Representation of the in situ spin resolved photoemission spectroscopy experiment.© Martin Jourdan.
Representation of the in situ spin resolved photoemission spectroscopy experiment.© Martin Jourdan.

Spin based electronics or spintronics rely on the magnetic moment of the electrons, their spin, for data storage and manipulation. As a quantum property the spin can only be aligned parallel (up) or antiparallel (down) with respect to a magnetic field.

A well-known application of this principle is the up-to-date read-head of hard disk drives, which is basically a tunneling junction of two ferromagnetic CoFeB layers separated by a thin insulating MgO barrier. With the magnetization direction of one of the ferromagnetic layers fixed, the second one follows the magnetic field generated by the magnetized domains of the hard disk. This way, a large tunneling magnetoresistance (TMR), that is, a difference in the read head resistance depending on the relative magnetization directions of the two ferromagnetic layers is obtained, which allows the read out of the information stored on the disk. However, due to the large area resistance of the tunneling junctions, further miniaturization, that is, increase of the storage density, is limited by the resulting huge absolute resistance of the device.

This problem could be solved by returning to the previous read-head technology, in which instead of the insulating tunneling barrier a thin non-magnetic metallic spacer separates the ferromagnetic layers of the read head. These so called spin valves were replaced by TMR junctions, because their magnetoresistance, based on the giant magnetoresistance effect (GMR), is much smaller than the TMR. However, the magnitude of the spin valve GMR depends on a central electronic property of the ferromagnetic electrodes: The spin polarization, that is, the normalized difference of the number of spin up and spin down electrons available for charge transport, which for conventional ferromagnets like CoFeB amounts to ≅60%. The availability of a material with an increased, ideally 100%, spin polarization would substantially increase the GMR and allow for a change of read head technology back to spin valves, with the possibility of further miniaturization of the read heads.

In addition to this most straight forward application of highly spin polarized materials there are many visionary spintronic applications like the spin-field effect transistor (Spin-FET, [1]) based on the Rashba-effect. This potential application requires the injection of a spin-polarized current in a semiconductor in which case the benefits from a near 100% spin polarization are most obvious: Theory predicts a direct injection efficiency which is two orders of magnitude smaller if the spin polarization is reduced from 98% to 80% [2].

It is obvious that materials with close to 100% spin polarization (so called half metals) at room temperature are highly desirable. Thus more than 10 years ago the theoretical prediction (e.g. [3]) of half metallicity in Heusler materials, intermetallic compounds with X2YZ composition characterized by four interpenetrating face centered cubic sublattices, raised considerable interest (for an overview of the properties of this class of materials see e.g. [4]). Experimentally, at very low temperatures various evidence for high spin polarizations were found, but disappointingly for unclear reasons the measured spin polarizations were always strongly reduced at elevated temperatures approaching room temperature.

Fortunately, recent spin-polarized photoemission spectroscopy experiments revived the hope, that Heusler materials represent a breakthrough in the field of spintronics: at Mainz University (JGU), Germany, investigating the Heusler compound Co2MnSi, 93% spin polarization at room temperature was measured directly [5]. In addition to the measurements, the theoretical description developed at Munich University (LMU), Germany, represents a major step forward. For the first time it allows a realistic description of the surface effects in Heusler compounds, which strongly contribute to the density of states derived photoemission probabilities. Comparing the experimental results with the calculations of the band structure and photoemission spectrum it was concluded that the Heusler compound Co2Mnsi is a true half-metal with a relatively small minority charge carrier gap in the bulk, which is widened in the surface region of the thin film samples due to a 100% spin polarized surface resonance. The directly measured 93% spin polarization is fully consistent with true half-metallicity considering the limited energy resolution of the experiments.

From the experiments it can also be concluded that the proper preparation of a well-ordered high purity surface of the Heusler compound is essential for a huge spin polarization. Thus with respect to spintronics applications the new results raise hope, but also indicate that very careful interface engineering will be required, which explains the absence of Heusler materials in today's memory devices. Additionally, it is still an open question if the large spin polarization measured on a free surface in ultrahigh vacuum can be maintained in contact with other metals, insulating tunneling barriers or semiconductors.

However, given that it was now finally possible to experimentally observe the long predicted room temperature half metallicity, such questions can be tackled with renewed enthusiasm.

Further reading

1. B. Datta, et al., Appl. Phys. Lett., 56 (1990), p. 665

2. G. Schmidt, et al., Phys. Rev. B, 62 (2000), p. R4790

3. I. Galanakis, et al., Phys. Rev. B, 66 (2002), p. 174429

4. T. Graf, et al., Prog. Solid State Chem., 39 (2011), p. 1
5. M. Jourdan, et al., Nat. Commun., 5 (2014), p. 3974

Read full text on ScienceDirect

DOI: 10.1016/j.mattod.2014.08.020