- Email: [email protected]
Investigation of phase selection hierarchy in Mn–Al alloys
Intermetallics 115 (2019) 106617
Contents lists available at ScienceDirect
Intermetallics journal homepage: http://www.elsevier.com/locate/intermet
Investigation of phase selection hierarchy in Mn–Al alloys Ayse Merve Genc a, 1, Ozgun Acar a, 1, Servet Turan b, Ilkay Kalay c, Umut Savacı b, Yunus Eren Kalay a, * a
Department of Metallurgical and Materials Engineering, Middle East Technical University, Ankara, 06800, Turkey Department of Materials Science and Engineering, Eskis¸ehir Technical University, Eskisehir, 26000, Turkey c Department of Materials Science and Engineering, Cankaya University, Ankara, 06790, Turkey b
A R T I C L E I N F O
A B S T R A C T
Keywords: Permanent magnets Ferromagnetic τ-phase Phase transitions Synchrotron radiation X-ray diffraction Electron microscopy
Primarily attributed to the formation of the ferromagnetic τ-phase, near equiatomic composition of Mn–Al have recently received much attention in permanent magnet industry. Several mechanisms have been proposed in literature for the τ-phase formation but controversy still arises regarding the dominating mode. In the current work, MnAl-based alloys having different compositions in a range between Mn50.5Al49.5 and Mn57Al43 have been studied by means of in-situ high energy X-ray diffraction, differential scanning calorimetry (DSC) and magnetic measurements. Synchrotron and DSC experiments showed the dependence of the τ-MnAl on the intermediate disordered ε0 -phase. Alloys having 53.4 at% or less Mn (S1, S2) followed a transformation route of εþε’→τ→βþγ2 upon annealing. Alloys having more than 53.4 at% of Mn had only ε-phase. High energy X-ray diffraction pat terns showed that ε-phase directly transformed into stable phases in the absence of ε0 -phase. It is observed that ε0 not only promoted the ferromagnetic τ-phase but also stabilized it by delaying the nucleation of stable phases.
1. Introduction Most of the current permanent magnet industry depends on rareearth (RE) based magnets. Despite their significantly high magnetic properties, strategical importance of RE elements has induced extensive research towards improvement of RE-free permanent magnets. As far as the magnetic properties of the RE and RE-free permanent magnets are concerned, there is a gap between the maximum energy product (BHmax) values. BHmax of RE magnets lies in between 30 and 45 MGOe whereas its closest successor AlNiCo alloy has a BHmax value of only 6 MGOe [1, 2]. MnAl-based alloys, having a theoretical BHmax value in between RE magnets and AlNiCo, are potential candidates for magnetic applications [3]. This novel alloy has attracted much interest owing to its low cost, high corrosion resistance, high energy per unit weight and high coer civity relative to the other conventional RE-free permanent magnets [4, 5]. MnAl-based alloys exhibit an attractive combination of magnetic properties for many technological applications. Its magnetism arises from a metastable tetragonal τ-phase [6]. As the τ-phase is metastable, it can be obtained with non-equilibrium solidification techniques. Conventionally τ-phase is achieved through quenching of ε-phase fol lowed by annealing at a relatively lower temperature. Several
fabrication methods such as melt spinning, splat quenching, gas atom ization and mechanical milling have been also in use since they allow access to thermodynamically metastable phases [7–10]. Production method and alloy composition are two significant factors affecting the structure. In the Mn–Al phase diagram, τ-phase formation range is marked by a line at compositions between 50 at.%Mn and 60 at. %Mn [4,11]. Thereby, most of the studies are concentrated within this range. Singh and co-researchers (2015) studied the effects of different processing techniques on Mn54Al46 alloy. The magnetic properties were found to depend on the crystallite size of the ε-phase and the density of defects. Mechanical alloying resulted in a microstructure which consists of small grain size, high defect density, but low amount of τ-phase. The arc-melted alloy had largest grain size and the highest fraction of τ-MnAl phase. The highest Curie temperature (638 K) and coercivity (1.8 kOe) were achieved for the samples prepared by a combination of arc-melting and ball milling. Improved magnetic properties are attributed to the pinning of ferromagnetic domains by the defects. These defects also believed to act as nucleation sites for the τ-phase [12]. Optimum mag netic properties of the same composition (Mn54Al46) were achieved after annealing the sample at 673 K (400 � C) and obtaining an average grain size of 300 μm in the study by Foreback et al. (2008) [13]. Similarly,
* Corresponding author. E-mail address: [email protected] (Y.E. Kalay). 1 These authors contributed equally to this work. https://doi.org/10.1016/j.intermet.2019.106617 Received 4 August 2019; Received in revised form 1 September 2019; Accepted 24 September 2019 Available online 30 September 2019 0966-9795/© 2019 Elsevier Ltd. All rights reserved.
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Zeng and Baker (2006) observed that magnetic properties are strongly dependent on both the fraction of τ-phase and the grain size. Mechanical alloying and mechanical milling were used to produce nanocrystalline microstructure. A coercivity of 4.8 kOe and a remanence of 87 emu/g were reported for Mn54Al46 powders annealed at 673 K (400 � C) for 10 min [14]. The ε→τ transition temperature of melt-spun Mn55Al45 decreased by 100 K due to fine ε-phase grain size [15]. In order to improve the magnetic properties of MnAl-based perma nent magnet alloys, the τ phase formation mechanism must be well established. There are three different theories proposed in literature attempting to explain mechanism of ε→τ phase transformation. The first theory states that the ferromagnetic τ-phase forms via a two step pro cess; metastable B19 ordering within the parent ε-phase (ε →ε0 ) followed by a displacive or martensitic shear transformation (ε’ →τ) [3,16–20]. ε0 is an intermediate phase and is not shown on the phase diagram. It is suggested that ε-phase already contains numerous small nuclei of the ε0 -phase. Upon heating, ordering takes place and high density of stacking faults show up in the εþε0 matrix. This faulting gives rise to the martensitic shear, which transforms ε0 into a ferromagnetic τ-phase. It is later observed that the second step involves a thermally activated reordering transformation in addition to the martensitic shear [19]. The second theory suggests that ε→τ phase transformation occurs via a compositionally invariant, diffusional transformation known as massive transformation [21–24]. It is observed that the τ-phase het erogeneously nucleates at the grain boundaries and grows behind advancing incoherent interfaces over the temperature range of 723–973 K (450–700 � C) [21]. The resultant microstructure of the τ-phase contains a high density of lattice defects i.e. dislocations and twins as supported by the electron microscopy observations [22]. Dur ing microstructural analysis, τ-MnAl was found to contain arrays of overlapping octahedral stacking faults, microtwins, thermal antiphase boundaries and dislocations. The origin of these defects was found to be the atomic attachment on {111} and {020}-type facets at the migrating massive growth interface [23]. Classical nucleation theory is applied in order to explain this behavior and it is suggested that the growth mechanism consists of random atomic jumps from the parent phase to nucleation sites [24]. The third theory involves a new transformation mechanism having both structural shear and atomic diffusion which is named as hybrid displacive-diffusional transformation. Wiezorek et al. (2011) carried out transmission electron microscopy (TEM) analysis to support this theory. It was asserted that the transformation kinetics are controlled by the massive transformation mode, whereas partial dislocation glide causes the τ-phase formation. TEM results showed that εꞌ-phase and ε-phase coexist in the same matrix, and εꞌ-phase is present due to the initial elastically accommodated anisotropic misfit strain. Stacking faults overlap each other which were observed in the ε þ εꞌ matrix causing the formation of τ phase. This was also confirmed by in-situ heating TEM imaging showing that there were numerous twins and stacking faults on the close-packed planes of ferromagnetic τ-phase [25]. Even though previous studies provided insights on MnAl-based al loys, the exact ε→τ transformation mechanism is still unclear as the composition dependency is considered. There are some studies which correlate synchrotron light source data with thermal analysis [9,26]. However, this correlation has never been used to investigate the effect of composition on the phase transformations for single MnAl alloys. In this work, the transformation sequence and structural changes in four different MnAl-based alloys were investigated by means of in-situ high energy X-ray diffraction, isochronal differential scanning calorimetry (DSC) and magnetic measurements.
argon atmosphere. These ingots were re-melted three times to ensure compositional homogeneity. After the production, alloys were sol utionized to produce single ε-phase at 1423 K (1150 � C) for an hour. Then, they were subsequently quenched in water to conserve the high temperature metastable ε-phase at room temperature. The presence of ε-phase was confirmed by Cu-Kα X-ray diffraction (XRD). The nominal compositions were determined using energy dispersive X-ray spectros copy (EDS) attached to FEG-SEM. Differential scanning calorimetry (DSC) was used to analyze thermal behavior during solid-state phase transformations. Powder specimens were filled in Al pans and covered with Al lids, and they were heated-up from room temperature to 773 K (500 � C) at 10 K/min under a protective N2 gas atmosphere. High energy X-ray diffraction (HEXRD) experiments were performed at BL04-MSPD beamline of ALBA Synchrotron Light Laboratory. In-situ continuous heating was applied up to 773 K (500 � C) with a heating rate of 10 K/min and the corresponding phase changes were observed. Samples in powder form were sealed into 1 mm diameter borosilicate capillaries. These specimens were exposed to X-rays with a wavelength of 0.4962 Å. Magnetic properties were measured with a vibrating sample magne tometer (VSM) under an applied field of 5 T at room temperature. TEM specimens were prepared by focused ion beam technique. In-situ TEM analyses were performed using Jeol JEM-2100F Transmission Electron Microscope (200 kV). Samples were heated from room temperature to 823 K (550 � C) with 10 K/min as in synchrotron experiments. 3. Results & discussion The compositions were purposely selected to represent possible
τ-phase forming chemistries according to Mn–Al phase diagram [4] and
coded as S1, S2, S3 and S4. Each specimen was solutionized at 1423 K (1150 � C) for an hour and quenched in water subsequently. Conven tional XRD showed the existence of single ε-phase in all samples as shown in Fig. 1. The nominal chemistries were verified using EDS and the results were tabulated in Table 1. Fig. 2 shows the HEXRD data on 2D-films obtained while heating the solutionized alloys from room temperature to 773 K (500 � C) with a heating rate of 10 K/min within a furnace attached to the synchrotron beamline. Diffraction patterns were found to be different for alloys with less Mn (S1 and S2) and for the alloys with larger amounts of Mn (S3 and S4). A broad peak having low intensity was observed around 19 nm 1 (Q) in S1 and S2 whereas that signal was missing for the specimens S3 and S4. It is known that ε and the intermediate ε0 -phase have over lapping diffraction patterns. The only characteristic peak of ε0 -phase is at 19.7 nm 1 (Q). 2D synchrotron film plots suggest that ε0 and ε-phase were present at room temperature in S1 and S2, on the other hand; asquenched state of S3 and S4 had only ε-phase. The sequence of phase transformations was found to be similar in S1 and S2 while it was different for S3 and S4 specimens (Fig. 2). Fig. 2a and b shows that ε and ε0 slowly disappear and the peaks of metastable τ-phase appear upon heating. εþε0 matrix completely diminishes after 740 K (467 � C). τ-phase formation was followed by the stable nonmagnetic phases at elevated temperatures. On the contrary, in S3 and S4 alloys (Fig. 2c and d) room temperature ε-phase directly transforms into stable phases with increasing temperature, surpassing the ferro magnetic τ-phase transformation. Detailed representation of the high energy X-ray pattern shows the change in the ε0 -phase peak intensity with respect to temperature (Fig. 3). It is clearly seen that ε0 -phase was already present in quenched state for S1 and S2. In alloy S1 (Fig. 3a), the hump at room temperature around 19.7 nm 1 (Q) sharpens at 633 K (360 � C) and vanishes around 724 K (451 � C). In alloy S2 (Fig. 3b), a similar pattern was observed but the temperature at which the hump grows into a peak was different. The intensity of ε0 -peak reaches its maximum value at 684 K (411 � C) and then the peak disappears at 733 K (460 � C). This suggests that matrix has already ε0 nuclei at room temperature. As the temperature increases, ε0 nuclei grows within the matrix up to a certain limit and disappears with
2. Experimental procedure Ingots of Mn–Al alloys within the composition range of 50.5–57.0 at % Mn - bal. Al were produced in a copper-hearth electric arc-melting furnace from high-purity Mn (99.95%) and Al (99.8%) elements under 2
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Fig. 1. XRD pattern of sample (a) S1, (b) S2, (c) S3, and (d) S4 after solutionizing.
composition. Formation and stability range of each phase is listed in Table 2 εþε0 matrix transforms into τ-phase at 633 K (360 � C) and 643 K (370 � C) for S1 and S2, respectively (Fig. 5a and b) whereas stable nonmagnetic phases form subsequently at 695 K (422 � C) and 700 K (427 � C) (Fig. 5a and b). Hence, there is a difference of 55–60 K between the formation of τ-phase and stable phases. Temperature at which stable phases formed was around 695 K (422 � C) for S1 and S2 alloys while transformation temperature de creases to 635 K (362 � C) for S3 and S4 (Fig. 5c and d). This suggests that presence of ε0 -phase not only promotes the τ-phase formation but also retards the transformation to the stable phases. DSC analysis was carried out to identify the thermal behavior. A similar experimental procedure with synchrotron analysis was used in DSC experiments for comparison and correlation with the HEXRD data. As-quenched samples were annealed up to 773 K (500 � C) with a heating rate of 10 K/min in DSC and quenched to maintain the microstructure. The purpose of using similar heating rates was to mimic HEXRD ex periments so that isochronal phase changes can be identified. XRD analysis prior to DSC experiments confirmed that ε-phase was obtained at room temperature. The heating curves obtained in DSC are given in Fig. 6. Most distinctive differences are the shape of the curves and the positions of the peaks. The heating pattern of the S1 and S2 are similar to each other but different than S3 and S4. For S1 and S2, a small hump at around 520 K (247 � C) was followed by a peak at around 680 K (407 � C). This finding is in good agreement with the synchrotron data showing a low intensity ε0 diffraction peak at 19.7 nm 1 (Q) in specimens S1 and S2 (Fig. 3a and b). The increase in the intensity of the ε0 peak was attributed to the growth of ε0 nuclei. It is likely that the hump observed in DSC scans of S1 and S2 alloys (520 K/247 � C) corresponds to the growth of the ε0 -phase. The small hump was missing in the DSC scans of specimens S3 and S4 and the major peak was observed at a relatively lower tem perature (640 K/367 � C). HEXRD patterns of S3 and S4 (Fig. 3c and d) do not have a Bragg reflection at 19.7 nm 1 (Q) which confirms the absence
Table 1 EDS results of the samples S1, S2, S3, and S4. Mn and Al ratios of the alloys are given in atomic and weight percentages. Sample Name S1 S2 S3 S4
Mn
Al
at%
wt%
at%
wt%
50.5 � 0.3 53.4 � 0.2 55.2 � 0.1 57.0 � 0.5
67.5 � 0.3 70.0 � 0.2 71.5 � 0.1 73.0 � 0.5
49.5 � 0.3 46.6 � 0.2 44.8 � 0.1 43.0 � 0.5
32.5 � 0.3 30.0 � 0.2 28.5 � 0.1 27.0 � 0.5
further increase. It should also be noted that the intensity of the ε0 -peak in S1 and S2 was relatively low that it is very difficult to distinguish its presence by any characterization technique other than high energy X-ray diffraction. Characteristic ε0 peak at 19.7 nm 1 (Q) corresponds to 27� (2ϴ) in conventional X-ray (CuKα). It is likely that Fig. 1a and b have an ε0 peak at 27� (2ϴ), but it is not possible to detect due to its low intensity. High energy X-ray diffraction patterns of the alloys S1, S2, S3 and S4 between Q values of 24.5–33.0 nm 1 is given in Fig. 4. In alloys S1 and S2, ferromagnetic τ-phase forms with increasing temperature (Fig. 4a and b); however, in alloys S3 and S4 (Fig. 4c and d) the τ-phase peaks were not observed (Fig. 4c and d). The change in transformation scheme indicates the significance of the ε0 -phase in terms of τ phase formation. When ε0 -phase was absent at the quenched state, as in the case of S3 and S4 alloys, τ-phase cannot be obtained. It is likely that ε0 -phase serve as nucleation sites for the τ-phase. Fig. 5 shows the high energy X-ray diffraction patterns of the alloys S1, S2, S3 and S4 between 10.0 and 24.0 nm 1 (Q). Stable phases were formed in all samples upon heating independent of composition. It is likely that the stable phases arise from a matrix which contains ε and ε0 phases. Fig. 5 indicates that the transformation temperatures vary with the 3
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Fig. 2. 2-D film synchrotron plots of (a) S1 (b) S2, (c) S3, and (d) S4. Red lines indicate the τ phase peak positions.
Fig. 3. High energy X-ray diffraction patterns of the alloys (a) S1, (b) S2, (c) S3 and (d) S4 within 18.2–21.2 nm different temperatures.
of ε’. Heating path of S1 and S2 alloys suggests an overlap of the main peak with a shoulder peak appearing at 650 K (377 � C). The shoulder peak is marked with an arrow on Fig. 6. Correlating with the synchrotron data (Fig. 4a and b), it can be speculated that the shoulder peak in S1 and S2 specimens correspond to the formation of τ-phase. On the other hand,
1
(Q) range showing the evolution of εꞌ peak at
DSC heating curve of S3 and S4 was found to display a sharp single peak leaning towards lower temperatures. The shoulder peak is absent in S3 and S4. The remaining transition can be attributed to the formation of stable phases. This finding is in good agreement with the X-ray diffraction data presented in Fig. 5g and d indicating that only stable phases are formed and τ-phase is missing. Onset temperature of the 4
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Fig. 4. High energy X-ray diffraction patterns of the alloys (a) S1, (b) S2, (c) S3 and (d) S4 within 24.5–33.0 nm upon heating.
1
Fig. 5. High energy X-ray diffraction patterns of the alloys (a) S1, (b) S2, (c) S3 and (d) S4 within 10.0–24.0 nm phases upon heating. 5
(Q) range showing the formation of τ-MnAl peak
1
(Q) range showing the formation of the stable
A.M. Genc et al.
Intermetallics 115 (2019) 106617
phases. As expected, alloys S1 and S2 showed ferromagnetic behavior due to the presence of τ-phase. Coercivity and remanence of S1 was found to be 640 Oe and 23 emu/g, respectively. S2 had relatively stronger magnetic behavior having a coercivity of 1560 Oe and rema nence of 26 emu/g. With further increase of Mn, ε0 -phase becomes ab sent thus, τ-phase cannot be obtained. Alloys S3 and S4 were paramagnetic in nature as τ-phase is not observed. In order to understand and visualize the temperature effects on ε→τ transformation, in-situ heating TEM experiments were carried out. Still frames extracted from the dynamic video sequence during in-situ heat ing TEM experiment of sample S2 are presented in Fig. 8. A dense pattern of planar defects was observed at room temperature. Electron diffraction patterns regarding to ε and εꞌ regions couldn’t be collected because of the camera positions configured in the TEM used in this study. However, based on our DSC and XRD analyses, it is hypothesized that planar defects may possess ε0 -phase, so ε and εꞌ-phases coexist in the same matrix. It should be noted that εꞌ regions may be nanosized, therefore, the arrows shown in Fig. 8 is intended for demonstration of possible phase locations. Further high resolution TEM and electron diffraction analyses should be conducted to monitor εꞌ regions. As temperature increases, defect arrays (ε0 ) become denser and eventually transforms into morphologically plate-like τ-phase at 390 � C. Upon further annealing, stable phases readily form at around 520 � C. At 550 � C, sample mostly consists of stable phases along with some τ-phase and trace amount of ε-phase residuals. Same experimental procedure was carried out for sample S4. Screenshots extracted from the dynamic video sequence during in-situ heating TEM experiment of sample S4 are shown in Fig. 9. Planar defect arrays observed at room temperature in S2 were missing in case of S4. In good agreement with XRD results, τ-MnAl formation was not observed and ε-phase directly transformed into stable phases. The transformation temperature found to be lower than S2. Stable phase transformation started at 520 � C in S2 whereas, the value dropped to 460 � C for S4 alloy. This finding is consistent with the synchrotron and DSC analyses suggesting that ε0 -phase retards the stable phase forma tion. At 550 � C, whole sample was decomposed into β-Mn and γ2. Previous studies indicated that ε→τ transformation may occur either by massive mode, shear mode or a combination of both (hybrid). Studies suggesting massive transformation claim that shear mechanism is invalid, ε→τ transformation is independent of the ε0 ordering and it can proceed without the ε0 -phase [17,21–23,27]. However, the synchrotron data presented in this work shows ε0 -phase is a strong prerequisite for
Table 2 Critical transformation temperatures of ε, ε0 , τ and stable phases for alloys S1, S2, S3 and S4. Sample Name
Phase Type and Range
ε
ε0
τ
Stable Phases
S1
300–741 K (27–468 � C)
634–773 K (361–500 � C)
695–773 K (422–500 � C)
S2
300–740 K (27–467 � C)
643–773 K (370–500 � C)
700–773 K (427–500 � C)
S3
300–729 K (27–456 � C) 300–729 K (27–456 � C)
300-633a724 K (27–360a451 � C) 300-684a733 K (27411-460 � C) –
–
635–773 K (362–500 � C) 639–773 K (366–500 � C)
S4 a
–
–
Temperature at which ε0 peak starts to intensify.
Fig. 6. DSC heating curves of S1, S2, S3 and S4 with a heating rate of 10 K/min from room temperature to 773 K (500 � C).
major peaks in S1, S2, S3, and S4 were calculated as 637 K (364 � C), 636 K (363 � C), 627 K (354 � C), and 621 K (348 � C), respectively. Magnetic hysteresis curves obtained for the samples after DSC analysis are given in Fig. 7. Alloys S1, S2, S3 and S4 were heated from room temperature to 773 K (500 � C) at 10 K/min under N2 gas atmo sphere and quenched subsequently. Magnetic properties of those sam ples were measured by vibrating sample magnetometer (VSM). According to HEXRD (Fig. 4a and b). S1 and S2 have ferromagnetic τ-phase and stable phases while S3 and S4 should have only stable
Fig. 8. Screenshots extracted from the dynamic video sequence during in-situ heating TEM experiment of S2 at temperatures of (a)30, (b) 390, (c) 520 and (d) 550 � C. Dashed lines indicate the grain boundary between ε and β-phases.
Fig. 7. Hysteresis curves of S1, S2, S3 and S4, inset showing the second quadrant of the B–H curve of S1 and S2 specimens. 6
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Declaration of competing interest 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. Acknowledgement In-situ high energy X-ray experiments in this research was under taken on the BL04-MSPD beamline at the ALBA Synchrotron facility, Spain. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.intermet.2019.106617. References [1] J.M.D. Coey, Permanent magnet applications, J. Magn. Magn. Mater. 248 (2002) 441–456, https://doi.org/10.1016/S0304-8853(02)00335-9. [2] J.M.D. Coey, Permanent magnets: plugging the gap, Scr. Mater. 67 (2012) 524–529, https://doi.org/10.1016/j.scriptamat.2012.04.036. [3] T. Ohtani, N. Kato, S. Kojima, K. Kojima, Y. Sakamoto, I. Konno, M. Tsukahara, T. Kubo, Magnetic properties of Mn-Al-C permanent magnet alloys, IEEE Trans. Magn. 13 (1977) 1328–1330, https://doi.org/10.1109/TMAG.1977.1059574. [4] Q. Zeng, I. Baker, J.B. Cui, Z.C. Yan, Structural and magnetic properties of nanostructured Mn–Al–C magnetic materials, J. Magn. Magn. Mater. 308 (2007) 214–226, https://doi.org/10.1016/j.jmmm.2006.05.032. [5] A. Chaturvedi, R. Yaqub, I. Baker, Microstructure and magnetic properties of bulk nanocrystalline MnAl, Metals (Basel) 4 (2014) 20–27, https://doi.org/10.3390/ met4010020. [6] H. K€ ono, On the ferromagnetic phase in MnAl system, J. Phys. Soc. Japan. 13 (1958) 1444–1451. [7] A. Yamaguchi, Y. Tanaka, K. Yanagimoto, I. Sakaguchi, N. Kato, Development of P/ M Mn-Al-C alloy permanent magnet, Bull. Jpn. Inst. Met. 28 (1989) 422. [8] E. Fazakas, L.K. Varga, F. Mazaleyrat, Preparation of nanocrystalline Mn–Al–C magnets by melt spinning and subsequent heat treatments, J. Alloy. Comp. 434–435 (2007) 611–613, https://doi.org/10.1016/j.jallcom.2006.08.313. [9] I. Janotova, P.S. Svec, P. Svec, I. Mat’ko, D. Janickovic, J. Zigo, M. Mihalkovic, J. Marcin, I. Skorvanek, Phase analysis and structure of rapidly quenched Al-Mn systems, J. Alloy. Comp. 707 (2017) 137–141, https://doi.org/10.1016/j. jallcom.2016.11.171. [10] J.Y. Law, J. Rial, M. Villanueva, N. Lopez, J. Camarero, L.G. Marshall, J. S. Blazquez, J.M. Borrego, V. Franco, A. Conde, L.H. Lewis, A. Bollero, Study of phases evolution in high-coercive MnAl powders obtained through short milling time of gas-atomized particles, J. Alloy. Comp. 712 (2017) 373–378, https://doi. org/10.1016/j.jallcom.2017.04.038. [11] T.B. Massalski, Binary Alloy Phase Diagrams, 1990. http://www.amazon.com/Bina ry-Alloy-Phase-Diagrams-Massalski/dp/087170403X. [12] N. Singh, V. Mudgil, K. Anand, A.K. Srivastava, R.K. Kotnala, A. Dhar, Influence of processing on structure property correlations in τ-MnAl rare-earth free permanent magnet material, J. Alloy. Comp. 633 (2015) 401–407, https://doi.org/10.1016/j. jallcom.2015.02.041. [13] B. Foreback, I. Baker, Permanent Magnets from Mechanically Milled Mn54Al46 Alloys, 2008. [14] Q. Zeng, I. Baker, Z.C. Yan, Nanostructured Mn-Al permanent magnets produced by mechanical milling, J. Appl. Phys. 99 (2006) 10–13, https://doi.org/10.1063/ 1.2159187. [15] F. Jim� enez-Villacorta, J. Marion, J. Oldham, M. Daniil, M. Willard, L. Lewis, Magnetism-Structure correlations during the ε→τ transformation in rapidlysolidified MnAl nanostructured alloys, Metals (Basel) 4 (2014) 8–19, https://doi. org/10.3390/met4010008. [16] J. Van Landuyt, G. Van Tendeloo, J.J. Van Den Broek, H. Donkersloot, Permanent magnetism and microstructure in τ-AlMn(C), J. Magn. Magn. Mater. 15–18 (1980) 1451–1452. [17] Y.J. Kim, J.H. Perepezko, The thermodynamics and competitive kinetics of metastable τ phase development in MnAl-base alloys, Mater. Sci. Eng. A 163 (1993) 127–134. [18] D.C. Crew, P.G. McCormick, R. Street, MnAl and MnAlC permanent magnets produced by mechanical alloying, Scr. Metall. Mater. 32 (1995) 315–318, https:// doi.org/10.1016/S0956-716X(99)80057-0. [19] P. Muellner, B.E. Buergler, H. Heinrich, A.S. Sologubenko, G. Kostorz, Observation of the shear mode of the ε→τ phase transformation in a Mn-Al-C single crystal, Philos. Mag. Lett. 82 (2002) 71–79, https://doi.org/10.1080/ 09500830110103225. [20] J.J.V.D.B. Broek, H. Donkersloot, Phase transformations in pure and carbon-doped Al45Mn55 alloys, Acta Metall. 27 (1979) 1497–1504. https://doi. org/10.1016/0001-6160(79)90172-X.
Fig. 9. Screenshots extracted from the dynamic video sequence during in-situ heating TEM experiment of S4 at temperatures of (a)30, (b) 400, (c) 460 and (d) 550 � C. Dashed lines indicate the grain boundary between ε and β-phases.
τ-phase formation. Alloys having less Mn (S1 and S2) already had ε0 -phase within the ε-matrix at room temperature. It was observed that ε-phase first transforms into the ordered ε0 -phase and then τ-phase was formed via martensitic shear of the ε’. Upon annealing, ε0 readily transforms into ferromagnetic τ-phase. This finding is in good correla tion with shear [3,16–20] and hybrid mode studies [25] which state that
τ-phase was generated due to the direct shear of the ordered ε0 -phase
regions. Through in-situ TEM experiments, it is speculated that the evolution of the τ-MnAl phase involves a nucleation and growth process wherein, the nucleation takes place heterogeneously along the ε0 phase grain boundaries. Grain boundaries with high interfacial energy and relatively weak bonding are often the preferred nucleation sites for heterogeneous nucleation. 4. Conclusions
Based upon the results obtained from in-situ high energy X-ray diffraction as well as differential scanning calorimetry, magnetic mea surements and in-situ TEM analyses of MnAl-based alloy, phase trans formation hierarchy of the compositions that lead to the formation of metastable ferromagnetic τ-phase have been developed. Overall results indicate that the composition of MnAl-based alloys affect both the sequence and the temperature of phase transformations. Alloys having Mn content less than 53.4 at% (S1, S2) has a parent phase mixture consisting of ε and ε0 phases at the quenched state. Upon annealing, growth of ε0 -phase was followed by a phase transformation into ferro magnetic τ-phase and stable phases. Alloys which have more than 53.4 at% of Mn (S3, S4) consist of only ε-phase at the quenched state. ε-phase directly transforms into stable phases with an increase in temperature. High energy X-ray diffraction patterns and DSC scans both confirmed that ε0 -phase was missing and τ-MnAl formation was surpassed. Alloys having τ-phase were found to be ferromagnetic in VSM. This finding suggest that magnetic τ-MnAl formation cannot proceed in the absence of ε0 -phase. In situ TEM imaging confirmed the transformation sequence of εþε’→τ→stable phases for S2 alloy. ε and ε0 -phases were observed at room temperature and transformed into τ-phase upon annealing. Stable phases followed τ-phase formation spontaneously. It is observed that compositional ordering do not take place as Mn content increases as in S4 alloy. Only a certain composition range results in the formation of the ε0 -phase and only that composition range results in the τ-phase forma tion. Mn70% (wt) was found to be the limit for ε0 and τ-phase formation.
7
A.M. Genc et al.
Intermetallics 115 (2019) 106617
[21] D.P. Hoydick, E.J. Palmiere, W.A. Soffa, On the formation of the metastable L1o phase in manganese-aluminum-base permanent magnet materials, Scr. Mater. 36 (1997) 151–156, https://doi.org/10.1016/S1359-6462(96)00356-9. [22] D.P. Hoydick, E.J. Palmiere, W.a. Soffa, Microstructural development in MnAl-base permanent magnet materials: new perspectives, J. Appl. Phys. 81 (1997) 5624, https://doi.org/10.1063/1.364619. [23] C. Yanar, V. Radmilovic, W.A. Soffa, J.M.K. Wiezorek, Evolution of microstructure and defect structure in L10-ordered manganese aluminide permanent magnet alloys, Intermetallics 9 (2001) 949–954, https://doi.org/10.1016/S0966-9795(01) 00095-4. [24] C. Yanar, J.M.K. Wiezorek, V. Radmilovic, W.A. Soffa, Massive transformation and the formation of the ferromagnetic L10 phase in manganese-aluminum-based
alloys, Metall. Mater. Trans. A 33A (2002) 2002–2413, https://doi.org/10.1007/ s11661-002-0363-3. [25] J.M.K. Wiezorek, A.K. Kulovits, C. Yanar, W.A. Soffa, Grain boundary mediated displacive-diffusional formation of τ-phase MnAl, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 42 (2011) 594–604, https://doi.org/10.1007/s11661-0100308-1. [26] H. Fang, J. Cedervall, F.J.M. Casado, Z. Matej, J. Bednarcik, J. Ångstr€ om, P. Berastegui, M. Sahlberg, Insights into formation and stability of τ-MnAlZx (Z ¼ C and B), J. Alloy. Comp. 692 (2017) 198–203, https://doi.org/10.1016/j. jallcom.2016.09.047. [27] V.M. Gundyrev, M.A. Uimin, A.E. Ermakov, O.B. Andreeva, Transformation in MnAl-C alloys, Phys. Status Solidi. 91 (1985) K55–K58, https://doi.org/10.1002/ pssa.2210910152.
8
Contents lists available at ScienceDirect
Intermetallics journal homepage: http://www.elsevier.com/locate/intermet
Investigation of phase selection hierarchy in Mn–Al alloys Ayse Merve Genc a, 1, Ozgun Acar a, 1, Servet Turan b, Ilkay Kalay c, Umut Savacı b, Yunus Eren Kalay a, * a
Department of Metallurgical and Materials Engineering, Middle East Technical University, Ankara, 06800, Turkey Department of Materials Science and Engineering, Eskis¸ehir Technical University, Eskisehir, 26000, Turkey c Department of Materials Science and Engineering, Cankaya University, Ankara, 06790, Turkey b
A R T I C L E I N F O
A B S T R A C T
Keywords: Permanent magnets Ferromagnetic τ-phase Phase transitions Synchrotron radiation X-ray diffraction Electron microscopy
Primarily attributed to the formation of the ferromagnetic τ-phase, near equiatomic composition of Mn–Al have recently received much attention in permanent magnet industry. Several mechanisms have been proposed in literature for the τ-phase formation but controversy still arises regarding the dominating mode. In the current work, MnAl-based alloys having different compositions in a range between Mn50.5Al49.5 and Mn57Al43 have been studied by means of in-situ high energy X-ray diffraction, differential scanning calorimetry (DSC) and magnetic measurements. Synchrotron and DSC experiments showed the dependence of the τ-MnAl on the intermediate disordered ε0 -phase. Alloys having 53.4 at% or less Mn (S1, S2) followed a transformation route of εþε’→τ→βþγ2 upon annealing. Alloys having more than 53.4 at% of Mn had only ε-phase. High energy X-ray diffraction pat terns showed that ε-phase directly transformed into stable phases in the absence of ε0 -phase. It is observed that ε0 not only promoted the ferromagnetic τ-phase but also stabilized it by delaying the nucleation of stable phases.
1. Introduction Most of the current permanent magnet industry depends on rareearth (RE) based magnets. Despite their significantly high magnetic properties, strategical importance of RE elements has induced extensive research towards improvement of RE-free permanent magnets. As far as the magnetic properties of the RE and RE-free permanent magnets are concerned, there is a gap between the maximum energy product (BHmax) values. BHmax of RE magnets lies in between 30 and 45 MGOe whereas its closest successor AlNiCo alloy has a BHmax value of only 6 MGOe [1, 2]. MnAl-based alloys, having a theoretical BHmax value in between RE magnets and AlNiCo, are potential candidates for magnetic applications [3]. This novel alloy has attracted much interest owing to its low cost, high corrosion resistance, high energy per unit weight and high coer civity relative to the other conventional RE-free permanent magnets [4, 5]. MnAl-based alloys exhibit an attractive combination of magnetic properties for many technological applications. Its magnetism arises from a metastable tetragonal τ-phase [6]. As the τ-phase is metastable, it can be obtained with non-equilibrium solidification techniques. Conventionally τ-phase is achieved through quenching of ε-phase fol lowed by annealing at a relatively lower temperature. Several
fabrication methods such as melt spinning, splat quenching, gas atom ization and mechanical milling have been also in use since they allow access to thermodynamically metastable phases [7–10]. Production method and alloy composition are two significant factors affecting the structure. In the Mn–Al phase diagram, τ-phase formation range is marked by a line at compositions between 50 at.%Mn and 60 at. %Mn [4,11]. Thereby, most of the studies are concentrated within this range. Singh and co-researchers (2015) studied the effects of different processing techniques on Mn54Al46 alloy. The magnetic properties were found to depend on the crystallite size of the ε-phase and the density of defects. Mechanical alloying resulted in a microstructure which consists of small grain size, high defect density, but low amount of τ-phase. The arc-melted alloy had largest grain size and the highest fraction of τ-MnAl phase. The highest Curie temperature (638 K) and coercivity (1.8 kOe) were achieved for the samples prepared by a combination of arc-melting and ball milling. Improved magnetic properties are attributed to the pinning of ferromagnetic domains by the defects. These defects also believed to act as nucleation sites for the τ-phase [12]. Optimum mag netic properties of the same composition (Mn54Al46) were achieved after annealing the sample at 673 K (400 � C) and obtaining an average grain size of 300 μm in the study by Foreback et al. (2008) [13]. Similarly,
* Corresponding author. E-mail address: [email protected] (Y.E. Kalay). 1 These authors contributed equally to this work. https://doi.org/10.1016/j.intermet.2019.106617 Received 4 August 2019; Received in revised form 1 September 2019; Accepted 24 September 2019 Available online 30 September 2019 0966-9795/© 2019 Elsevier Ltd. All rights reserved.
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Zeng and Baker (2006) observed that magnetic properties are strongly dependent on both the fraction of τ-phase and the grain size. Mechanical alloying and mechanical milling were used to produce nanocrystalline microstructure. A coercivity of 4.8 kOe and a remanence of 87 emu/g were reported for Mn54Al46 powders annealed at 673 K (400 � C) for 10 min [14]. The ε→τ transition temperature of melt-spun Mn55Al45 decreased by 100 K due to fine ε-phase grain size [15]. In order to improve the magnetic properties of MnAl-based perma nent magnet alloys, the τ phase formation mechanism must be well established. There are three different theories proposed in literature attempting to explain mechanism of ε→τ phase transformation. The first theory states that the ferromagnetic τ-phase forms via a two step pro cess; metastable B19 ordering within the parent ε-phase (ε →ε0 ) followed by a displacive or martensitic shear transformation (ε’ →τ) [3,16–20]. ε0 is an intermediate phase and is not shown on the phase diagram. It is suggested that ε-phase already contains numerous small nuclei of the ε0 -phase. Upon heating, ordering takes place and high density of stacking faults show up in the εþε0 matrix. This faulting gives rise to the martensitic shear, which transforms ε0 into a ferromagnetic τ-phase. It is later observed that the second step involves a thermally activated reordering transformation in addition to the martensitic shear [19]. The second theory suggests that ε→τ phase transformation occurs via a compositionally invariant, diffusional transformation known as massive transformation [21–24]. It is observed that the τ-phase het erogeneously nucleates at the grain boundaries and grows behind advancing incoherent interfaces over the temperature range of 723–973 K (450–700 � C) [21]. The resultant microstructure of the τ-phase contains a high density of lattice defects i.e. dislocations and twins as supported by the electron microscopy observations [22]. Dur ing microstructural analysis, τ-MnAl was found to contain arrays of overlapping octahedral stacking faults, microtwins, thermal antiphase boundaries and dislocations. The origin of these defects was found to be the atomic attachment on {111} and {020}-type facets at the migrating massive growth interface [23]. Classical nucleation theory is applied in order to explain this behavior and it is suggested that the growth mechanism consists of random atomic jumps from the parent phase to nucleation sites [24]. The third theory involves a new transformation mechanism having both structural shear and atomic diffusion which is named as hybrid displacive-diffusional transformation. Wiezorek et al. (2011) carried out transmission electron microscopy (TEM) analysis to support this theory. It was asserted that the transformation kinetics are controlled by the massive transformation mode, whereas partial dislocation glide causes the τ-phase formation. TEM results showed that εꞌ-phase and ε-phase coexist in the same matrix, and εꞌ-phase is present due to the initial elastically accommodated anisotropic misfit strain. Stacking faults overlap each other which were observed in the ε þ εꞌ matrix causing the formation of τ phase. This was also confirmed by in-situ heating TEM imaging showing that there were numerous twins and stacking faults on the close-packed planes of ferromagnetic τ-phase [25]. Even though previous studies provided insights on MnAl-based al loys, the exact ε→τ transformation mechanism is still unclear as the composition dependency is considered. There are some studies which correlate synchrotron light source data with thermal analysis [9,26]. However, this correlation has never been used to investigate the effect of composition on the phase transformations for single MnAl alloys. In this work, the transformation sequence and structural changes in four different MnAl-based alloys were investigated by means of in-situ high energy X-ray diffraction, isochronal differential scanning calorimetry (DSC) and magnetic measurements.
argon atmosphere. These ingots were re-melted three times to ensure compositional homogeneity. After the production, alloys were sol utionized to produce single ε-phase at 1423 K (1150 � C) for an hour. Then, they were subsequently quenched in water to conserve the high temperature metastable ε-phase at room temperature. The presence of ε-phase was confirmed by Cu-Kα X-ray diffraction (XRD). The nominal compositions were determined using energy dispersive X-ray spectros copy (EDS) attached to FEG-SEM. Differential scanning calorimetry (DSC) was used to analyze thermal behavior during solid-state phase transformations. Powder specimens were filled in Al pans and covered with Al lids, and they were heated-up from room temperature to 773 K (500 � C) at 10 K/min under a protective N2 gas atmosphere. High energy X-ray diffraction (HEXRD) experiments were performed at BL04-MSPD beamline of ALBA Synchrotron Light Laboratory. In-situ continuous heating was applied up to 773 K (500 � C) with a heating rate of 10 K/min and the corresponding phase changes were observed. Samples in powder form were sealed into 1 mm diameter borosilicate capillaries. These specimens were exposed to X-rays with a wavelength of 0.4962 Å. Magnetic properties were measured with a vibrating sample magne tometer (VSM) under an applied field of 5 T at room temperature. TEM specimens were prepared by focused ion beam technique. In-situ TEM analyses were performed using Jeol JEM-2100F Transmission Electron Microscope (200 kV). Samples were heated from room temperature to 823 K (550 � C) with 10 K/min as in synchrotron experiments. 3. Results & discussion The compositions were purposely selected to represent possible
τ-phase forming chemistries according to Mn–Al phase diagram [4] and
coded as S1, S2, S3 and S4. Each specimen was solutionized at 1423 K (1150 � C) for an hour and quenched in water subsequently. Conven tional XRD showed the existence of single ε-phase in all samples as shown in Fig. 1. The nominal chemistries were verified using EDS and the results were tabulated in Table 1. Fig. 2 shows the HEXRD data on 2D-films obtained while heating the solutionized alloys from room temperature to 773 K (500 � C) with a heating rate of 10 K/min within a furnace attached to the synchrotron beamline. Diffraction patterns were found to be different for alloys with less Mn (S1 and S2) and for the alloys with larger amounts of Mn (S3 and S4). A broad peak having low intensity was observed around 19 nm 1 (Q) in S1 and S2 whereas that signal was missing for the specimens S3 and S4. It is known that ε and the intermediate ε0 -phase have over lapping diffraction patterns. The only characteristic peak of ε0 -phase is at 19.7 nm 1 (Q). 2D synchrotron film plots suggest that ε0 and ε-phase were present at room temperature in S1 and S2, on the other hand; asquenched state of S3 and S4 had only ε-phase. The sequence of phase transformations was found to be similar in S1 and S2 while it was different for S3 and S4 specimens (Fig. 2). Fig. 2a and b shows that ε and ε0 slowly disappear and the peaks of metastable τ-phase appear upon heating. εþε0 matrix completely diminishes after 740 K (467 � C). τ-phase formation was followed by the stable nonmagnetic phases at elevated temperatures. On the contrary, in S3 and S4 alloys (Fig. 2c and d) room temperature ε-phase directly transforms into stable phases with increasing temperature, surpassing the ferro magnetic τ-phase transformation. Detailed representation of the high energy X-ray pattern shows the change in the ε0 -phase peak intensity with respect to temperature (Fig. 3). It is clearly seen that ε0 -phase was already present in quenched state for S1 and S2. In alloy S1 (Fig. 3a), the hump at room temperature around 19.7 nm 1 (Q) sharpens at 633 K (360 � C) and vanishes around 724 K (451 � C). In alloy S2 (Fig. 3b), a similar pattern was observed but the temperature at which the hump grows into a peak was different. The intensity of ε0 -peak reaches its maximum value at 684 K (411 � C) and then the peak disappears at 733 K (460 � C). This suggests that matrix has already ε0 nuclei at room temperature. As the temperature increases, ε0 nuclei grows within the matrix up to a certain limit and disappears with
2. Experimental procedure Ingots of Mn–Al alloys within the composition range of 50.5–57.0 at % Mn - bal. Al were produced in a copper-hearth electric arc-melting furnace from high-purity Mn (99.95%) and Al (99.8%) elements under 2
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Fig. 1. XRD pattern of sample (a) S1, (b) S2, (c) S3, and (d) S4 after solutionizing.
composition. Formation and stability range of each phase is listed in Table 2 εþε0 matrix transforms into τ-phase at 633 K (360 � C) and 643 K (370 � C) for S1 and S2, respectively (Fig. 5a and b) whereas stable nonmagnetic phases form subsequently at 695 K (422 � C) and 700 K (427 � C) (Fig. 5a and b). Hence, there is a difference of 55–60 K between the formation of τ-phase and stable phases. Temperature at which stable phases formed was around 695 K (422 � C) for S1 and S2 alloys while transformation temperature de creases to 635 K (362 � C) for S3 and S4 (Fig. 5c and d). This suggests that presence of ε0 -phase not only promotes the τ-phase formation but also retards the transformation to the stable phases. DSC analysis was carried out to identify the thermal behavior. A similar experimental procedure with synchrotron analysis was used in DSC experiments for comparison and correlation with the HEXRD data. As-quenched samples were annealed up to 773 K (500 � C) with a heating rate of 10 K/min in DSC and quenched to maintain the microstructure. The purpose of using similar heating rates was to mimic HEXRD ex periments so that isochronal phase changes can be identified. XRD analysis prior to DSC experiments confirmed that ε-phase was obtained at room temperature. The heating curves obtained in DSC are given in Fig. 6. Most distinctive differences are the shape of the curves and the positions of the peaks. The heating pattern of the S1 and S2 are similar to each other but different than S3 and S4. For S1 and S2, a small hump at around 520 K (247 � C) was followed by a peak at around 680 K (407 � C). This finding is in good agreement with the synchrotron data showing a low intensity ε0 diffraction peak at 19.7 nm 1 (Q) in specimens S1 and S2 (Fig. 3a and b). The increase in the intensity of the ε0 peak was attributed to the growth of ε0 nuclei. It is likely that the hump observed in DSC scans of S1 and S2 alloys (520 K/247 � C) corresponds to the growth of the ε0 -phase. The small hump was missing in the DSC scans of specimens S3 and S4 and the major peak was observed at a relatively lower tem perature (640 K/367 � C). HEXRD patterns of S3 and S4 (Fig. 3c and d) do not have a Bragg reflection at 19.7 nm 1 (Q) which confirms the absence
Table 1 EDS results of the samples S1, S2, S3, and S4. Mn and Al ratios of the alloys are given in atomic and weight percentages. Sample Name S1 S2 S3 S4
Mn
Al
at%
wt%
at%
wt%
50.5 � 0.3 53.4 � 0.2 55.2 � 0.1 57.0 � 0.5
67.5 � 0.3 70.0 � 0.2 71.5 � 0.1 73.0 � 0.5
49.5 � 0.3 46.6 � 0.2 44.8 � 0.1 43.0 � 0.5
32.5 � 0.3 30.0 � 0.2 28.5 � 0.1 27.0 � 0.5
further increase. It should also be noted that the intensity of the ε0 -peak in S1 and S2 was relatively low that it is very difficult to distinguish its presence by any characterization technique other than high energy X-ray diffraction. Characteristic ε0 peak at 19.7 nm 1 (Q) corresponds to 27� (2ϴ) in conventional X-ray (CuKα). It is likely that Fig. 1a and b have an ε0 peak at 27� (2ϴ), but it is not possible to detect due to its low intensity. High energy X-ray diffraction patterns of the alloys S1, S2, S3 and S4 between Q values of 24.5–33.0 nm 1 is given in Fig. 4. In alloys S1 and S2, ferromagnetic τ-phase forms with increasing temperature (Fig. 4a and b); however, in alloys S3 and S4 (Fig. 4c and d) the τ-phase peaks were not observed (Fig. 4c and d). The change in transformation scheme indicates the significance of the ε0 -phase in terms of τ phase formation. When ε0 -phase was absent at the quenched state, as in the case of S3 and S4 alloys, τ-phase cannot be obtained. It is likely that ε0 -phase serve as nucleation sites for the τ-phase. Fig. 5 shows the high energy X-ray diffraction patterns of the alloys S1, S2, S3 and S4 between 10.0 and 24.0 nm 1 (Q). Stable phases were formed in all samples upon heating independent of composition. It is likely that the stable phases arise from a matrix which contains ε and ε0 phases. Fig. 5 indicates that the transformation temperatures vary with the 3
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Fig. 2. 2-D film synchrotron plots of (a) S1 (b) S2, (c) S3, and (d) S4. Red lines indicate the τ phase peak positions.
Fig. 3. High energy X-ray diffraction patterns of the alloys (a) S1, (b) S2, (c) S3 and (d) S4 within 18.2–21.2 nm different temperatures.
of ε’. Heating path of S1 and S2 alloys suggests an overlap of the main peak with a shoulder peak appearing at 650 K (377 � C). The shoulder peak is marked with an arrow on Fig. 6. Correlating with the synchrotron data (Fig. 4a and b), it can be speculated that the shoulder peak in S1 and S2 specimens correspond to the formation of τ-phase. On the other hand,
1
(Q) range showing the evolution of εꞌ peak at
DSC heating curve of S3 and S4 was found to display a sharp single peak leaning towards lower temperatures. The shoulder peak is absent in S3 and S4. The remaining transition can be attributed to the formation of stable phases. This finding is in good agreement with the X-ray diffraction data presented in Fig. 5g and d indicating that only stable phases are formed and τ-phase is missing. Onset temperature of the 4
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Fig. 4. High energy X-ray diffraction patterns of the alloys (a) S1, (b) S2, (c) S3 and (d) S4 within 24.5–33.0 nm upon heating.
1
Fig. 5. High energy X-ray diffraction patterns of the alloys (a) S1, (b) S2, (c) S3 and (d) S4 within 10.0–24.0 nm phases upon heating. 5
(Q) range showing the formation of τ-MnAl peak
1
(Q) range showing the formation of the stable
A.M. Genc et al.
Intermetallics 115 (2019) 106617
phases. As expected, alloys S1 and S2 showed ferromagnetic behavior due to the presence of τ-phase. Coercivity and remanence of S1 was found to be 640 Oe and 23 emu/g, respectively. S2 had relatively stronger magnetic behavior having a coercivity of 1560 Oe and rema nence of 26 emu/g. With further increase of Mn, ε0 -phase becomes ab sent thus, τ-phase cannot be obtained. Alloys S3 and S4 were paramagnetic in nature as τ-phase is not observed. In order to understand and visualize the temperature effects on ε→τ transformation, in-situ heating TEM experiments were carried out. Still frames extracted from the dynamic video sequence during in-situ heat ing TEM experiment of sample S2 are presented in Fig. 8. A dense pattern of planar defects was observed at room temperature. Electron diffraction patterns regarding to ε and εꞌ regions couldn’t be collected because of the camera positions configured in the TEM used in this study. However, based on our DSC and XRD analyses, it is hypothesized that planar defects may possess ε0 -phase, so ε and εꞌ-phases coexist in the same matrix. It should be noted that εꞌ regions may be nanosized, therefore, the arrows shown in Fig. 8 is intended for demonstration of possible phase locations. Further high resolution TEM and electron diffraction analyses should be conducted to monitor εꞌ regions. As temperature increases, defect arrays (ε0 ) become denser and eventually transforms into morphologically plate-like τ-phase at 390 � C. Upon further annealing, stable phases readily form at around 520 � C. At 550 � C, sample mostly consists of stable phases along with some τ-phase and trace amount of ε-phase residuals. Same experimental procedure was carried out for sample S4. Screenshots extracted from the dynamic video sequence during in-situ heating TEM experiment of sample S4 are shown in Fig. 9. Planar defect arrays observed at room temperature in S2 were missing in case of S4. In good agreement with XRD results, τ-MnAl formation was not observed and ε-phase directly transformed into stable phases. The transformation temperature found to be lower than S2. Stable phase transformation started at 520 � C in S2 whereas, the value dropped to 460 � C for S4 alloy. This finding is consistent with the synchrotron and DSC analyses suggesting that ε0 -phase retards the stable phase forma tion. At 550 � C, whole sample was decomposed into β-Mn and γ2. Previous studies indicated that ε→τ transformation may occur either by massive mode, shear mode or a combination of both (hybrid). Studies suggesting massive transformation claim that shear mechanism is invalid, ε→τ transformation is independent of the ε0 ordering and it can proceed without the ε0 -phase [17,21–23,27]. However, the synchrotron data presented in this work shows ε0 -phase is a strong prerequisite for
Table 2 Critical transformation temperatures of ε, ε0 , τ and stable phases for alloys S1, S2, S3 and S4. Sample Name
Phase Type and Range
ε
ε0
τ
Stable Phases
S1
300–741 K (27–468 � C)
634–773 K (361–500 � C)
695–773 K (422–500 � C)
S2
300–740 K (27–467 � C)
643–773 K (370–500 � C)
700–773 K (427–500 � C)
S3
300–729 K (27–456 � C) 300–729 K (27–456 � C)
300-633a724 K (27–360a451 � C) 300-684a733 K (27411-460 � C) –
–
635–773 K (362–500 � C) 639–773 K (366–500 � C)
S4 a
–
–
Temperature at which ε0 peak starts to intensify.
Fig. 6. DSC heating curves of S1, S2, S3 and S4 with a heating rate of 10 K/min from room temperature to 773 K (500 � C).
major peaks in S1, S2, S3, and S4 were calculated as 637 K (364 � C), 636 K (363 � C), 627 K (354 � C), and 621 K (348 � C), respectively. Magnetic hysteresis curves obtained for the samples after DSC analysis are given in Fig. 7. Alloys S1, S2, S3 and S4 were heated from room temperature to 773 K (500 � C) at 10 K/min under N2 gas atmo sphere and quenched subsequently. Magnetic properties of those sam ples were measured by vibrating sample magnetometer (VSM). According to HEXRD (Fig. 4a and b). S1 and S2 have ferromagnetic τ-phase and stable phases while S3 and S4 should have only stable
Fig. 8. Screenshots extracted from the dynamic video sequence during in-situ heating TEM experiment of S2 at temperatures of (a)30, (b) 390, (c) 520 and (d) 550 � C. Dashed lines indicate the grain boundary between ε and β-phases.
Fig. 7. Hysteresis curves of S1, S2, S3 and S4, inset showing the second quadrant of the B–H curve of S1 and S2 specimens. 6
A.M. Genc et al.
Intermetallics 115 (2019) 106617
Declaration of competing interest 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. Acknowledgement In-situ high energy X-ray experiments in this research was under taken on the BL04-MSPD beamline at the ALBA Synchrotron facility, Spain. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.intermet.2019.106617. References [1] J.M.D. Coey, Permanent magnet applications, J. Magn. Magn. Mater. 248 (2002) 441–456, https://doi.org/10.1016/S0304-8853(02)00335-9. [2] J.M.D. Coey, Permanent magnets: plugging the gap, Scr. Mater. 67 (2012) 524–529, https://doi.org/10.1016/j.scriptamat.2012.04.036. [3] T. Ohtani, N. Kato, S. Kojima, K. Kojima, Y. Sakamoto, I. Konno, M. Tsukahara, T. Kubo, Magnetic properties of Mn-Al-C permanent magnet alloys, IEEE Trans. Magn. 13 (1977) 1328–1330, https://doi.org/10.1109/TMAG.1977.1059574. [4] Q. Zeng, I. Baker, J.B. Cui, Z.C. Yan, Structural and magnetic properties of nanostructured Mn–Al–C magnetic materials, J. Magn. Magn. Mater. 308 (2007) 214–226, https://doi.org/10.1016/j.jmmm.2006.05.032. [5] A. Chaturvedi, R. Yaqub, I. Baker, Microstructure and magnetic properties of bulk nanocrystalline MnAl, Metals (Basel) 4 (2014) 20–27, https://doi.org/10.3390/ met4010020. [6] H. K€ ono, On the ferromagnetic phase in MnAl system, J. Phys. Soc. Japan. 13 (1958) 1444–1451. [7] A. Yamaguchi, Y. Tanaka, K. Yanagimoto, I. Sakaguchi, N. Kato, Development of P/ M Mn-Al-C alloy permanent magnet, Bull. Jpn. Inst. Met. 28 (1989) 422. [8] E. Fazakas, L.K. Varga, F. Mazaleyrat, Preparation of nanocrystalline Mn–Al–C magnets by melt spinning and subsequent heat treatments, J. Alloy. Comp. 434–435 (2007) 611–613, https://doi.org/10.1016/j.jallcom.2006.08.313. [9] I. Janotova, P.S. Svec, P. Svec, I. Mat’ko, D. Janickovic, J. Zigo, M. Mihalkovic, J. Marcin, I. Skorvanek, Phase analysis and structure of rapidly quenched Al-Mn systems, J. Alloy. Comp. 707 (2017) 137–141, https://doi.org/10.1016/j. jallcom.2016.11.171. [10] J.Y. Law, J. Rial, M. Villanueva, N. Lopez, J. Camarero, L.G. Marshall, J. S. Blazquez, J.M. Borrego, V. Franco, A. Conde, L.H. Lewis, A. Bollero, Study of phases evolution in high-coercive MnAl powders obtained through short milling time of gas-atomized particles, J. Alloy. Comp. 712 (2017) 373–378, https://doi. org/10.1016/j.jallcom.2017.04.038. [11] T.B. Massalski, Binary Alloy Phase Diagrams, 1990. http://www.amazon.com/Bina ry-Alloy-Phase-Diagrams-Massalski/dp/087170403X. [12] N. Singh, V. Mudgil, K. Anand, A.K. Srivastava, R.K. Kotnala, A. Dhar, Influence of processing on structure property correlations in τ-MnAl rare-earth free permanent magnet material, J. Alloy. Comp. 633 (2015) 401–407, https://doi.org/10.1016/j. jallcom.2015.02.041. [13] B. Foreback, I. Baker, Permanent Magnets from Mechanically Milled Mn54Al46 Alloys, 2008. [14] Q. Zeng, I. Baker, Z.C. Yan, Nanostructured Mn-Al permanent magnets produced by mechanical milling, J. Appl. Phys. 99 (2006) 10–13, https://doi.org/10.1063/ 1.2159187. [15] F. Jim� enez-Villacorta, J. Marion, J. Oldham, M. Daniil, M. Willard, L. Lewis, Magnetism-Structure correlations during the ε→τ transformation in rapidlysolidified MnAl nanostructured alloys, Metals (Basel) 4 (2014) 8–19, https://doi. org/10.3390/met4010008. [16] J. Van Landuyt, G. Van Tendeloo, J.J. Van Den Broek, H. Donkersloot, Permanent magnetism and microstructure in τ-AlMn(C), J. Magn. Magn. Mater. 15–18 (1980) 1451–1452. [17] Y.J. Kim, J.H. Perepezko, The thermodynamics and competitive kinetics of metastable τ phase development in MnAl-base alloys, Mater. Sci. Eng. A 163 (1993) 127–134. [18] D.C. Crew, P.G. McCormick, R. Street, MnAl and MnAlC permanent magnets produced by mechanical alloying, Scr. Metall. Mater. 32 (1995) 315–318, https:// doi.org/10.1016/S0956-716X(99)80057-0. [19] P. Muellner, B.E. Buergler, H. Heinrich, A.S. Sologubenko, G. Kostorz, Observation of the shear mode of the ε→τ phase transformation in a Mn-Al-C single crystal, Philos. Mag. Lett. 82 (2002) 71–79, https://doi.org/10.1080/ 09500830110103225. [20] J.J.V.D.B. Broek, H. Donkersloot, Phase transformations in pure and carbon-doped Al45Mn55 alloys, Acta Metall. 27 (1979) 1497–1504. https://doi. org/10.1016/0001-6160(79)90172-X.
Fig. 9. Screenshots extracted from the dynamic video sequence during in-situ heating TEM experiment of S4 at temperatures of (a)30, (b) 400, (c) 460 and (d) 550 � C. Dashed lines indicate the grain boundary between ε and β-phases.
τ-phase formation. Alloys having less Mn (S1 and S2) already had ε0 -phase within the ε-matrix at room temperature. It was observed that ε-phase first transforms into the ordered ε0 -phase and then τ-phase was formed via martensitic shear of the ε’. Upon annealing, ε0 readily transforms into ferromagnetic τ-phase. This finding is in good correla tion with shear [3,16–20] and hybrid mode studies [25] which state that
τ-phase was generated due to the direct shear of the ordered ε0 -phase
regions. Through in-situ TEM experiments, it is speculated that the evolution of the τ-MnAl phase involves a nucleation and growth process wherein, the nucleation takes place heterogeneously along the ε0 phase grain boundaries. Grain boundaries with high interfacial energy and relatively weak bonding are often the preferred nucleation sites for heterogeneous nucleation. 4. Conclusions
Based upon the results obtained from in-situ high energy X-ray diffraction as well as differential scanning calorimetry, magnetic mea surements and in-situ TEM analyses of MnAl-based alloy, phase trans formation hierarchy of the compositions that lead to the formation of metastable ferromagnetic τ-phase have been developed. Overall results indicate that the composition of MnAl-based alloys affect both the sequence and the temperature of phase transformations. Alloys having Mn content less than 53.4 at% (S1, S2) has a parent phase mixture consisting of ε and ε0 phases at the quenched state. Upon annealing, growth of ε0 -phase was followed by a phase transformation into ferro magnetic τ-phase and stable phases. Alloys which have more than 53.4 at% of Mn (S3, S4) consist of only ε-phase at the quenched state. ε-phase directly transforms into stable phases with an increase in temperature. High energy X-ray diffraction patterns and DSC scans both confirmed that ε0 -phase was missing and τ-MnAl formation was surpassed. Alloys having τ-phase were found to be ferromagnetic in VSM. This finding suggest that magnetic τ-MnAl formation cannot proceed in the absence of ε0 -phase. In situ TEM imaging confirmed the transformation sequence of εþε’→τ→stable phases for S2 alloy. ε and ε0 -phases were observed at room temperature and transformed into τ-phase upon annealing. Stable phases followed τ-phase formation spontaneously. It is observed that compositional ordering do not take place as Mn content increases as in S4 alloy. Only a certain composition range results in the formation of the ε0 -phase and only that composition range results in the τ-phase forma tion. Mn70% (wt) was found to be the limit for ε0 and τ-phase formation.
7
A.M. Genc et al.
Intermetallics 115 (2019) 106617
[21] D.P. Hoydick, E.J. Palmiere, W.A. Soffa, On the formation of the metastable L1o phase in manganese-aluminum-base permanent magnet materials, Scr. Mater. 36 (1997) 151–156, https://doi.org/10.1016/S1359-6462(96)00356-9. [22] D.P. Hoydick, E.J. Palmiere, W.a. Soffa, Microstructural development in MnAl-base permanent magnet materials: new perspectives, J. Appl. Phys. 81 (1997) 5624, https://doi.org/10.1063/1.364619. [23] C. Yanar, V. Radmilovic, W.A. Soffa, J.M.K. Wiezorek, Evolution of microstructure and defect structure in L10-ordered manganese aluminide permanent magnet alloys, Intermetallics 9 (2001) 949–954, https://doi.org/10.1016/S0966-9795(01) 00095-4. [24] C. Yanar, J.M.K. Wiezorek, V. Radmilovic, W.A. Soffa, Massive transformation and the formation of the ferromagnetic L10 phase in manganese-aluminum-based
alloys, Metall. Mater. Trans. A 33A (2002) 2002–2413, https://doi.org/10.1007/ s11661-002-0363-3. [25] J.M.K. Wiezorek, A.K. Kulovits, C. Yanar, W.A. Soffa, Grain boundary mediated displacive-diffusional formation of τ-phase MnAl, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 42 (2011) 594–604, https://doi.org/10.1007/s11661-0100308-1. [26] H. Fang, J. Cedervall, F.J.M. Casado, Z. Matej, J. Bednarcik, J. Ångstr€ om, P. Berastegui, M. Sahlberg, Insights into formation and stability of τ-MnAlZx (Z ¼ C and B), J. Alloy. Comp. 692 (2017) 198–203, https://doi.org/10.1016/j. jallcom.2016.09.047. [27] V.M. Gundyrev, M.A. Uimin, A.E. Ermakov, O.B. Andreeva, Transformation in MnAl-C alloys, Phys. Status Solidi. 91 (1985) K55–K58, https://doi.org/10.1002/ pssa.2210910152.
8