Voltage control of magnetocrystalline anisotropy in ferromagnetic-semiconductor-piezoelectric hybrid structures

Abstract

We demonstrate dynamic voltage control of the magnetic anisotropy of a (Ga,Mn)As device bonded to a piezoelectric transducer. The application of a uniaxial strain leads to a large reorientation of the magnetic easy axis which is detected by measuring longitudinal and transverse anisotropic magnetoresistance coefficients. Calculations based on the mean-field kinetic-exchange model of (Ga,Mn)As provide microscopic understanding of the measured effect. Electrically induced magnetization switching and detection of unconventional crystalline components of the anisotropic magnetoresistance are presented, illustrating the generic utility of the piezo voltage control to provide new device functionalities and in the research of micromagnetic and magnetotransport phenomena in diluted magnetic semiconductors.

Interface Magnetism in Ferromagnetic Metal–compound Semiconductor Hybrid Structures

“…Their ferromagnetic resonance measurements indicated that whilst the free-energy density for magnetizing the film along the in-plane ⟨100⟩ directions were equivalent, due to the anticipated cubic magnetocrystalline anisotropy in their Fe films, the in-plane ⟨110⟩ directions were inequivalent -indicative of an additional uniaxial magnetic anisotropy term K U with easy-axis oriented along the [110] and hard-axis along the [110] direction. The origin of this uniaxial magnetic anisotropy which is found in these, and related, films has escaped explanation: it is toward gaining an understanding of this anisotropy we will focus the remainder of this review.…”

“…It is obvious that care must be taken in applying the correct expression: further difficulties may arise when the interfacial uniaxial easy axis is no-longer oriented along the [110] direction, but rather along [110], as is the case in films deposited on, for example, InAs(001). When considering such an admixture of cubic magnetocrystalline and interfacial uniaxial magnetic anisotropies, three distinct situations may occur: i) when the ferromagnetic film is very thick, in which case |K eff U | ≪ |K eff 1 |, the uniaxial anisotropy is negligible and the film behaves as if it has only cubic in-plane magnetic anisotropy; ii) in the case when the ferromagnetic film is very thin, in which case |K eff U | ≫ |K eff 1 |, the cubic anisotropy is negligible and the film behaves as if it has only uniaxial in-plane magnetic anisotropy; and, iii) when the film is of a thickness where |K eff U | ≈ |K eff 1 |, both terms contribute, which results in two-stage magnetization reversal for fields applied along the uniaxial hard-axis (UHA), depicted schematically in the inset to the lower right frame of figure 2.…”

“…Entries for the III-Vs are ordered by increasing lattice parameter, and we see that the lattice mismatch for the epitaxial Fe overlayer passes through a minimum, being roughly zero for the In 20 Ga 80 As ternary compound. One very interesting point to note is the dependence of the uniaxial easy-axis direction on the lattice mismatch: if one were to assume that the interface anisotropy arises due to epitaxial strain or anisotropic stress relaxation, as have previously been suggested in the literature, one may think that a positive lattice mismatch (2a Fe > a) results in interface anisotropy along [110]. From this, we would then anticipate the interface anisotropy vanishing when the lattice mismatch is zero -leaving only the cubic magnetocrystalline anisotropy: however, this is not the case 42 .…”

“…From this, we would then anticipate the interface anisotropy vanishing when the lattice mismatch is zero -leaving only the cubic magnetocrystalline anisotropy: however, this is not the case 42 . Indeed, for both In 50 Ga 50 As and Al 48 In 52 As ternary compounds, the lattice mismatch is negative (2a Fe < a) but the easy-axis of the uniaxial magnetic anisotropy remains along the [110] direction. This shows that it is unlikely that epitaxial strain is the origin of the uniaxial interface anisotropy, as is now the generally-held consensus 33,41,55,56 .…”

“…where now ϕ is the angle between the magnetization and the [110] direction, and α is the angle between H and the [110] direction. It is obvious that care must be taken in applying the correct expression: further difficulties may arise when the interfacial uniaxial easy axis is no-longer oriented along the [110] direction, but rather along [110], as is the case in films deposited on, for example, InAs(001).…”