Movatterモバイル変換

[0]ホーム
URL:

Skip to main content
NCBI home page
Search in PMCSearch
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health.
Learn more:PMC Disclaimer | PMC Copyright Notice
Scientific Reports logo

Interfacial structures and energetics of the strengthening precipitate phase in creep-resistant Mg-Nd-based alloys

D Choudhuri1,a,R Banerjee1,b,S G Srinivasan1,c
1Department of Materials Science and Engineering, University of North Texas, Denton, TX, 76191, USA
a

Email:deep.choudhuri@gmail.com

b

Email:Raj.Banerjee@unt.edu

c

Email:Srinivasan.Srivilliputhur@unt.edu

Received 2016 Aug 23; Accepted 2016 Dec 7; Collection date 2017.

Copyright © 2017, The Author(s)

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visithttp://creativecommons.org/licenses/by/4.0/

PMCID: PMC5240141  PMID:28094302

Abstract

The extraordinary creep-resistance of Mg-Nd-based alloys can be correlated to the formation ofnanoscale-platelets of β1-Mg3Nd precipitates, that grow along 〈11Inline graphic0〉Mg in bulkhcp-Mg and on dislocation lines. The growth kinetics of β1 is sluggish even at high temperatures, and presumably occurs via vacancy migration. However, the rationale for the high-temperature stability of precipitate-matrix interfaces and observed growth direction is unknown, and may likely be related to the interfacial structure and excess energy. Therefore, we study two interfaces– {112}β1/{Inline graphic100}Mg and {111}β1/{11Inline graphic0}Mg– that are commensurate with β1/hcp-Mg orientation relationship via first principles calculations. We find that β1 acquires plate-like morphology to reduce small lattice strain via the formation of energetically favorable {112}β1/{Inline graphic100}Mg interfaces, and predict that β1 grows along 〈11Inline graphic0〉Mg on dislocation lines due to the migration of metastable {111}β1/{11Inline graphic0}Mg. Furthermore, electronic charge distribution of the two interfaces studied here indicated that interfacial-energy of coherent precipitates is sensitive to the population of distorted lattice sites, and their spatial extent in the vicinity of interfaces. Our results have implications for alloy design as they suggest that formation of β1-like precipitates in thehcp-Mg matrix will require well-bonded coherent interface along precipitate broad-faces, while simultaneously destabilizing other interfaces.

Magnesium (Mg) alloys have tremendous potential as structural materials for automotive applications in engine block, transmission parts due to their lightweight. The concomitant fuel savings is however, thwarted by their unfavorable creep strength1,2,3,4. Mg-rare earth (RE) alloys are exceptions with demonstrated excellent creep resistance1,2,3. Their superior creep properties are correlated to the formation of high volume fractions of strengthening precipitates based on RE intermetallic compounds4,5. However, the scarcity and high cost of RE elements limits the development of creep–resistant Mg-RE alloys. This work is motivated by gaining an atomistic understanding of commonly found strengthening precipitates in creep-resistant Mg-RE alloys, and inform Mg-alloy design approaches by providing energetics and bonding characteristics associated with their presence within thehcp-Mg matrix. Therefore, we have investigated the structure of interface between intermetallic precipitate phase β1 (an ordered cubic structure) andhcp-Mg. This precipitate typically forms in Mg-Nd-based alloys like Mg-Nd, Mg-Nd-Y and Mg-Nd-Y-Zr, e.g. commercial WE43 and WE54 5,6,7. The coherent β1 phase has a plate-like morphology, and shares an (0001)Mg//(011)β1, [Inline graphic100]Mg//[21Inline graphic]β1 and [11Inline graphic0]Mg//[1Inline graphic1]β1 orientation relationship (OR) withhcp-Mg5,6,7,8,9,10,11,12. TEM results also indicate that plate-like β1 precipitates always forms with a high aspect ratio, and its broad-face {21Inline graphic}β1 parallel to {Inline graphic100}Mg 4,5,6,7,8,9,10,11. Furthermore, β1 nucleating within the Mg-matrix is associated with the orthorhombic β′ precipitate phase5,7, or as self-accommodating β1 triads5,11, and heterogeneously along dislocation lines via accommodation of β1 stress-free transformation strains associated with β1 formation7,8,9,13,14. Heterogeneous formation is particularly useful in restraining high-temperature creep deformation, because its dynamic nucleation arrests dislocation movement and enhances the creep-resistance of Mg-RE alloys4,9,10,11.

The formation tendencies of β1 precipitates have been characterized experimentally5,6,7,8,9,10,11, and its evolution within Mg matrix and on dislocations studied via phase-field models12,13,14. Despite extensive studies on β1 phase, basic thermodynamic quantities like β1/Mg interfacial energies are yet to be determined unequivocally. Quantitative understanding of β1/Mg interfaces is crucial for three reasons. First, atomic resolution transmission electron microscopy (TEM) studies have shown that the{112}β1/{Inline graphic100}Mg interfaces form both longer and shorter edges of β1 plates11. This is surprising. One would have expected at least an additional {111}β1/{11Inline graphic0}Mg interface, because both {112}β1 and {111}β1 planes take part in forming thehcp-Mg/β1 orientation relationship. Second, a well-developed β1 plate shares a substantial interfacial-area with the parent Mg-matrix. Thus the interfacial energy, in conjunction with transformation strain energy, will likely determine the plate-like morphology12,13,14,15. Therefore, determination ofhcp-Mg/β1 interfacial energies will provide valuable inputs to refine phase field models12,13,14,15. Third, this study allows us to rationalize whether these coherent interfaces have lower excess energies16.

Density functional theory (DFT)-based first principles calculations have been used in the past to determine interfacial energies of coherent precipitates in Mg-RE alloys14,15,17. So far, DFT studies on Mg-Nd have largely examined the stability of the β″ (orderedhcp structure5) and the β′ precipitate phases15,17; both precipitate phases form at early stages prior to β1 precipitation. For example, Issaet al. reported that β″ have positive and negative interfacial energies, when it joins with basal {0001}Mg and prismatic {Inline graphic100}Mg planes respectively15,17. Based on these interfacial energy calculations they proposed that β″ is unstable in Mg-Nd17, which was later corroborated with TEM observations18. The good agreement with experiments validates the available RE pseudopotentials with frozen-core f-electrons, and permits us to study other systems. However, the methodology used for determining negative interfacial energies needs further investigation because one intuitively associates an interface with positive excess energy. Therefore, the rationale for predicted negative interfacial energy was also investigated from the perspective of β1/Mg interfaces, especially since β1 is a key-strengthening phase in many commercial alloys17.

Previous DFT studies on Mg-Nd have not investigated β1 formation, its interfacial structure, and the associated bonding environment14,15,17. Furthermore, they did not examine thermodynamic quantities like surface excess energies - a difficult quantity to measure experimentally in the larger system limit14,15,17. Thus, guided by TEM studies5,6,7,8,9,10,11, we have investigated the energetics and structure of two types of β1/Mg interfaces, namely {112}β1/{Inline graphic100}Mg and {111}β1/{11Inline graphic0}Mg, via DFT calculations.

Computational Methodology

First principles computations were carried out with the Vienna Ab-initio Simulation Package (VASP) employing the projector-augmented plane-wave (PAW) method18, with electron exchange and correlation described by the generalized gradient approximation (GGA)19,20. All calculations were performed at 0 K, with a cut-off energy of 360 eV, k-point spacing of 0.1 Å−1, 0.2 eV Methfessel-Paxton smearing width and Brillouin zone integration at 10−7 eV convergence threshold. Groundstate structures of supercells were obtained by relaxing ionic, shape, and volumetric degrees of freedom till global energy convergence was achieved to within 1 meV/atom, and the Hellmann-Feynman forces on the atoms were less than 1 meV/Å. Visualization and subsequent analysis of the relaxed structures was performed using Vesta21 and Ovito22.

Results

Construction of initial supercells

Figure 1a shows β1 Mg3Nd with a BiF3 structure, lattice parameter 7.391 Å, space group FmInline graphicm (#226, similar tofcc structures), and Pearson symbol cF16 (corresponding tobcc-ordered D03strukturbericht)23. We constructed an orthorhombic β1 supercell whose edges are parallel to 〈011〉β1, 〈111〉β1 and 〈112〉β1. This choice corresponds to β1 -hcp-Mg orientation relationship seen in experiments5,6,7,8,9,10,11. Two orthogonal supercells containing 96 atoms, and lengths 35 (Fig. 1c) and 50 Å (Fig. 1d) were used in our analysis. In the resulting supercells (shown inFig. 1c and d) 12 Nd atoms were replaced with Mg in one portion to create a “Mg-slab”, while retaining enough Nd atoms in the remainder to maintain Mg3Nd stoichiometry corresponding to β1. The resulting structures created {112}β1 (hereafter referred to as “{112}” supercell) and {111}β1 interfaces between Mg and the β1 precipitate as shown inFig. 1c and d respectively. Note that in these initial supercells, Mg matrix has a “bcc” structure.

Figure 1.

Figure 1

Crystal structures of (a) cubic β1 and (b) β1 in an orthorhombic cell. Pre-relaxed supercells extended along: (c) 〈112〉β1 and without any prior vacancies, and (d) 〈111〉β1 and showing the locations of vacancy placed within the “bcc” Mg-region.

The Mg-slab was further modified based on the β1/Mg interfacial structure proposed by Nie and Muddle5. They suggested that β1 formation is associated with vacancies near the β1/Mg interface along 〈111〉β1 5. Concomitantly, recent analytical calculations, involving stress-free transformations strains during β1 formation11,13, indicated the presence of a tensile strain along 〈111〉β1. Such tensile strains likely facilitate vacancy creation along 〈111〉β1 by expanding the volume near the interfaces. Therefore, to understand such effects a Mg vacancy was introduced near the β1/Mg interface (square,Fig. 1d) and at the center of Mg (diamond,Fig. 1d) in separate supercells. An Mg vacancy is placed inside “Mg” slab, rather than in the intermetallic β1 compound, was guided by β1/Mg interfacial structure proposed in literature6,11. Also, this reasonable because Mg vacancy formation energies in β1 (1.24 eV) is greater than inhcp-Mg (0.82 eV). The resulting vacancy concentration was ~1 at% for the entire supercell, and 2at% for the Mg slab. Hereon the vacancy containing supercells - near β1/Mg interface, and center of Mg slab - will be referred to as “{111}β1(Iv)” and “{111}β1(Cv)” supercells respectively.

Relaxed structures

The relaxed structures of three supercells with β1/Mg interfaces are shown inFig. 2. In the {112}, {111}β1(Iv) and {111}β1(Cv) supercells (Figs 1a and2b,c respectively), the Mg slab with an initialbcc structure transforms to the correcthcp crystal structure indicated by hexagons inFig. 2. The β1 however, retains itsbcc-ordered structure as confirmed by common neighbor analysis (CNA)22,24. In this manner we created “equilibrium” supercells comprising of two types of β1/hcp-Mg interfaces: {112}β1//{Inline graphic100}Mg and {111}β1(Iv/Cv)//{11Inline graphic0}Mg. The structural relaxation of {111}β1(Iv) and {111}β1(Cv) supercells (Fig. 2b and c), with an added vacancy (Fig. 1d), also created a region with vacancy-like excess free volume marked by dotted polygons inFig. 2b and c. This occurs even when a single vacancy was initially placed at the center of the Mg slab (Fig. 1d). The excess volume ofhcp-Mg inFig. 2b and c measured by Voronoi tessellation21, was ~30 Å3, and ~23 Å3. In comparison, Mg slabs without prior vacancies inside the {111} supercell, did not have an excess volume regions and neither did their initialbcc Mg relax to ahcp structure (Supplementary Figure S1) – unlike those presented inFig. 2b and c. Thus vacancies aid in the transformation ofbcc-Mg tohcp.

Figure 2. Relaxed supercells showing β1/Mg interfaces.

Figure 2

(a) {112}β1/{Inline graphic100}Mg interface, (b) {111}β1(Iv)/{11Inline graphic0}Mg interface, with a vacancy that was initially placed near the interface, and (c) {111}β1(Cv)/{11Inline graphic0}Mg interface with a vacancy was initially placed at the center of Mg-region.

In the relaxed {112} supercell,Fig. 2a, certain crystallographic orientations of Mg and β1 were mutually coincident and exhibited the following orientation relationship (OR): (0001)Mg//(011)β1, [Inline graphic100]Mg//[21Inline graphic]β1, and [11Inline graphic0]Mg//[1Inline graphic1]β1. Electron diffraction patterns from literature have indicated that the β1-Mg OR causes a ~5.26° misorientation between (0001)Mg and (011)β1 planes5,6,7,8,9,10,11, which is in excellent agreement with DFT value of 5.3° indicated inFig. 2a, and, also validates our approach. The misorientation can be quantified as the angle between [1Inline graphic00]β1 and [Inline graphic110]Mg.

A comparison of β1 and Mg crystallographic directions within {111}β1(Iv) and {111}β1(Cv) supercells showed that, eventhough (0001)Mg and (011)β1 planes are parallel, the other crystallographic axes involved in β1-Mg OR are not coincident. For example, angle between [Inline graphic100]Mg and [21Inline graphic]β1 in {111}Iv and {111}Cv is ~15°, while those axes are coincident in case of the {112} supercell. In other words, DFT calculations suggest that a creation of only {111}β1(Iv/Cv)/{11Inline graphic0}Mg interfaces may not yield the experimentally seen β1-Mg OR atleast near the interfacial regions. This result is surprising because, based on the β1-Mg orientation relationship, one would have expected a mutual alignment of β1 and Mg orientations in the {111}β1(Iv/C/v) supercells.

Implication of this “off β1-Mg OR” misalignment in the {111}β1(Iv/C/v) supercells was further probed by comparing it’s local atomic distortions with the {112} supercells. Mg-Mg, Mg-Nd and Nd-Nd bond-lengths measured from regions 3–4 atomic layers away from the interfaces were used to calculate the lattice strainInline graphic, whereaβ1/Mg-interface andabulk are the lattice parameters of each phase in the supercells, and their bulk equilibrium lattice respectively.Table 1 lists the β1 and Mg lattice parameters, and the lattice strains. Lattice strains, 5–10 times larger that the bulk systems, were found in {111}β1(Iv) and {111}β1(Cv) supercells along the a-axis of Mg or 〈11Inline graphic0〉 and 〈100〉 of β1. We further note that placement of vacancy away from {111}β1(Iv)/{11Inline graphic0}Mg interface (Fig. 1d) increases the β1 lattice strain by ~33% in {111}β1(Cv) in comparison to {111}β1(Iv). In other words, higher lattice strains are associated with the formation of {111}β1/{11Inline graphic0}Mg interfaces than {112}β1/{01Inline graphic0}Mg. Regardless, presence of vacancy-like excess volume in the {111}β1(Iv/Cv) supercells strains the Mg and β1 lattice, and will be shown to affect the β1/Mg interfacial energies.

Table 1. Table of lattice parameters and strain in Mg and β1 slabs of the relaxed structures.

InterfaceMg matrixβ1 precipitate a (Å)/strain (%)
a (Å)/strain (%)c(Å)/strain (%)
{211}β1/{0110}Mg – No vacancy3.186/−0.285.206/+0.517.385/− 0.37
{111}β1/{1120}Mg – Vacancy initially at the interface3.282/+2.735.202/+0.437.282/−1.76
{111}β1/{1120}Mg – Vacancy initially inside Mg3.276/+2.535.220/+0.787.239/−2.34

Calculated lattice parameters of bulk phases: Mg − 3.195 Å and 5.180 Å and β1 − 7.412 Å.

β1/Mg interfacial energies

The geometry used for calculating β1/Mg interfacial energies (γβ1/Mg) is schematically shown inFig. 3a as vacuum-space/bulk-β1/bulk-Mg/vacuum-space. Here the interacting β1/Mg surfaces are {111}β1(Iv/Cv)/{11Inline graphic0}Mg and {112}β1/{01Inline graphic0}Mg (seeSupplementary Figure 2). This methodology was previously used by Liuet al. to examine the interface between an ordered intermetallic compound and the disordered matrix25. This approach further avoids the development of excess artificial interfacial strains by having vacuum on both sides of the unconstrained interfacial supercell. Thus, one need not include additional coherency strain energy terms (e.g. ref.17) to evaluate the interfacial energies. Furthermore, since β1 and Mg are exposed to the vacuum on either ends (Fig. 3a), we are required to determine the “free” surface of the participating {111}β1, {11Inline graphic0}Mg, {112}β1 and {Inline graphic100}Mg planes. The free surface energy, γS-p, for a phase “p” was calculated using the expressionInline graphic, where Ep(N) is the calculated ground state energy for a “N” atomic layer supercell slab of phase “p”, EBulk-P is the bulk energy of phase “p”, and A is the surface area25,26,27. The calculations involved a maximum of 24 (~35 Å) and 60 (~50 Å) layers for Mg and β1 respectively, and a minimum of ~12 Å of vacuum space. EBulk-P was determined from the slope of Ep(N) vs N plot. For pure metals, EBulk-P corresponds to their cohesive energy27, and our calculations yielded 1.53 eV/atom for Mg - in excellent agreement with literature26,27,28.Figure 3b shows that plot of γS-p vs. N for different β1 and Mg planes, and the results indicated that the γS-p values, for {111}β1(Iv/Cv), {11Inline graphic0}Mg, {112}β1 and {Inline graphic100}Mg planes rapidly converge within the chosen range of N.

Figure 3.

Figure 3

(a) Configuration used for calculating β1/hcp-Mg interfacial energy. (b) Free surface energies (γS-p) of different β1 and Mg surface planes plotted as a function of number of atomic layers (N). (c) γS-p vs 1/N plots to determine γS-p in the larger size limit. The dotted line fits corresponding to R2 values of 0.94. The intercept of the linear fits, 1/N → 0 yield excess surface energy in the large system limit (N → ∞), which is a well-defined thermodynamic quantity. (d) Bar chart showing the interfacial energies of threeβ1/Mg interfaces reveal {112}β1/{Inline graphic100}Mg interfaces are energetically more stable than {111}β1/{11Inline graphic0}Mg.

The γS-p values for the Mg and β1 planes were calculated using linear fits (R2 ≈ 0.94) to γS-p vs. 1/N plots (Fig. 3c). We then extrapolated the fits for 1/N → 0, which correspond to N → ∞ layers or the “bulk”. Our surface energy γS-p of (0001)Mg was 0.52 J/m2, agreed well with literature reports of 0.52 and 0.55 J/m2 26,27.Figure 3b shows that β1 has higher excess surface energies (γS-p) than pure Mg. This indicates a stronger bonding within β1, and that it is difficult to cleave or fracture β1 during deformation. We further note that such surface excess energies have not been reported for any other precipitate phases in Mg-RE systems14,15,16.

The plots inFig. 3b and c helps estimate the optimal value of “N” required for the interfacial energy calculations that reasonably accounts for the bulk phases along with the free-surface and interfacial energies. Thus, 12 (~15 Å) and 24 (~25 Å) layers was chosen for each phase making up the {112}β1/{Inline graphic100}Mg and {111}β1/{11Inline graphic0}Mg interfaces respectively, and ~12 Å of vacuum was maintained above β1 and Mg slabs (Fig. 3a). β1/Mg interfacial energy (Inline graphic) was estimated using the expression22:

graphic file with name srep40540-m45.jpg

where γS-β1 and γS-Mg are the free surface energies of β1 and Mg respectively, Esupercell is the ground state energy of the β1-Mg supercell while EBulk-β1 and EBulk-Mg are the bulk energies of β1 and Mg respectively, and A is the surface area. The γβ1/Mg values of these three interfaces are plotted inFig. 3d as a bar chart. Their respective relaxed structures are also shown inSupplementary Figure 2. The calculated γβ1/Mg values for {112}β1, {111}β1(Cv) and {111}β1(Iv) supercells were 98, 282 and 762 mJ/m2 respectively. Thus, the {112}β1/{Inline graphic100}Mg interfaces are more energetically stable than {111}β1/{11Inline graphic0}Mg. Also note that inclusion of excess surfaces energies for estimating interfacial energies (equation 1) will result in only positive excess energy values, which is thermodynamically reasonable.

Structure and bonding environment near β1/Mg interfaces

To explain the lower interfacial energy of {112}β1/{Inline graphic100}Mg inFig. 3d, we have compared the bonding environment of these interfaces. The analysis correlates the interfacial atomic-coordination (Fig. 4a and c), with electron charge density distribution (Fig. 4b and d).

Figure 4.

Figure 4

(a,b) and (c,d) show structure and electron charge distributions of {112}β1/{Inline graphic100}Mg and {111}β1(Iv)/{11Inline graphic0}Mg interfaces respectively. In (a) and (b) red and blue colors indicate atomic sites inhcp-Mg and β1 structures respectively, while green indicates “distorted” lattice sites. (e) shows the legend for electron charge densities in (b) and (d). (e) shows the structure of{111}β1(Cv)/{11Inline graphic0}Mg interface. Interfacial width is indicated by translucent boxes, and the contours in the electron charge density plots engulf regions with lower charge densities compared to the bulk.

InFig. 4a and c the atom colors are assigned by CNA environment type18,20 and the bonds depict in-plane – i.e. (0001)Mg//(011)β1 – coordination around an atomic site: (i) red denoteshcp-Mg with 6 in-plane coordination, (ii) blue denotesbcc-ordered β1 with 4 in-plane coordination, and (iii) green depicts sites which were not identified as eitherhcp-Mg or β1. Thus, the green-colored bonds/atoms near the interfacial regions - indicated with shaded boxes acrossFig. 4a–d - can be interpreted as distortion of the lattice sites, from bulkhcp-Mg and β1 structures.

A comparison of the interfacial regions, and their corresponding bonding environment reveal that the distorted lattice sites (shown by green colored bonds) are invariably located around regions with significant depletion of electron charge density distributions (marked by red arrows inFig. 4b and d). In other words, interfacial lattice distortions correlate well with lower electron density or weak bonding29. Furthermore, the electron density distributions near the interfacial regions of {111}β1(Iv)/{11Inline graphic0}Mg was discernibly lower (compareFig. 4b and c with legend at the bottom), and spread over a wider region than {112}β1/{Inline graphic100}Mg. Correspondingly, lattice significantly distorted near the {111}β1(Iv)/{11Inline graphic0}Mg interface. Therefore, considering the larger differences in the energies of {112}β1/{Inline graphic100}Mg (98 mJ/m2) and {111}β1(Iv)/{11Inline graphic0}Mg (282 mJ/m2) interfaces, it is evident that smaller population of distorted lattice sites and narrow spatial extent near the interface correlate with lower interfacial energy. This is further confirmed for {111}β1(Cv)/{11Inline graphic0}Mg system (762 mJ/m2) inFig. 4e, which displayed a wider spatial extent of distorted interfacial lattice sites (overlapping bothhcp-Mg and β1 structures) than the other two interfaces.

Discussion

One of our motivations was to rationalize the existence of only {112}β1/{Inline graphic100}Mg interfaces seen in Mg-Nd alloys, which was contrary to the exceptions based on β1/Mg OR11. The DFT calculations revealed that, unlike {112}β1/{Inline graphic100}Mg, formation of {111}β1/{11Inline graphic0}Mg requires the presence of vacancy-like excess volume near the interfaces, which causes interfacial distortions, higher lattice strains and presumably strain energy within the mating phases. Such pronounced distortion/strain increases the interatomic spacings near the interface, reduces charge density between the atoms, and creates weaker interfacial bonding in {111}β1/{11Inline graphic0}Mg. The relatively poor bonding increases the excess energy of {111}β1/{11Inline graphic0}Mg incomparison to the {112}β1/{Inline graphic100}Mg interface, and makes the latter an energetically preferable interface.”

Despite the fact that excess volume results in an energetically unstable {111}β1/{11Inline graphic0}Mg interface. However, the presence of such excess volume near the interfaces may explain the growth of plate-like β1 within Mg-matrix and along dislocation lines5,6,7,8,9,10,11. Nie and Muddle have proposed that shear transformation strains are associated with β1 formation, because of which excess vacancy concentration is created at the short ends oriented along 〈11Inline graphic0〉Mg. It is likely β1 growth can occur via vacancy migration to the interfaces. However, to maintain elongated plate-like morphology during the growth process, β1 needs a stable interface in its long edge. Our DFT results suggest that a low lattice strain and interfacial energy of {112}β1/{100}Mg interfaces allows β1 precipitates to acquire a plate-like morphology with a high aspect ratio and possess a broad {112}β1-face parallel to {1Inline graphic00}Mg as seen in experiments4,5,6,7,8,9,10,11,12. In the case of dislocation assisted β1 form as linear chains along 〈11Inline graphic0〉Mg 10,13,30, which are also observed in our creep-tested microstructures10. We postulate that such linear growth of β1 along dislocation lines occur via transfer of vacancies to metastable interfaces like {111}β1(Iv)/{11Inline graphic0}Mg (Fig. 2b), which tends to have lower interfacial energy than vacancy formation away form the β1/Mg interface (i.e. {111}β1(Cv)/{11Inline graphic0}Mg inFigs 2c and3d). Since dislocations can rapidly supply vacancies via dislocation pipe diffusion, the diffusion assisted movement of {111}β1(Iv)/{11Inline graphic0}Mg type interfaces rapidly consumes a large portion of the host dislocation line, and orients β1 along 〈11Inline graphic0〉Mg. Finally, since {111}β1(Iv)/{11Inline graphic0}Mg interfaces are energetically unfavorable, interfacial atomic rearrangements, e.g. via shear process5,6, produces {112}β1/{Inline graphic100}Mg interfaces in the final microstructure.

From the perspective of alloy design approaches; the contribution of this report is two fold. We demonstrate that in order to develop alloys with advantages associated with “β1-like” precipitation one needs to (i) retain a well bonded coherent interface at the precipitate broad faces, e.g. {112}β1/{Inline graphic100}Mg, while simultaneously (ii) form mobile yet unstable interfaces on the short edges to allow precipitate growth along dislocation lines. However, finding non-RE substitutions, which fits these two criteria, is our next challenge, and is currently being tackled by coupling evolutionary algorithms with first principle calculations.

Conclusion

Formation of nanoscale β1-Mg3Nd precipitates inhcp-Mg matrix is known to enhance the creep resistance of Mg-alloys because of their tendency to form on dislocation lines and sluggish coarsening kinetics when they form in thehcp alloy matrix. Using first principles calculations, we have compared the excess energy and the structure of {112}β1/{Inline graphic100}Mg and {111}β1/{11Inline graphic0}Mg interfaces. Our calculations revealed that the formation of {112}β1/{Inline graphic100}Mg interfaces, compared to {111}β1/{11Inline graphic0}Mg interface, is associated with significant reduction in both the interfacial energy and lattice strains in the adjacent β1 and Mg matrix. The favorable formation energetics of the {112}β1/{Inline graphic100}Mg interface, in conjunction with small lattice strains, influences β1 acquiring a plate-like morphology with its broad-face {112}β1 parallel to {Inline graphic100}Mg. Electronic structure of these interfaces revealed that a lower interfacial energy also correlates with a smaller population of distorted lattice sites near the interfacial regions. Additionally, our DFT investigation informs that creep-resistant Mg alloys will benefit from “β1-like” precipitation because of two key features (i) retain a well bonded coherent interface at the precipitate broad faces, e.g. {112}β1/{Inline graphic100}Mg, and (ii) destabilize other interfaces (corresponding to precipitate-Mg OR) to promote growth along a desired direction, e.g 〈11Inline graphic0〉Mg.

Additional Information

How to cite this article: Choudhuri, D.et al. Interfacial structures and energetics of the strengthening precipitate phase in creep-resistant Mg-Nd-based alloys.Sci. Rep.7, 40540; doi: 10.1038/srep40540 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Material

Supplementary Materials
srep40540-s1.pdf (1.2MB, pdf)

Acknowledgments

The authors used UNT’s Talon2 High Performance Computing cluster, and Texas Advance Computing Center’s Stampede. D.C., R.B. and S.G.S. also acknowledge support from grant# NSF-DMR-1435611.

Footnotes

Author Contributions D.C., S.G.S., and R.B. conceptualized this work. D.C. ran the simulations. All authors contributed in writing this manuscript.

References

  1. Luo A. & Pekguleryuz M. O.Cast magnesium alloys for elevated temperature applications. J. Mater. Sci.29, 5259–5271 (1994). [Google Scholar]
  2. Mordike B. L.Creep-resistant magnesium alloys. Mater. Sci. Eng A. A324, 103–112 (2002). [Google Scholar]
  3. Vagarli S. S. & Langdon T. G.Deformation mechanism in H.C.P. metals at elevated temperatures – I. Creep behavior of magnesium. Acta Metall. 29, 1969–1982 (1981). [Google Scholar]
  4. Vagarli S. S. & Langdon T. G.Deformation mechanism in H.C.P. metals at elevated temperatures – II. Creep behavior of Mg-0.8%Al solid solution alloy. Acta Metall. 30, 1157–1170 (1982). [Google Scholar]
  5. Nie J. F.Precipitation and hardening in magnesium alloys. Metall. Mater. Trans. A.43, 3891–3939 (2012). [Google Scholar]
  6. Nie J. F. & Muddle B. C.Characterization of strengthening precipitate phases in a Mg–Y–Nd alloy. Acta. Mater.48, 1691–1703 (2000). [Google Scholar]
  7. Saito K. & Kenji H.The structures of precipitates in an Mg-0.5 at% Nd age-hardened alloy studied by HAADF-STEM technique. Mater. Trans.52, 1860–1867 (2011). [Google Scholar]
  8. Choudhuri D.et al. Evolution of a honeycomb network of precipitates in a hot-rolled commercial Mg–Y–Nd–Zr alloy. Phil. Mag. Lett.93, 395–404 (2013). [Google Scholar]
  9. Zhu S. M., Gibson M. A., Easton M. A. & Nie J. F.The relationship between microstructure and creep resistance in die-cast magnesium–rare earth alloys. Scripta Mater.63, 698–703 (2010). [Google Scholar]
  10. Choudhuri D.et al. Homogeneous and heterogeneous precipitation mechanisms in a binary Mg–Nd alloy. J. Mater. Sci.49, 6986–7003 (2014). [Google Scholar]
  11. Zhou X., Weyland M. & Nie J. F.On the strain accommodation of β1 precipitates in magnesium alloy WE54. Acta Mater.75, 122–133 (2014). [Google Scholar]
  12. Gao Y.et al. Simulation study of precipitation in an Mg–Y–Nd alloy. Acta Mater.60, 4819–4832 (2012). [Google Scholar]
  13. Liu H.et al. A simulation study of β1 precipitation on dislocations in an Mg–rare earth alloy. Acta Mater.77, 133–150 (2014). [Google Scholar]
  14. Liu H., et al. A simulation study of the shape of β′ precipitates in Mg–Y and Mg–Gd alloys. Acta Mater.61, 453–4662 (2013). [Google Scholar]
  15. Ji Y. Z.et al. Predictingβ′ precipitate morphology and evolution in Mg–RE alloys using a combination of first-principles calculations and phase-field modeling. Acta Materi. 76, 259–271 (2014). [Google Scholar]
  16. Sutton A. P. & Balluffi R. W.Overview no. 61 On geometric criteria for low interfacial energy. Acta Metall., 35, 2177–2201 (1987). [Google Scholar]
  17. Issa A., Saal J. E. & Wolverton C.Formation of high-strengthβ′ precipitates in Mg–RE alloys: The role of the Mg/β precipitate instability. Acta Mater.83, 75–83 (2015). [Google Scholar]
  18. Nie J. F., Wilson N. C., Zhu Y. M. & Xu Z.Solute clusters and GP zones in binary Mg–RE alloys. Acta Mater.106, 260–271 (2016). [Google Scholar]
  19. Kresse G. & Hafner J.Ab inito molecular dynamics for metals. Phys. Rev. B.47, 558–561 (1993). [DOI] [PubMed] [Google Scholar]
  20. Perdew J. P., Kieron B. & Ernzerhof M.Generalized gradient approximation made simple. Phys. Rev. Lett.77, 3865–3868 (1996). [DOI] [PubMed] [Google Scholar]
  21. Koichi M. & Izumi F.VESTA: a three-dimensional visualization system for electronic and structural analysis. J. Appl. Cryst.41, 653–658 (2008). [Google Scholar]
  22. Stukowski A.Visualization and analysis of atomistic simulation data with OVITO–the Open Visualization Tool. Modelling Simul. Mater. Sci. Eng. 18, 015012 (2009). [Google Scholar]
  23. Villaes P. & Calvent L. D.Person’s Handbook of Crystallographic Data for Intermetallic phases, vol. 23. The Materials Information Society. 2691 (1996). [Google Scholar]
  24. Stukowski A.Structure identification methods for atomistic simulations of crystalline materials. Modelling Simul. Mater. Sci. Eng. 20, 045021 (2012). [Google Scholar]
  25. Liu W., Li J. C., Zheng W. T. & Jiang Q.Ni Al (110)/Cr (110) interface: A density functional theory study. Phys. Rev. B.73, 205421 (2006). [Google Scholar]
  26. Wachowicz E. & Kiejna A.Bulk and surface properties of hexagonal-close-packed Be and Mg. J. Phys. Condens. Matter. 13, 10767 (2001). [Google Scholar]
  27. Markus J. & Groß A.Microscopic properties of lithium, sodium, and magnesium battery anode materials related to possible dendrite growth. J. Chem. Phys.141, 174710 (2014). [DOI] [PubMed] [Google Scholar]
  28. Kittel C.Introduction to Solid State Physics, 8th ed. (John Wiley & Sons, New York, 2004). [Google Scholar]
  29. Sutton Adrian P.Electronic structure of materials. (Clarendon Press, pp-291993). [Google Scholar]
  30. Zhu Y. M., Liu H., Xu Z., Wang Y. & Nie J. F.Linear-chain configuration of precipitates in Mg–Nd alloys. Acta Mater.83, 239–247 (2015). [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Materials
srep40540-s1.pdf (1.2MB, pdf)

Articles from Scientific Reports are provided here courtesy ofNature Publishing Group

ACTIONS

RESOURCES

[8]ページ先頭
©2009-2025 Movatter.jp