Materials World Network: Spin dynamics of the ferromagnet/antiferromagnet interface studied by time-resolved x-ray magnetic dichroism

Lead Research Organisation: University of Exeter
Department Name: Physics


The field of spintronics aims to deliver new device function by controlling the motion of an electron through its magnetic moment, or "spin", as well as through its electric charge. The outstanding success of spintronics has been the use of Giant Magnetoresistance (GMR) in the spin-valve sensors used to read data from hard disk drives. The spin-valve consists of two ferromagnetic (F) layers separated by a non-magnetic layer. Each F layer has a magnetic moment that acts as a compass needle and reorients in response to an applied magnetic field, such as that generated by the bits of data stored on a hard disk. GMR occurs if the two compass needles change their relative orientation. If one of the compass needles is kept fixed while the other is free to reorient in the magnetic field then data can be read out as a change in the electrical resistance of the sensor.
This project is concerned with the means by which one of the compass needles is fixed. The established method is to deposit an antiferromagnetic (AF) layer on the outside of one of the F layers. The F/AF interface generates a strong effective magnetic field that fixes the orientation of the F layer magnetization. This effect, known as "exchange bias", is widely used but poorly understood in detail. Within a F material each atom has a magnetic moment, and every such magnetic moment, or compass needle, is forced to align in the same direction by the powerful "exchange interaction". Within an AF material adjacent magnetic moments instead align anti-parallel to each other. At the F/AF interface the magnetic moments of the F become fixed relative to those in the interfacial AF layer. The AF material has no net magnetic moment, so is largely unaffected by applied magnetic fields, and its magnetism is more difficult to observe. Little is known about the magnetic moments (spins) of the AF at an F/AF interface, particularly when structural imperfections are present. Spin-valves are required to change their magnetic alignment on sub-nanosecond timescales, where the motion of the magnetic moments within the AF and their influence upon the F are completely unexplored.
We will use synchrotron x-ray radiation to make the first measurements of the motion of the magnetic moments at the F/AF interface at GHz frequencies. In particular we will make use of the x-ray magnetic circular and linear dichroism effects, known as XMCD and XMLD respectively. The F/AF samples consist of atoms in which a nucleus is surrounded by filled and partially filled shells of electrons. The energy required to excite an electron from a filled to a partially filled shell has an energy that is specific to a particular atom, while the energy of the x-rays from the synchrotron can be tuned so as to study only that atom. The x-rays are produced in pulses of sub-nanosecond duration. By synchronizing the x-rays with a magnetic field that has a sinusoidal time variation, the instantaneous state of the sample may be determined at a given point in its cycle of oscillation. Specifically the XMCD and XMLD effects allow the magnetic state of the F and AF layers to be determined independently.
We will use the Advanced Light Source (ALS) in Berkeley and the Diamond Light Source in the UK to apply this measurement technique to samples of the highest structural quality, fabricated by molecular beam epitaxy at the University of California, Berkeley. The GHz frequency dynamics of the F layer will first be characterized by time resolved magneto-optical measurements at Exeter. Both x-ray and magneto-optical measurements will be performed as a function of temperature so as to compare the response when the AF layer has different degrees of antiferromagnetic order. We will hence obtain much deeper insight into how the AF layer controls the response of the F layer to a high frequency magnetic field.

Planned Impact

The proposed project aims to provide improved an understanding of the ferromagnet/antiferromagnet (F/AF) interface and particularly the response of its constituent magnetic moments (spins) to magnetic fields in the 1-10 GHz frequency range. The F/AF interface generates the "exchange bias" that pins the fixed layer of a spin or tunnel valve structure. This effect has huge commercial importance since the spin-valve has been used as the recording sensor in every hard disk drive manufactured since around the year 2000. Spin valves are also used in Magnetic Random Access Memory (MRAM) manufactured by companies such as Everspin, as tunable microwave sources fabricated by SMEs such as Nanosc, and in a variety of magnetic field sensing applications ranging from navigation to automotive engine management. While exchange bias has been widely exploited, our understanding of the magnetic state of the AF is limited. The exchange bias field is usually a small fraction of its maximum theoretical value, suggesting that interfacial magnetic moments are frustrated, that AF domains are present but uncontrolled, and that the AF state is at best metastable. The stability of the exchange bias and its robustness under thermal cycling and in extreme environments are important questions that are currently treated empirically. Within data storage and microwave device applications the magnetic layers of the spin-valve may be excited at GHz frequencies. It is essential to understand whether the static and dynamic exchange bias fields are really the same and why exchange biased F layers typically exhibit enhanced damping. The enhanced damping may be an advantage in a sensor plagued by parasitic spin transfer torques effects, or else a disadvantage in an oscillator application where a narrow linewidth is required. Modest improvements to the exchange bias mechanism can have a commercial value of millions of dollars, such is the scale of the magnetic recording industry. Within the UK, Seagate's Springtown operation in Northern Ireland manufactures up to 400M recording heads per annum, some 25% of the world supply, and contributes some £100M p.a. to the UK economy.
Our approach is to obtain understanding through the study of model systems where the magnetic state of the F/AF system is unobscured by structural imperfections. We do not pretend that these materials will find immediate commercial application, rather that they will yield understanding that can be applied to materials of practical relevance. Our results will be made available to all through open access publications in mainstream condensed matter physics and electrical engineering journals that are read by both academics and industrial research engineers. Knowledge will also be transferred to industrial beneficiaries by seminars at industrial facilities and through personal interaction since Exeter has a very strong track record of collaboration with industry. Hicken has been directly funded by Seagate Technology for a total of more than 7 years through 3 separate grants under the Seagate plan. In 2003 Hicken also spent 2 months at the Hitachi Global Storage Technologies San Jose research centre that led to an enduring collaboration that has produced 10 joint publications. Furthermore Hitachi fabricated the samples for the XFMR studies conducted within our previous EPSRC grant.
Finally, the skills and expertise in growth, optical, x-ray and high frequency metrology, and modeling gained by the young researchers within the project will make them highly attractive for employment within the magnetic recording or other advanced materials industries.