Ferromagnetism and Superconductivity
Hybrid Ferromagnet / Superconductor Devices

 
It has long been accepted that superconductivity and ferromagnetism are two mutually exclusive phenomena that generally do not coexist. Recently, however, sufficient evidence has been obtained to indicate that in some unconventional superconducting systems superconductivity and ferromagnetism, as well as superconductivity and anti-ferromagnetism can coexist. We are interested in some of the aspects of this interaction.
We are also interested in device applications of hybrid superconductor/ferromagnetic structures. For instance, we have recently proposed a magneto-superconducting switch. The device, which is based on Andreev reflection at the superconductor/ferromagnet contact, combines high efficiency with non-volatility.
The low-impedance state of the device corresponds to the normal state of the superconductor, whereas the high-impedance state corresponds to the superconducting state. The proposed device does not require high-quality Andreev contacts; on the contrary, interface scattering significantly increases the efficiency of the device. Up to 1000% - 2500% efficiency can be achieved with the existing ferromagnetic materials. The device can be used as a basic element for non-volatile logic and memory.
The proposed device, which we call a Non-Volatile Andreev Switch (NOVAS) consists of an Andreev nanocontact between a highly spin-polarized magnetic metal and a superconductor, which can be switched from superconducting to normal state by the edge field of a ferromagnetic control film positioned on top of the contact. The edge field of the control film can be modulated from ~ 0 to several kOe by rotating the magnetic moment M in the x-y plane of the film by 90º (see Fig.1). While these edge fields are large enough to quench the superconductivity in Sn or Pb, for example, the magnetization rotation can be done by using a much smaller in-plane field if the ferromagnetic control film is fabricated of a soft magnetic material (e.g. permalloy). The suppression of Andreev reflection in a ferromagnet/superconductor (FS) contact leads to high impedance when the superconductor is in the superconducting state and to low impedancewhen the superconductor is in the normal state. Nominally, for a half-metal the efficiency Rhigh/Rlow of such a device is infinite at T = 0 (the conductance at zero bias is zero). In reality, this ratio will be limited by spin-flips, spin-orbit interaction, and other secondary processes. While the efficiency of conventional magnetic devices decline rapidly when P is reduced from 100%, a NOVAS will retain its high efficiency even for materials with P ~ 70%-80%, which are readily available. As can be seen from Fig.2, the device efficiency can be further increased by using low quality (high scattering) contacts.
 
Schematic of Non-Volatile Andreev Switch (NOVAS)
Fig. 1 - Schematic of Non-Volatile Andreev Switch (NOVAS). The device consists of an Andreev nanocontact between a highly spin-polarized ferromagnetic film (bottom electrode) and a superconducting film (top electrode). A soft ferromagnetic control film (F), e.g. permalloy, is fabricated on top of the superconductor. The resistance of the device is modulated by the F-film, whose fringe field (approximately several kOe) suppresses superconductivity when its magnetic moment is in-plane (a). The normal area (shown by a black rectangle) is spanned over a distance comparable with the thickness of the F-film, typically about 100 nm, which is generally much lager than the contact size, a ~ 3 -30 nm.
 
Impedance ratios of superconducting-to-normal states of NOVAS as a function of spin polarization P for Z=0 (two bottom curves) and Z=1 (two top curves) for the ballistic and diffusive cases. For a reduced barrier transparency (Z=1) the ratio in both cases significantly increases.
Fig. 2 - Impedance ratios of superconducting-to-normal states of NOVAS as a function of spin polarization P for Z=0 (two bottom curves) and Z=1 (two top curves) for the ballistic and diffusive cases. For a reduced barrier transparency (Z=1) the ratio in both cases significantly increases.

 
 
Copyright ©2006 Boris Edward Nadgorny