Unprecedentedly Wide Curie‐Temperature Windows as Phase‐Transition Design Platform for Tunable Magneto‐Multifunctional Materials

Abstract

A series of unprecedentedly wide Curie-temperature windows (CTWs) between 40 and 450 K are realized by employing the isostructural alloying principle for the strongly coupled magnetostructural phase transitions in a single host system. The CTWs provide a design platform for magneto-multifunctional multiferroic alloys that can be manipulated in a quite large temperature space in various scales and patterns, as well as by multiple physical fields. A coupling of ferroelasticity and ferromagnetism1, 2 can lead to a multiferroic behavior of first-order magnetostructural phase transition (MST) in magnetoelastic materials. Attractive physical effects, such as ferromagnetic shape memory,3 magnetostrain,4 magnetocaloric effect (MCE),5, 6 magnetoresistance,7 and exchange bias,8 are observed based on the MSTs. These effects are receiving increasing attentions from the applications in actuating,9 sensing,10 magnetic cooling,11 heat pump,12 and energy conversion.13 As an important class of MSTs, ferromagnetic martensitic transformations (FMMTs) are widely found in Heusler, Fe-based, and MM′X alloys, and produce diversiform physical discontinuities due to the significant alterations in crystallographic, electronic, orbital, and magnetic structures in the systems. Extraordinary magneto-multifunctional properties emerge thereby and can be tuned in different ways. Profiting from FMMTs, the magnetic-field-induced shape memory effect and giant magnetostrain/magnetostriction have been extensively studied with promising potential in micromechanical controls and strain outputs. The MCE,14 which happens in a magnetic transition, can be enhanced appreciably by first-order FMMTs with great changes in structural entropy in spin-lattice coupled systems.15-17 Materials bearing FMMTs are thus further considered as candidates for caloric applications, probably combined with mechanocaloric or electrocaloric effects.18, 19 Very recently, an inspiring application of electric power generation,20, 21 using first-order FMMT ferromagnets driven by high-temperature heat resources, becomes increasingly attractive for the energy conversion, which demonstrates an exciting advance in magneto-multifunctionalities of magnetoelastic alloys. For all the magneto-multifunctionalities, the magnetostructural coupling plays a fundamental role, since only in this case can the structural and magnetic transitions modulate each other. In order to approach FMMTs, the martensitic transformations should happen below the Curie temperature (TC) of the system so that they couple with magnetic state changes. Along with the increasing demands of diversified functional applications, searching for new materials with highly tunable FMMTs, especially in a much wider working temperature range, is ever growing. However, challenges exist in practice. One challenge is to enlarge the magnetization change (ΔM) across FMMTs to maximize the magnetic-energy change especially in a moderate magnetic field. The other challenge is to broaden the temperature range where the FMMTs can take place, similar to the case in magnetocaloric materials Gd-Si-Ge between 20 and 290 K,23 or the case in Mn-Fe-P-As between 150 and 350 K.24 For first-order FMMTs, the structural transitions are often limited in a temperature range between just above room temperature and liquid-nitrogen temperature. If TC of an FMMT material could be tailored to higher temperatures, the temperature space in which magnetostructural coupling can occur will be expanded largely. As a rising family of FMMT materials, the hexagonal MM′X (M, M′ = transition metals, X = carbon or boron group elements) compounds have been intensively studied in recent years.25-27 The FMMTs in these compounds are characterized by small thermal hysteresis, high temperature-sensitivity, high Curie-temperatures, gigantic anisotropic strains, huge volume expansions, large magnetization jump, magnetic-field-induced shape memory effects, and giant MCEs,26, 28-30 which collectively make MM′X compounds rather different from the conventional FMMT materials (for example, Heusler alloys). Many methods, including physical pressure, chemical modification, atom-vacancy introducing, and quench-relaxation annealing, have been adopted to tune the FMMTs. In our previous works,26, 29 we proposed an isostructural alloying principle to guide the chemical substitution during the material design. On this principle, we alloyed two isostructural compounds that have the same crystal structure but distinct phase stability and different magnetic behaviors together to form a new compound, by which we were able to manipulate both the phase transition and the magnetic exchange interactions simultaneously in a material host. Recently, this effective principle has generated active influences on design and realization of the desired MST materials.30-36 In the light of this principle, in our previous work26 we have alloyed MnNiGe with isostructural FeNiGe (Mn1-yFeyNiGe) and established a stable magnetostructural coupling in a broad Curie-temperature window (CTW), as illustrated in left diagram in Scheme 1, in which tunable large MCEs have been obtained. Upon increasing Fe content (y), Tt of FMMT was efficiently lowered below TCM of martensite (point B in Scheme 1) and was continuously decreased down to the magnetic-frozen temperature (TgA, point C in Scheme 1) of the spin-glass-like austenite. Simultaneously, for martensite phase introducing of Fe atom gradually converted the spirally antiferromagnetic (AFM) coupling to ferromagnetic (FM) one, with TNM becoming TCM, and finally resulted in a high magnetization when y > 0.22 (near point C at y1 = 0.26, deep blue region in the left diagram in Scheme 1). As a result, a CTW was established between TCM ≈ 350 K and TgA ≈ 70 K. In the CTW, the strong ferromagnetism of martensite, featured by decreasing saturation field (HS) and increasing saturation magnetization (MS), shows a positive correlation with Fe content. This work26 provides a fundamental understanding on the tuning of magnetostructural coupling under the isostructural alloying principle and opens up a way to explore large CTWs for strongly coupled MST materials. In this work, in line with the important success mentioned above, we design new material systems step by step by taking the strategy shown in Scheme 1. If one keeps the strong ferromagnetism around point C, at the same time tailors the transition (Tt) to high temperatures, one may obtain a series of strongly coupled MSTs with low-field effects and large ΔM. TCM, which determines the upper critical temperature (Tcrup, point D) of the CTW, should be simultaneously raised from point B′ to high temperatures. By achieving this, abundant magneto-multifunctional applications can be expanded and promoted freely over a very broad temperature range from cryogenic to high temperatures. According to the isostructural alloying principle, to realize the above purposes the isostructural counterpart to Mn1-yFeyNiGe system must meet three conditions: (1) A high Tt, which allows us to tailor the FMMT (Tt) of the alloyed system from low to high temperatures; (2) A high TCM, to upraise the Tcrup of CTW; (3) A strong ferromagnetic coupling in martensite phase, to gain large ΔM, low HS and desired magneto-multifunctional effects. Among Ni2In-type hexagonal compounds, MnNiSi can be identified as a promising isostructural counterpart. Both its Tt and TCM are as high as 1200 and 600 K, respectively.37, 38 Its martensite phase is a strong ferromagnet with a low HS and a large MS. The high Tt hopefully drives up the FMMT from point C (Scheme 1) continuously to high temperatures with increasing alloying level of Si content (x) (red line with dual-arrows, Scheme 1). Similarly, TCM may be upraised as well. It is desirable that Tt and TCM curves would cross over at high temperatures, resulting in a brand-new CTW between two far-apart points C and D. After determining the isostructural counterpart MnNiSi, Mn0.74Fe0.26NiGe was taken as the first starting alloy. For MnNiSi, we just alloyed it with Mn0.74Fe0.26NiGe by simply substituting Ge for Si atoms. An isostructural system of Mn0.74Fe0.26NiGe1-xSix was created. Although the FMMTs vanish in Mn1-yFeyNiGe compounds when y > 0.26 (for example, y = 0.36, 0.46, 0.55), these compositions were further taken as our starting alloys, with a consideration that an intrinsic strong ferromagnetism can be expected owing to the ferromagnetic coupling between Fe–Mn atoms in the martensite form.26 In this work, dual-variable Mn1-yFeyNiGe1-xSix (y = 0.26, 0.36, 0.46, 0.55; 0 ≤ x ≤ 1) systems were studied systematically. A crystallographic and magnetic phase diagram was proposed in Figure 1a,b, based on the data from the structural (x-ray diffraction (XRD)), magnetic (M(T) curve), and thermal (differential scanning calorimetry (DSC)) measurements (more data in Supporting Information, Figures S1, S3, S5, and S6). The alloy series of (y = 0.26, x) is taken as an example (Figure 1a). Upon introducing Si atoms on Ge sites, Tt begins to increase from low temperature of 74 K for x = 0 to high temperature up to 1000 K for x = 1.0, which is very close to Tt (≈1200 K) of stoichiometric MnNiSi.37 Meanwhile, TCM is also increased, indicating Si substitution results in a significant enhancement in magnetic exchange interactions between Mn/Fe–Mn/Fe atoms in martensite phase. One can further see that the slope of Tt-(Si) x curve is much larger than that of TCM-(Si) x one. This implies that the Si substitution imposes more influence on the structural transition than on the magnetic coupling. The crossover point D mentioned in Scheme 1 appears at 434 K, which is much higher than TCM (350 K) of starting alloy Mn1-yFeyNiGe.26 For higher Fe contents (y = 0.36, 0.46, 0.55), a similar behavior is observed. Both Tt and TCM are simultaneously raised to high temperatures simply by Si substitution and the crossover points are thus obtained above 400 K. An amplified phase diagram is depicted in Figure 1b. For the alloy series of (y = 0.26, x), a CTW with a width of 360 K between the lower critical temperature (Tcrdn) = 74 K and Tcrup = 434 K is obtained. From Figure 1b, one sees that Tcrup (crossover point) goes along a kappa-like curve with a maximum value 448 K in (y = 0.46, x) series, while with increasing Si content Tcrdn further decreases from 74 K to about 38 K. These Tcrdn, below which the displacements of alloy atoms will be frozen, are the lowest transition temperatures reported in first-order FMMT systems. As an important result, a series of unprecedentedly wide CTWs with a maximal width of 405 K between Tcrdn and Tcrup are obtained in Mn1-yFeyNiGe1-xSix system (also see inset to Figure 1b). These CTWs span from below liquid nitrogen temperature (≈38 K) to room temperature and to the highest temperature of 448 K, which are much wider than the temperature distribution ranges of many other FMMTs.17, 22-24, 27-31, 33-36 This is the first realization that highly tunable MSTs can be gained in such wide temperature range in one material system. A strong magnetostructural coupling of ferroelasticity and ferromagnetism in Mn1-yFeyNiGe1-xSix system is thus obtained over a quite large range, connecting the ultralow and the high temperatures. In Figure 1b, between the window boundaries (Tcrdn-Tcrup) (the light-yellow region), the strongly coupled MSTs can be obtained in any composition point with random (y, x) values. The MST behaviors within the CTWs are presented by magnetization measurements. Figure 1c shows the thermomagnetic M(T) curves of (y = 0.36, x) alloy series (more data in Supporting Information, Figure S3). The M(T) curves with abrupt magnetization jump from paramagnetic (PM) state to FM one, and clear but small thermal hysteresis demonstrate the first-order FMMT behavior between 0.15 ≤ x ≤ 0.60. During the Si substitution, the FMMT can be tailored from 51 K to high temperatures, showing a high tunability. These transitions involve two structures, Ni2In-type hexagonal parent phase and TiNiSi-type orthorhombic martensite, as illustrated in Figure 1d (see structural analysis by XRD in Supporting Information, Figure S1). The FMMT occurs from the parent phase to the martensite via distortions of Ni-Ge/Si hexagonal rings and Mn/Fe–Mn/Fe zigzag chains, with a giant volume expansion of around 3% (more data in Supporting Information, Figure S2 and Table S1). Based on this structural transition, a high temperature-sensitivity and a large magnetization jump can be observed. Furthermore, very high magnetizations are seen even in a low field of 1 kOe. Large ΔMs across the FMMTs up to 65 emu g−1 are measured in the CTW, as shown in Figure 1e (more data in Supporting Information, Table S2). These ideal features will promote the magneto-multifunctional properties such as giant MCEs and electric power generation at low fields with low energy consumptions, especially in a wide temperature range including the important high-temperature region. For higher Si substitution with x = 0.65 and 0.70, the thermal hysteresis becomes zero in the M(T) curves as the MST decouples above the upper critical temperature (Tcrup = 448 K). In order to understand the origin of ferromagnetism in martensite phase, the magnetizing behaviors at 5 K of Mn1-yFeyNiGe1-xSix were analyzed. Here we first look back the magnetic properties of Mn1-yFeyNiGe.26 As shown in the left part of Figure 1e, MS increases to a plateau with a value of about 2.65 μB upon increasing Fe content (y). Fe substitution can efficiently convert AFM couplings in MnNiGe to FM ones. HS decreases monotonically and reaches a minimum of 5 kOe at y = 0.24 where the AFM coupling is almost overcome by FM couplings. In Mn1-yFeyNiGe system, the high magnetizations and low saturation fields are achieved gradually with increasing Fe substitution. After introducing Si at Ge site, in sharp contrast, the Mn1-yFeyNiGe1-xSix (y = 0.26, 0.36, 0.46, 0.55) coheres the lowest HS and highest MS of Mn1-yFeyNiGe system and maintains them at values of HS ≈ 1.35 to 2.12 kOe and MS ≈ 2.2 to 2.6 μB. M(H) curves of (y = 0.26, x) (0.10 ≤ x ≤ 0.60) series are given in Figure 1f (more data in Supporting Information, Figure S4). One can see all samples are easily magnetized in a low magnetic field and gain large saturation magnetizations of about 80 emu g−1 (2.6 μB). The features of low HS and high MS in Mn1-yFeyNiGe1-xSix are highly expected to facilitate the magneto-multifunctional properties, especially the desired low-field effects. The magnetic structure in martensite phase of Mn1-yFeyNiGe1-xSix system was further revealed by first-principles calculations. Figure 1g shows the partial density of states (PDOS) of 3d-metal atoms in (y = 0.5, x = 0.5) alloy, from which one sees the remarkable spin polarizations on both Mn and Fe atoms. This accounts for the large magnetic moments of ≈2.8 μB and ≈1.8 μB on Mn and Fe atoms, respectively (Supporting Information, Table S3). Compared with Mn atom, Fe atom has a weaker polarization and a clear PDOS peak exists well below the Fermi level in spin-down state, which results in a smaller moment on Fe atom. Ni and Ge/Si atoms carry near-zero moments (Supporting Information, Table S3) as the spin polarizations are very weak due to strong covalent bonds between Ni–Si/Ge atoms. Figure 1h depicts the spin electron density of sample (y = 0.5, x = 0.5), which shows the magnetization distribution in the compound. Strong localization of (positive) spin electron density values exist around both Mn/Fe atoms, indicating significant localized moments with a parallel alignment in the zigzag chains. In Mn1-yFeyNiGe1-xSix, the magnetizations, mainly originating from Mn and Fe moments with FM exchange interactions, keep high and stable values during the whole Si substitution. In this section, we present the desired magnetic functional properties in CTWs. Isothermal M(H) curves across the FMMT of sample (y = 0.26, x = 0.30) were measured (Figure 2a). As illustrated, a metamagnetic behavior is observed between K with a critical field of about kOe and K with a rather low of 5 kOe. This metamagnetic behavior the magnetic-field-induced martensitic structural transition from hexagonal parent phase to an FM orthorhombic further characterized the MCEs upon this The magnetic entropy changes of (y = 0.26, x = 0.30) alloy were from the M(H) curves using (see the As shown in Figure a value of = was obtained at a field change = kOe and = K, showing a giant conventional the alloy shows a giant of at moderate = 20 kOe and = K. These giant MCEs are from the abrupt and large magnetization changes during the FMMTs as well as the low HS as the of on the we alloys with (y = 0.36, x = 0.5, in the CTW to the M(H) curves and the as shown in Figure It is that on temperature very in the temperature range between and K, and a value of at = and a value of at moderate = 20 kOe. giant MCEs are obtained in these materials. large MCEs are found above point (see It is important to that these giant MCEs the whole CTWs are of the caloric effects from both the structural and magnetic transitions and more two caloric effects are in the same owing to the in crystallographic and magnetic This effect would the caloric of strongly coupled 29 present the of maximum of on the we values of Figure in Figure with the A between maximum of and is observed up to = kOe. This that the temperature of magnetization across the FMMTs is field which may be to the of the transitions in these materials. The data in Figure can be to a = where the can be considered as the that strong the maximum of on The values of obtained from are in Figure as a of with the that the of a magnetic The high values of and obtained on (y = 0.36, x) series as shown in Figure the large MCEs the wide CTWs. As well the functional behavior is critical for In order to the on the FMMTs, thermal were on for compositions of (y = 0.26, x = and (y = 0.36, x = Figure shows curves of alloy (y = 0.26, x = peak on martensitic During in this the transition were in the peak and peak which demonstrate that the phase transitions and The same were on other two alloys (y = 0.26, x = and (y = 0.36, x = The of Tt, and are presented in Figure (more data in Supporting Information, Table S4). For the transitions at low K) and room K) temperatures (y = 0.26, x = Tt keeps almost the whole with a small than K (for sample with x = the of Tt ≈ K indicating an even functional stability than the This stability is further by a in by a saturation for the transitions at K in alloy (y = 0.36, x = only very changes of Tt K) and can be observed. These changes may be to the atom by high-temperature thermal It is clear that the studied alloys a functional stability even at high temperatures, which will the applications of the materials. Figure shows a of the maximum entropy changes of our materials and materials based on data (Supporting Information, Table The is in with the one reported by In this and MCEs are observed in many and compounds, including the first-order FMMT materials such as Heusler and MM′X alloys. However, the MCEs in each family of material are limited in temperature since many MSTs can only be tuned in a limited temperature It can be seen that different transition temperatures but giant MCEs occur below K. In the MCEs of MM′X alloys also exist mainly in a range around room temperature. In sharp contrast, our Mn1-yFeyNiGe1-xSix single material system shows giant MCEs over the very wide temperature range from 40 to 450 K. these wide the applications in of or can also be It is further that large MCEs and effects tuned by and MCEs by can be freely at any desired temperature within such broad In to the as magnetic the giant materials can also be as promising working for magnetic heat or electric power 20 For these applications, the MSTs are to occur at high temperatures in order to gain the efficiently and from driven by or from or or by the from applications of and For the high-temperature MSTs between and 448 K, the materials an increasingly low due to the for Ge In the principle of isostructural we have realized the strongly coupled magnetostructural transitions within a series of unprecedentedly wide Curie-temperature with as large as 400 K, in a single host system This clear physical of the tuning in material design and the Curie-temperature windows for phase The of low-field large caloric effects, wide working and functional stability the materials promising for various applications including and energy The unprecedentedly wide Curie-temperature windows provide a broad design platform of magnetostructural transitions for tunable magneto-multifunctional properties of the multiferroic magnetoelastic alloys. The strongly coupled magnetostructural transitions can be manipulated in of and for potential applications, including 51 functional and multiferroic were under was for and was over each were in with at K for 5 and down to room temperature XRD analysis were on a with at room temperature. measurements including thermomagnetic curves and magnetization curve were on a and physical system curves with temperature above 400 K were on a sample with a maximum field of kOe. analysis was using scanning calorimetry and The is 5 K By an two the change by magnetic transition can be on the The magnetic entropy changes across magnetostructural transitions were from the magnetization curves using the The is as where and are two temperatures of the maximum of the temperature was to the magnetization curves with an of 2 K for these FMMTs with a clear thermal The first-principles on density of states magnetic moment and spin electron density were using based on the density The exchange and correlation were using the in the and in the were for a of the The for the field was at 1 atom. This was by of and of and and of of This was on to the of on As a to our and this provides by the materials are and may be for but are or from than should be to the The is for the content or of any by the than should be to the for the