Robust Bain distortion in the premartensite phase of a platinum-substituted Ni2MnGa magnetic shape memory alloy

The premartensite phase of shape memory and magnetic shape memory alloys (MSMAs) is believed to be a precursor state of the martensite phase with preserved austenite phase symmetry. The thermodynamic stability of the premartensite phase and its relation to the martensitic phase is still an unresolved issue, even though it is critical to the understanding of the functional properties of MSMAs. We present here unambiguous evidence for macroscopic symmetry breaking leading to robust Bain distortion in the premartensite phase of 10% Pt-substituted Ni2MnGa. We show that the robust Bain-distorted premartensite (T2) phase results from another premartensite (T1) phase with preserved cubic-like symmetry through an isostructural phase transition. The T2 phase finally transforms to the martensite phase with additional Bain distortion on further cooling. Our results demonstrate that the premartensite phase should not be considered as a precursor state with the preserved symmetry of the cubic austenite phase.

Thirdly, it is not clear what is the broader perspective. The authors have failed to convincingly describe the role of Pt in stabilizing the premartensite phase. What is the insight, with regard to alloy design, that is being brought. For example, what other alloying elements can have a similar effect.
Finally, the authors have stated in the abstract that the results gives better understanding of functional properties of the alloy. This is not clear. What properties are being described and what role the premartensite phase plays in them? Therefore, while the result for an additional premartensite phase with stable Bain distortion in itself is interesting, the underlying physics behind it has not been described adequately and the broader implication for this is not clear. In view of this, I regret I cannot recommend publication of the article in Nature Communications in its present form.

Reviewer #2 (Remarks to the Author)
The manuscript presents a detailed study of the sequence of transformations in a magnetic shape memory alloy of Ni2MnGa with 10% of Pt. The authors evidence, by means of high resolution synchrotron X-ray diffraction, the presence of a premartensitic phase with a Bain distortion. Even though the characteristics of the intermediate phase between austenite and martensite has been extensively studied in Ni-Mn-Ga alloys, there is still controversy about the origin of the modulation shown by the low temperature phases. In this sense, the present work shows a clear evidence of a new transformation sequence: austenite--undistorted premartensite---distorted premartensite---martensite, not reported until now. This finding can provide important information about the transformation mechanism and the origin of the martensitic phase modulation. Taking this into account, the paper deserves publication in Nature Communications. In order to be published the following questions should be addressed. In a previous work, ref 28, on the Ni2MnGa system, only the undistorted premartensite phase is observed in the transformation sequence. If the presence of a Bain distorted premartensitic phase is more favorable to maintain an invariant habitat plan, could one expect the presence of this phase in Ni2MnGa even in a very narrow temperature range? Could the observed transformation sequence be generalized to all systems with premartensite phase? Or is it the result of the modification in the magnetoelastic coupling due to the disorder and volume change induced by the Pt addition? The authors describe some differences in the magnetic measurements between NiPtMnGa and Ni2MnGa, i.e. the absence of dip in the susceptibility at the transition temperature of the premartensitic phase and a huge drop in the susceptibility linked to this phase. Could the authors give some explanation to these differences? Taking into account the large susceptibility and volume changes, fig. 1 and fig 6, Why the authors consider type II premartensite and not an intermediate martensitic phase with modulation 3M?

Reviewer #1 (Remarks to the Author):
Comment: The article by Singh et al. reports on the discovery of two premartensite phases in a Ni2MnGa alloy with 10% Pt substitution. The interesting new observation is that there occurs a second premartensite phase with intermediate Bain distortion, specifically for the studied alloy. However, the authors have failed to explain the detailed mechanism of the transformation and also failed to put this information in broader perspective.

Reply:
We appreciate referee's comments about our manuscript "The interesting new observation is that there occurs a second premartensite phase with intermediate Bain distortion". In the following, we have tried our best to provide point-by-point response to the referee's comments/suggestions.

Comment 1:
First, what is the atomic mechanism for the two transformations? What are the different atomic displacements? Are the atomic displacements same or different? This kind of information can truly provide new insights into the physics of the problem. The authors did carry out Rietveld refinement, but these details are not available from their analysis.

Reply:
We have now provided the detailed results of Rietveld refinement (atomic displacement, atomic position, symmetry, lattice parameters etc.) for the two premartensite phases (i) PM (T1), where the cubic symmetry is maintained with no evidence of Bain distortion and (ii) PM (T2), which exhibits robust Bain distortion. The results clearly indicate that PM (T1) to PM (T2) transformation is isostructural type as the crystal symmetry and Wyckoff positions are same for both the phases and only the atomic displacements are different. We have compared the atomic displacement in the two premartensite phases in the revised manuscript in Fig.6 c-e. Details of Rietveld refinement results are provided in the supplementary file of the revised manuscript. Reply: The premartensite phase, which has been extensively studied in Ni-Mn-Ga magnetic shape memory alloys appears in between the parent austenite and martensite phases and has always been considered as a precursor effect with preserved symmetry of the high temperature cubic phase. This is because there is no evidence of symmetry breaking or in other words any evidence for Bain distortion in the premartensite phase. In the present work we have presented direct evidence for robust Bain distortion (i.e., broken symmetry) in the premartensite phase. This shows that premartensite is an independent thermodynamic phase, which further transforms to the martensite phase. Although the modulated premartensite phase has been observed in many alloys, a clear relationship between the modulated premartensite phase and the origin of the modulation of martensite phase is still an open issue. The present work shows clear evidence that Bain distorted martensite phase results from the Bain distorted premartensite phase, which has been never reported earlier. This observation also indicates that adaptive modulations of martensites are not applicable for the present alloy as no direct interface is formed between cubic austenite and tetragonal martensite. Therefore present finding provide clear information about the transformation mechanism and the origin of the modulation in the martensitic phase as soft phonon mode based model. The results of ref. 40 and also Pramanick Phys. Rev. B 85, 144412 (2012) have not addressed this issue of the premartensite phase, which gives direct evidence about the phase transformation sequence and hence the origin of modulation.

Reviewer2 has also mentioned this in his/her comment "Even though the characteristics of the intermediate phase between austenite and martensite has been extensively studied in Ni-Mn-Ga alloys, there is still controversy about the origin of the modulation shown by the low temperature phases. In this sense, the present work shows a clear evidence of a new transformation sequence: austenite--undistorted premartensite---distorted premartensite---martensite, not reported until now. This finding can provide important information about the transformation mechanism and the origin of the martensitic phase modulation."
Comment 3: Thirdly, it is not clear what is the broader perspective. The authors have failed to convincingly describe the role of Pt in stabilizing the premartensite phase. What is the insight, with regard to alloy design, that is being brought. For example, what other alloying elements can have a similar effect.
Reply: To understand the exact mechanism responsible for the stabilization of Bain distorted premartensite in Ni 1.9 Pt 0.1 MnGa, we have performed new ab initio calculations (please see modified Fig.8 and the corresponding text on pages 14-16 in the revised version of the manuscript). Our calculations indicate that volume expansion due to larger Pt atoms is responsible for the Bain distortion in Ni 1.83 Pt 0.17 MnGa (slightly different composition from experiment is chosen to minimize computational effort). We have demonstrated this by calculating the energy as a function of c/a for Ni 1.83 Pt 0.17 MnGa at the lower optimum volume of Ni 2 MnGa. These calculations at the lower volume do not show any Bain distortion for the premartensite phase. To establish the mechanism more clearly we have calculated the premartensite phase of the parent compound Ni 2 MnGa at the larger optimum volume of Ni 1.83 Pt 0.17 MnGa. Our calculations clearly show that a Bain distorted premartensite phase is also stable for Ni 2 MnGa at the larger volume while no indications of Bain distortion can be seen for the parent compound at its equilibrium volume. Further, our investigations on the magnetic properties show that the magnetization in Ni 2 MnGa increase at the larger volume. This indicates that higher magnetization at the larger volume helps in symmetry breaking and producing robust Bain distortion in the premartensite phase due to strong magnetoelastic coupling. Since the volume of Ni 1.83 Pt 0.17 MnGa (Ni 1.9 Pt 0.1 MnGa in experiments) is larger than that of Ni 2 MnGa, similar mechanism of magnetoelastic coupling is also responsible for the Bain distortion of its premartensite phase.
The stability of the premartensite phase is extremely sensitive to alloy composition and external fields such as pressure, temperature and magnetic fields. In the present case, Pt is non-magnetic and has same number of valence electrons as that of Ni. Hence, Pt influenced the transformation process only due to its larger atomic size. Substitution of a different element will most likely change the valence electron concentration, atomic size and possibly magnetic properties. Hence, it is extremely difficult to make a general statement about the effect of any alloying element. While the present study shows the effect of volume, systematic studies in future is required to understand the effect of other alloying elements.
Comment 4: Finally, the authors have stated in the abstract that the results give better understanding of functional properties of the alloy. This is not clear. What properties are being described and what role the premartensite phase plays in them?
Reply: The stoichiometric Ni 2 MnGa is an important system for application as magnetic actuator due to its large magnetic field induced strain (MFIS). This system exhibits not only a large MFIS but also magnetocaloric effect (MCE) and topological spin textures. Besides Ni 2 MnGa, the offstoichiometric Ni-Mn-Ga together with other Ni-Mn-X (X= In, Sn, Sb) based magnetic shape memory alloys (MSMA's) are extensively investigated due to their large MCE. However, in contrast to Ni 2 MnGa, they show a large hysteresis and irreversible behavior, which is a major drawback for their use in practical application [Phys. Rev B 92, 020105, (2015)]. It is interesting to note that in Ni 2 MnGa austenite phase transforms first to the premartensite phase, which further transforms to the modulated martensite phase. On the other hand, the off stoichiometric Ni-Mn-X (X= In, Sn, Sb) directly transform from austenite to the martensite phase without showing any precursor effect [Nature Mater. 4, 450 (2005)]. In the Pt substituted Ni 2 MnGa (studied here), the martensite phase originates from the larger Bain distorted premartensite phase suggesting that the Bain distortion appears in steps to facilitate the emergence of an invariant habit plane. Therefore, the premartensite phase may be able to adapt low inputs of magnetic or elastic energy change without undergoing irreversible effects. However, the main emphasis of this paper is to show that robust Bain distortion is possible in premartensite phase also and the consequence of this in the functional properties needs to be investigated.
In view of the fact that large MFIS is observed in Ni 2 MnGa is intimately linked with the existence of the modulated structure, which exhibits austenite-premartensite(PM)-martensite phase transformation sequence the successive change in modulation wave vector is going from PM (T1)--PM (T2)--martensite is likely to affect the observed MFIS. In the conclusion of manuscript we have mentioned "This basic understanding of the origin of modulation in these alloys provides a pathway to design new MSMAs since the large MFIS in Ni 2 MnGa is intimately linked with the existence of the modulated structure." We have also modified now the related sentence in the abstract as "Our work is a major breakthrough in the field of MSMAs as it gives better understanding of the premartensite phase and its relation to the martensite phase transition, which is necessary for engineering the functional properties of these alloys." However, if the reviewer considers it necessary, we will include an elaborative discussion in the manuscript.
Comment 5: Therefore, while the result for an additional premartensite phase with stable Bain distortion in itself is interesting, the underlying physic behind it has not been described