Breakthrough in magnetism: Evidence for simultaneous structural change at ferromagnetic transition

For more than a century, ferromagnetic transition has been considered to involve only an ordering of magnetic moment, without any change in the host crystal structure or symmetry; thus a cubic paramagnet (like Fe, CoFe₂O₄ and so on) has been considered to transform into a cubic ferromagnet upon ferromagnetic transition. Such a conclusion has been backed by numerous crystallography data by conventional XRD. It has been a foundation of our present-day understanding of ferromagnetism. However, this century-old belief has been challenged by our recent high-resolution experiment, and hence an alteration of this basic belief about ferromagnetic transition becomes inevitable.

Very recently, with high-resolution synchrotron XRD, we show direct evidence for the non-cubic symmetry of several typical “cubic” ferromagnets like CoFe₂O₄ and Tb_0.3Dy_0.7Fe₂ (Fig.1), and a simultaneous structural change at ferromagnetic transition temperature. These facts strongly suggest that ferromagnetic transition is also a structural transition, yielding a low crystallographic symmetry that conforms to the spontaneous magnetization (M_S) direction.

Figure 2 shows the magnitude of lattice distortion in the ferromagnetic state for these typical “cubic” ferromagnets as a function of temperature. Clearly, ferromagnetic transition does involve a structural change like a ferroelectric transition, but the change is usually too small to detect even by synchrotron XRD. This is the reason why for more than a century the ferromagnetic transition had been regarded as involving no change in crystal structure.

The discovery of the structure part of ferromagnetic transition alters the basic view about ferromagnetic transition, and may provide new insight into many issues in ferromagnetism, where magnetism alone cannot solve. Firstly, it unifies the mesoscopic explanation for both magnetostrictive effect in ferromagnetic materials and for the electrostrain effect in ferroelectric materials. Secondly, the structure change upon ferromagnetic transition may provide new clues for developing highly magneto-responsive materials, including the recently reported giant magnetocaloric and magnetoresistive materials. Thirdly, the concomitant strain-magnetic ordering indicates an interesting possibility that a ferromagnetic transition may be a 1st order transition rather than the well-accepted 2nd order one.

See S. Yang and X. Ren, Phys. Rev. B 77, 014407 (2008) for details.

Figure 2 Lattice distortion and crystal symmetry for several typical pseudocubic ferromagnetic/ferroelectric materials. Resolutions of conventional XRD and synchrotron XRD are indicated to show why the low symmetry ferromagnetic phases were not recognized in the past.