skip to primary navigationskip to content
 

Superconducting spintronics

Mark Blamire & Jason Robinson

Superconductivity originates in the coupling of electrons to form Cooper pairs. At least in conventional superconductors, such as the elemental materials (Al, Pb, Nb etc.) and in the majority of high temperature oxide superconductors such as YBCO, these pairs are formed with electrons which have opposite spin. This means that in a magnetic field, the Zeeman effect changes the relative energy and momentum of the two electron spin directions and so gives rise to the suppression of superconductivity by magnetic field.

It also means that the strong exchange fields within a ferromagnet strongly suppress the proximity effect between a superconductor and a ferromagnet: unlike a superconductor / normal metal system in which Cooper pairs can diffuse many nm into the normal metal, ferromagnetic material suppress Cooper pair density extremely rapidly. In fact, until 2001 no superconducting coupling to a ferromagnet had been detected. The first experiment to show this relied on the use of the weak ferromagnet CuNi through which pairs could pass from one superconducting electrode to another.

Our work on SF systems covers a number of different aspects.

S/F/S Josephson junctions

The conventional Cooper pair with zero net spin is a singlet state. As these singlet pairs pass through a ferromagnetic barrier the effective phase of the pair changes linearly so that, for certain thicknesses of the barrier the phase is opposite that of the starting pair; this is illustrated in figures 1(a) and (b). This Pi-state is interesting for certain applications, but it’s effect is most obvious in a plot of the critical (i.e. zero-voltage) current (Ic) vs barrier thickness. For a non-magnetic barriers this is an exponential decay determined by the normal metal coherence length; in the SFS case the critical current also decays exponentially, but it also reverses sign between the zero and pi phase-shift states. Experiments only usually measure the magnitude of Ic and not the sign and so a plot of Ic vs barrier thickness in this case shows a regular series of minima as Ic crosses zero.

In our programme we have now made SFS junctions with a variety of ferromagnetic barriers in which we have been able to accurately control the barrier thickness and hence the phase of the junction. [1,2,3] One of the issues in understanding the properties of SFS junctions is the formation of “dead layers” at the interface in which the magnetism is apparently suppressed. [3] We have shown that in the case of cobalt the dead layer can be eliminated completely by adding a 1nm thick layer of Rh at both interfaces. [4]

Multiple barrier junctions

It has been widely predicted that the dephasing effects of ferromagnets on singlet Cooper pairs can be eliminated by creating a barrier in which the pairs pass through magnetic layers in which the moments are antiparallel. Effectively the phase difference created by the first layer can be undone by passing through a second layer in which the exchange field has the opposite sign and hence the effects on the two spin signs are reversed. Our initial attempt to create such a device relied on using two different magnetic materials, permalloy and cobalt so that a small magnetic field switched the former by not the latter. [5] This device showed a much higher Ic in the antiparallel state of the magnetic layers compared with the parallel, but we could not make an quantitative comparisons because the coupling between the pairs and the individual layers was different. More recently we have created junctions in which tow identical Co layers were held in either parallel or antiparallel configurations be exchange coupling through a chromium spacer layer. A series of devices with progressively thicker Cr layers switched between parallel and antiparallel configurations showed a corresponding switch between low and high critical current, which could be accurately modelled [6].

Antiferromagnetic barriers and triplet pairs

We have created junctions with several antiferromagnetic materials including FeMn, [7] Cr, [8] and the weakly ferromagnetic Ho. Although these junctions showed interesting features, one of our primary interests in antiferromagnetic materials was the potential for creating spin triplet pairs in which the spins were not antiparallel. The triplet pair is believed to avoid the dephasing effects of the exchange field and so should have a much longer coherence length in a ferromagnet. We have created structures in which thin Ho layers are added at the surface of Co barriers. In comparison with the plain Co barriers the Ics for certain Ho thicknesses are very much larger, and for large Co thickness show a supercurrent when the plain Co layer does not (see Fig. 2). This is the first definitive evidence of how a triplet pairs can be controlled. [9]

Magnetic pinning of flux vortices

The increase of the critical current density of superconductor is dependent on the optimisation of the pinning of flux vortices. Normally the pinning centres are microstructural defects or inert second phase inclusions in the superconductor, but we have explored the potential for increasing the pinning strength by making the second phase particles magnetic. Our results have shown some interesting basic effects, [10] but to date experimental results have not matched our theoretical predictions. [11]  
 

[1]       C. Bell, R. Loloee, G. Burnell, & M. G. Blamire, Phys. Rev. B 71, 180501(R) (2005).
[2]       J. W. A. Robinson, S. Piano, G. Burnell, C. Bell, & M. G. Blamire, Phys. Rev. Lett. 97, 177003 (2006).
[3]       J. W. A. Robinson, S. Piano, G. Burnell, C. Bell, & M. G. Blamire, Phys. Rev. B 76, 094522 (2007).
[4]       J. W. A. Robinson, Z.H Barber, & M. G. Blamire, Appl. Phys. Lett. 95, 192509 (2009).
[5]       C. Bell, G. Burnell, C. Leung, E. J. Tarte, D. Kang, & M. G. Blamire, Appl. Phys. Lett. 84, 1153-1155 (2004).
[6]       J. W. A. Robinson, G. B. Halász, A. I. Buzdin, & M. G. Blamire, Phys. Rev. Lett. 104, 207001 (2010).
[7]       C. Bell, E. J. Tarte, G. Burnell, C. W. Leung, D. J. Kang, & M. G. Blamire, Phys. Rev. B 69, 109903 (2003).
[8]       J. W. A. Robinson, G. B. Halasz, & M. G. Blamire, Phys. Rev. Lett. 103, 207002 (2009).
[9]       J. W. A. Robinson, J. D. S. Witt, & M. G. Blamire, Science 329, 59-61 (2010).
[10]     A. Palau, H. Parvaneh, N. Stelmashenko, H. Wang, J. Driscoll, M.G. Blamire, Phys. Rev. Lett. 98, 117003 (2007).
[11]     M. G. Blamire, R. B. Dinner, S. C. Wimbush, & J. L. MacManus-Driscoll, Supercond. Sci. Tech. 22, 025017 (2009).