
13ErOs2_2C14C1413ErOs2_2C14C14
14ErRe2_2C14C1414ErRe2_2C14C14
15ErTc2_2C14C1415ErTc2_2C14C14
24HoRe2_2C14C14
25HoTc2_2C14C14
26KKNa2_2C14C14
27LaLaAl2_2C15C15
28NdNdAl2_2C15C15
29PmPmAl2_2C15C15
30PrPrAl2_2C15C15
31SmSmAl2_2C15C15
32SmOs2_2C14C14
33TbTbAl2_2C15C15
34TbOs2_2C14C14
35Th2_2Fe7_7A2B7A_2B_7
36ThFe5_5AB5AB_5
37ThThOs2_2C14C14
38ThRe2_2C14C14
39ThTc2_2C14C14

Table 1. High-throughput screening results of stable TCP nanoplates in Mg-XX binary and Mg-XX-YY ternary alloys.

16EuEuAl2_2C1540YYAl2_2†C15
17EuGdAl2_2†C1541YYOs2_2C14
18GdGdOs2_2C1442YYRe2_2C14
19GdGdRe2_2C1443YYTc2_2C14
20HoHoAl2_2C1544YbYbAl2_2C15

Table 1 (PDF p. 10).

† represents the TCP phase that has been observed in experiments.

Table 1 summarizes the high-throughput screening results for stable TCP phases in Mg-XX and Mg-XX-Y alloys. In binary systems, Mg2_2Ca, Mg2_2Yb, and MgCo2_2 were identified as potential precipitates, with Mg2_2Ca nanoplates have been experimentally confirmed. For ternary Mg-XX-Y alloys, 44 TCP phases, including Al2_2Ca [5], Al2_2Y [28], and Al2_2Gd [29] (Laves precipitates widely observed in previous research), were found, demonstrating the reliability of our screening approach. As TCP nanoplates are known to enhance the strength and creep resistance of Mg alloys, the newly discovered TCP phases could expand the chemical space for alloy design. Table 1 further suggests that REAl2_2 Laves phases may widely precipitate in Mg-RE-Al series alloys. In these alloys[30], TCP nanoplates typically form at lower aging temperatures (~200°C), while at higher temperatures (~500°C), long-period stacking ordered (LPSO) phases become more common.

A distinct class of single-unit-cell-thick metastable TCP phases, termed γ′′\gamma^{\prime\prime} phases, are often observed in Mg-RE-Zn (Ag) alloys. Owing to their unique ability to maintain constant thickness during aging, these γ′′\gamma^{\prime\prime} phases have received widespread attention over the past decade[4, 10, 18]. Regarding their thermodynamic origins, Zhao et al.[31] recently suggest that γ′′\gamma^{\prime\prime} phases in Mg-Nd-Ag alloys act as metastable intermediates for the equilibrium non-TCP NdAgMg11_{11} phase, based on HAADF-STEM observations and atomic-scale energy-dispersive X-ray spectroscopy. This reveals a unique precipitation pathway where non-TCP equilibrium phases use TCP-structured intermediates to mediate their precipitation. Applying the two-step screening strategy, we extended our search to potential stable TCP phases in hcp-structured Ti and Zr alloys in the Supplementary Information. Our results indicate that only the TiCr2_2 and ZrMn2_2 Laves phases (see Table S1) may precipitate from Ti-Cr and Zr-Mn binary alloys. This is mainly because, under the formation of unstable MRM_{\mathrm{R}}-type hcp-ordering, all large elements in the Ti (Zr) matrix cannot interact with the surrounding matrix to induce hcp→\toTCP transitions (see Figure S4), unlike those in the soft Mg matrix. Meanwhile, we emphasize that these screening results offer only preliminary guidance for future experimental explorations, as many intricate factors, such as solid solubility of relevant adding elements, alloy composition, thermal processing history, etc., also significantly affect the precipitation of TCP plates.

Table S1. High-throughput screening results of the stable TCP nanoplates in Ti-X and Zr-X binary alloys.

In this study, using DFT calculations to construct convex hull diagrams and assess phase stability, we rectified that the metastable β2′\beta'_{2} nanoplates in Mg-Zn alloys are Mg(Mg,Zn)2\mathrm{Mg(Mg,Zn)}_{2} rather than pure MgZn2\mathrm{MgZn}_{2}. We also clarified that the stoichiometry of the experimentally observed metastable Mg2Nd\mathrm{Mg}_{2}\mathrm{Nd} phase in Mg-Nd-Ca alloys is Mg2(Nd,Mg,Ca)\mathrm{Mg}_{2}(\mathrm{Nd,Mg,Ca}) and explained why this phase does not form in binary Mg-Nd alloys. This highlights the importance of accurately determining the composition of nanoprecipitates in experiments. Furthermore, by integrating the thermodynamic and kinetic conditions for TCP precipitation, we developed a two-step high-throughput screening strategy to identify stable TCP nanoplates in Mg alloys. The newly discovered TCP phases may facilitate the design of high-strength and creep-resistant Mg alloys. These findings enhance our atomic-level understanding of precipitation behavior in various hcp-based alloys and lay a foundation for designing advanced materials incorporating coherent TCP nanoplates.

Acknowledgments

This research is supported by the National Key Research and Development Program of China (2023YFB3710902), Fundamental Research Funds for the Central Universities of China (N2102011, N2007011, N160208001), and National 111 Project (B20029).

Author contributions

G. Qin and J. Bai conceived the original idea and designed the work. J. Bai conducted the simulations with help from X. Pang. J. Bai wrote the paper. G. Qin supervised the project and revised the manuscript.

Declaration of Competing Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Code availability

The code in this study is available from the authors upon reasonable request.

Supplementary Information for Decoding Complex Compositions in Topologically Close-Packed Nanoplates of Magnesium Alloys: A High-Throughput Route to Stable Precipitates

Junyuan Bai; Xueyong Pang; Gaowu Qin

Corresponding author.

Email: [email protected] (X.Y. Pang) [email protected] (G. W. Qin)

Two-step screening strategy for identifying effective alloying elements in binary and ternary alloys

Figure S3. A two-step screening strategy to identify effective alloying elements in (Ti, Zr)-X binary and (Ti, Zr)-X-Y ternary alloys that can form coherent stable TCP nanoplates.

No.Ti-X binary systemsStructural typeNo.Zr-X binary systemsStructural type
1TiCr2_2C151ZrMn2_2C15

Table 1 (PDF p. 19).

Methods

DFT calculations were conducted using the Vienna ab initio simulation package (VASP)[1, 2] employing Blochl’s projector augmented wave (PAW) potential method[3]. The exchange-correlation energy functional was described with the generalized gradient approximation (GGA) as parameterized by Perdew-Burke-Ernzerhof (PBE)[4]. The frozen core pseudopotentials were used for RE elements, as these pseudopotentials have been shown to replicate the thermodynamics and elastic properties of rare-earth intermetallics [5, 6]. A 520 eV plane wave cutoff was adopted, with convergence criteria for energy and atomic force set as 10−610^{-6} eV and 10−210^{-2} eV/Å, respectively. Partial occupancies were determined using the first-order Methfessel-Paxton method with a smearing width of 0.2 eV[7]. Relaxations of atomic coordinates and optimizations of the shape and size of the model were adopted for all calculations. A Γ\Gamma-centered kk-point mesh of 18×18×1018 \times18 \times10 was adopted for the HCP primitive cell, with other supercells appropriately scaled to maintain a constant kk-point density. Large numbers of symmetrically distinct Mg(Zn, Mg)2_2, Mg2_2(Nd, Mg), and Mg2_2(Nd, Ca, Mg) structures based on C14, C15, and C36 types were generated with the Clusters Approach to Statistical Mechanics (CASM) code[8].

The unit-cell structure of MRM_\mathrm{R}-type hcp-ordering was constructed based on their aa-vector, bb-vector, and cc-vector parallel to the [101ˉ0]hcp[10\bar{1}0]_\mathrm{hcp}, [011ˉ0]hcp[01\bar{1}0]_\mathrm{hcp} and [0001]hcp[0001]_\mathrm{hcp} directions, respectively. During screening, we chose the θ\theta value as a structural indicator to determine whether there were hcp→\rightarrowTCP transitions. In Figure S2, we tested θ\theta variation across different model sizes for three types of matrices. The results indicate that while smaller models, like the 162-atom (3×3×33 \times3 \times3) model (Figure S1), amplify the θ\theta value for the Mg matrix scenario due to periodic interactions, our calculations (Figure S4) based on this 162-atom model still display a clustered data distribution that can effectively identify candidate elements from a large pool of elements. Hence, to balance computational accuracy and cost, we employed this 162-atom (3×3×33 \times3 \times3) model containing MRM_\mathrm{R}-type XX-ordering with one unit-cell size or larger-sized (i.e., clusters in a 3D environment) for high-throughput screenings of hcp-Mg, Ti, and Zr alloys.

Convergence tests of $\theta$ value for model size across different matrices

Figure S2. Convergence tests of θ\theta value for model size across different matrices.

Models containing MRM_{\mathrm{R}}-type XX-ordering with a single unit-cell size (Figure S1, MgSmall_{\mathrm{Small}}-X situations) were employed to examine Mg-free binary phases such as the Al2_2Ca phase, while the models containing the MRM_{\mathrm{R}}-type XX-ordering with larger size (Figure S1, MgLarge_{\mathrm{Large}}-X situations) were used to examine the Mg-containing phases where Mg acts as the large atoms, e.g., MgCo2_2 phases. The same rule was also applied to hcp-based Ti and Zr matrices. The calculated structural stability results of these MRM_{\mathrm{R}}-type XX-ordering across MgSmall_{\mathrm{Small}}-X and MgLarge_{\mathrm{Large}}-X situations within the Mg matrix, as well as in the matrices of hcp-based Ti, and Zr, are shown in Figure S4. The atomic radii of different elements RelementR_{\mathrm{element}} in various hcp-matrix were calculated by measuring the average nearest distance of neighboring matrix atoms in a 54-atom model.

The formation energy EfE_{\mathrm{f}} of a configuration AxByA_xB_y, for instance, was calculated relative to the zero kelvin total energies of pure elements A and B as follows:

Ef=E(AxBy)−NxEA−NyEBNx+Ny(1)E_{\mathrm{f}}=\frac{E(A_xB_y)-N_xE_{\mathrm{A}}-N_yE_{\mathrm{B}}}{N_x+N_y} \tag*{(1)}

Where E(AxBy)E(A_xB_y) denotes the total energies of a configuration, and EAE_{\mathrm{A}} and EBE_{\mathrm{B}} are the total energies per atom for pure elements AA and BB, respectively. NxN_x and NyN_y represent the number of elements AA and BB in the configuration. In this way, the formation energy of a configuration with ternary or more constituents can be obtained. The zero temperature convex hull diagrams were constructed by using the pymatgen code [9].

Supplementary References:

  1. G. Kresse, J. Hafner, Ab initio molecular dynamics for open-shell transition metals, Phys Rev B 48 (1993) 13115–13118.

  1. G. Kresse, J. Hafner, Ab initio molecular dynamics for liquid metals, Phys Rev B 47 (1993) 558–561.

  1. P.E. Blöchl, Projector augmented-wave method, Phys Rev B 50 (1994) 17953–17979.

  1. J.P. Perdew, K. Burke, M. Ernzerhof, Generalized gradient approximation made simple, Phys Rev Lett 77 (1996) 3865–3868.

  1. M.C. Gao, A.D. Rollett, M. Widom, Lattice stability of aluminum-rare earth binary systems: A first-principles approach, Phys Rev B Condens Matter Mater Phys 75 (2007).

  1. L. Huber, I. Elfimov, J. Rottler, M. Militzer, Ab initio calculations of rare-earth diffusion in magnesium, Phys Rev B Condens Matter Mater Phys 85 (2012) 1–7.

References

  1. [1]J.F. Nie, Physical Metallurgy of Light Alloys, in: Physical Metallurgy: Fifth Edition, Elsevier Inc., 2014: pp. 2009–2156.
  2. [2]J.F. Nie, X. Gao, S.M. Zhu, Enhanced age hardening response and creep resistance of Mg-Gd alloys containing Zn, Scr Mater 53 (2005) 1049–1053.DOI
  3. [3]X. Gao, S.M. Zhu, B.C. Muddle, J.F. Nie, Precipitation-hardened Mg-Ca-Zn alloys with superior creep resistance, Scr Mater 53 (2005) 1321–1326.DOI
  4. [4]J.F. Nie, K. Oh-ishi, X. Gao, K. Hono, Solute segregation and precipitation in a creep-resistant Mg-Gd-Zn alloy, Acta Mater 56 (2008) 6061–6076.DOI
  5. [5]M. Li, D. Xie, J. Li, H. Xie, Q. Huang, H. Pan, G. Qin, Realizing ultra-fine grains and ultra-high strength in conventionally extruded Mg-Ca-Al-Zn-Mn alloys: The multiple roles of nano-precipitations, Mater Charact 175 (2021) 111049.DOI
  6. [6]A. Suzuki, N.D. Saddock, J.W. Jones, T.M. Pollock, Solidification paths and eutectic intermetallic phases in Mg-Al-Ca ternary alloys, Acta Mater 53 (2005) 2823–2834.DOI
  7. [7]A. Suzuki, N.D. Saddock, J.R. TerBush, B.R. Powell, J.W. Jones, T.M. Pollock, Precipitation strengthening of a Mg-Al-Ca - Based AXJ530 die-cast alloy, Metall Mater Trans A Phys Metall Mater Sci 39 A (2008) 696–702.DOI
  8. [8]X. Gao, J.F. Nie, Characterization of strengthening precipitate phases in a Mg-Zn alloy, Scr Mater 56 (2007) 645–648.
  9. [9]J. Buha, Reduced temperature (22-100 °C) ageing of an Mg-Zn alloy, Materials Science and Engineering: A 492 (2008) 11–19.DOI
  10. [10]J. Bai, X. Pang, X. Meng, H. Xie, H. Pan, Y. Ren, M. Jiang, G. Qin, Anomalous crystal structure of γ″ phase in the Mg-RE-Zn(Ag) series alloys: Causality clarified by ab initio study, J Mater Sci Technol 36 (2020) 167–175.DOI
  11. [11]H. Xie, H. Pan, Y. Ren, S. Sun, L. Wang, H. Zhao, B. Liu, X. Qi, G. Qin, Magnesium alloys strengthened by nanosaucer precipitates with confined new topologically close-packed structure, Cryst Growth Des 18 (2018) 5866–5873.DOI
  12. [12]J. Bai, H. Xie, X. Pang, M. Jiang, G. Qin, Structural pathway for nucleation and growth of topologically close-packed phase from parent hexagonal crystal, Acta Mater 274 (2024) 119999.
  13. [13]C.J. Bartel, Review of computational approaches to predict the thermodynamic stability of inorganic solids, J Mater Sci 57 (2022) 10475–10498.DOI
  14. [14]J.E. Saal, S. Kirklin, M. Aykol, B. Meredig, C. Wolverton, Materials design and discovery with high-throughput density functional theory: The open quantum materials database (OQMD), JOM 65 (2013) 1501–1509.DOI
  15. [15]S. Kirklin, J.E. Saal, B. Meredig, A. Thompson, J.W. Doak, M. Aykol, S. Rühl, C. Wolverton, The Open Quantum Materials Database (OQMD): Assessing the accuracy of DFT formation energies, NPJ Comput Mater 1 (2015) 15010.DOI
  16. [16]C.J. Bartel, Review of computational approaches to predict the thermodynamic stability of inorganic solids, J Mater Sci 57 (2022) 10475–10498.DOI
  17. [17]H. Okamoto, Comment on Mg-Zn (Magnesium-Zinc), Journal of Phase Equilibria 15 (1994) 129–130.
  18. [18]J.F. Nie, Precipitation and hardening in magnesium alloys, Metall Mater Trans A Phys Metall Mater Sci 43 (2012) 3891–3939.
  19. [19]A. Suzuki, N.D. Saddock, J.W. Jones, T.M. Pollock, Structure and transition of eutectic (mg,al)2ca laves phase in a die-cast mg-al-ca base alloy, Scr Mater 51 (2004) 1005–1010.
  20. [20]A. Orthacker, G. Haberfehlner, J. Taendl, M.C. Poletti, B. Sonderegger, G. Kothleitner, Diffusion-defining atomic-scale spinodal decomposition within nanoprecipitates, Nat Mater 17 (2018) 1101–1107.DOI
  21. [21]J.Z. Liu, J.H. Chen, Z.R. Liu, C.L. Wu, Fine precipitation scenarios of AlZnMg(Cu) alloys revealed by advanced atomic-resolution electron microscopy study: Part II: Fine precipitation scenarios in AlZnMg(Cu) alloys, Mater Charact 99 (2015) 142–149.DOI
  22. [22]J.Z. Liu, J.H. Chen, D.W. Yuan, C.L. Wu, J. Zhu, Z.Y. Cheng, Fine precipitation scenarios of AlZnMg(Cu) alloys revealed by advanced atomic-resolution electron microscopy study Part I: Structure determination of the precipitates in AlZnMg(Cu) alloys, Mater Charact 99 (2015) 277–286.DOI
  23. [23]G.G.C. Peterson, J. Brgoch, Materials discovery through machine learning formation energy, J Phys Energy 3 (2021).
  24. [24]W. Lou, H. Xie, X. Zhao, J. Bai, H. Zhang, Y. Wang, X. Li, H. Pan, Y. Ren, G. Qin, Variable precipitation behaviors of Laves phases in an ultralight Mg-Li-Zn alloy, Journal of Magnesium and Alloys 11 (2023) 2018–2026.
  25. [25]W.R. Ren, Z.Q. Li, A.P. Zhang, H.W. Chen, Precipitation of Mg2Nd Laves phases in a Mg-Nd-Ca alloy, Vacuum 218 (2023).DOI
  26. [26]A.R. Natarajan, E.L.S. Solomon, B. Puchala, E.A. Marquis, A. Van Der Ven, On the early stages of precipitation in dilute Mg-Nd alloys, Acta Mater 108 (2016) 367–379.DOI
  27. [27]S.J. Kang, T.H. Kim, C.W. Yang, J.I. Lee, E.S. Park, T.W. Noh, M. Kim, Atomic structure and growth mechanism of T1 precipitate in Al-Cu-Li-Mg-Ag alloy, Scr Mater 109 (2015) 68–71.
  28. [28]Y. Chen, Q. Li, Y. Li, W. Zheng, J. Wang, X. Zeng, Phase equilibria of long-period stacking ordered phase in the ternary Mg-Y-Al alloys, J Mater Sci Technol 126 (2022) 80–92.DOI
  29. [29]X. Dong, L. Feng, S. Wang, G. Ji, A. Addad, H. Yang, E.A. Nyberg, S. Ji, On the exceptional creep resistance in a die-cast Gd-containing Mg alloy with Al addition, Acta Mater 232 (2022).DOI
  30. [30]H. Zhang, C.Q. Liu, Y.M. Zhu, H.W. Chen, L. Bourgeois, J.F. Nie, Revisiting building block ordering of long-period stacking ordered structures in Mg–Y–Al alloys, Acta Mater 152 (2018) 96–106.
  31. [31]X. Zhao, Z. Li, A. Zhang, L. Hao, H. Chen, J.F. Nie, Atomic-scale investigation of precipitate phases in QE22 Mg alloy, J Mater Sci Technol 177 (2024) 114–127.DOI

Paper details

vlnc:W_gdndye84
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents