X-ray Scattering Techniques
Skip to section: |
Introduction |
Instruments |
Research |
The X-ray sub-group is headed by Dr Elizabeth Blackburn, and includes Prof. Ted Forgan, along with Erik Jellyman and Emma Campillo Munoz (research students). Our research in this field has resulted in numerous high-profile publications.
Electronic effects studied using x-rays
X-rays interact with the electrons in matter, and are routinely used to establish the physical structure of materials, i.e. where the electrons in a given compound are spatially. The x-ray photons can also interact directly with the spin of the electrons, and so they can scatter off the magnetic structure. However, the cross-section for this interaction is very small, and usually impractical to measure unless one has a large amount of time or a large amount of photons. Modern synchrotron sources certainly provide the latter.
There is another trick for looking at both magnetism and more subtle electronic effects, which involves exploiting resonant effects. If the incident photon energy is just right, it can excite a core atomic electron into the valence band. This is useful for identifying particular elements, as each element has its own characteristic energies. If the material is magnetic, then the valence bands may be split (as in the Zeeman effect) and one becomes sensitive to the magnetization in the material.
The key experimental necessity to exploit this feature is the ability to select the desired incident photon energy, according to the content of a given sample. Fortunately, this is possible with the insertion devices available at synchrotron sources.
Instruments
There are many synchrotron sources all over the world, with a variety of instruments optimized to answer different questions. Typical instruments for resonant absorption and scattering studies include I16 at Diamond, Oxford (UK), BM28, and ID12 at the European Synchrotron Radiation Facility, Grenoble (France).
Research Examples
Which atoms are magnetic in magnetocaloric materials? ALS feature
The magnetocaloric effect is defined as the reversible change in temperature of a magnetic material due to a change in the applied magnetic field. This change in temperature occurs because when the magnetic field is changed adiabatically, the spins in the magnetic material align, decreasing the magnetic entropy. As the process is adiabatic, the total entropy cannot change, and so the decrease in the magnetic entropy must be matched by an increase in lattice or electronic entropy, which is usually manifest as heat. This will happen in all magnetic materials, but leads to a significant temperature change in so-called magnetocaloric materials. Such materials have obvious uses in refrigeration systems.
One such material is obtained by substituting various elements for Mn in the compound Ni2MnGa (itself a very interesting shape memory alloy) - see S. Stadler et al., Applied Physics Letters 88, 192511 (2006).
One way to work out which atoms carry a magnetization is to probe the material at the resonant absorption edges of the composite atoms. Using circularly polarized light, a signal proportional to the magnetization can be obtained. If a complete measurement is made, the spin and orbital components of this magnetization can be extracted. By looking at the profile as a function of incident photon energy, it is possible to gain information on the local environment, and discern which electronic states are implicated. Link for more information. This technique is known as x-ray magnetic circular dichroism (XMCD).
Such measurements have been carried out on the parent compound, Ni2MnGa, (G. Jakob et al., Physical Review B 76, 174407 (2007)) which show that although most of the magnetization is carried by the Mn, it is the Ni that shows most sensitivity to the local environment, and it is the Ni that generates anisotropy in the material. We have recently completed a study (S. Roy et al., Physical Review B 79, 235127 (2009)) of the effect of substituting Cu for Mn. This causes the magnetism to become more delocalized - a key property of the parent compound Ni2MnGa – driving the magnetic transition temperature down. This delocalization reinforces the Ni-Ga chemical bond, raising the temperature of the structural transition. At 25% doping, the two transitions coincide (maximizing the magnetocaloric effect) at 44°C – a very suitable temperature for cooling down overworked bits of semiconductor! Work is ongoing in developing suitable thin film variants of this compound.
What happens at the interface of a ferromagnet and an antiferromagnet?
By sandwiching together various combinations of ferromagnets and antiferromagnets, a wide range of interesting, and useful, phenomena can be observed. One example is giant magnetoresistance, another is exchange bias. One thing that is important is the interface between the various layers. Due to the constraints of sample growth, wherein one layer of thin film is deposited on top of another - these interfaces are usually buried, and difficult to probe directly.
One way to get at this problem is to use penetrating radiation probes, such as x-rays or neutrons. Studies of thin film sandwiches have been carried out using both the XMCD technique described above, and also reflectivity measurements at energies specific to particular elements. A reflectivity measurement is pretty much as it sounds; the beam of x-rays is reflected off the surface sample. Below certain angles, it is completely reflected (total external reflection), but beyond the critical angle, the amount of reflected light depends on the depth profile of the electron density (and magnetization density, if at a resonant edge). By measuring over a wide range of angles, the interference fringes due to various layers can be measured, and the electron density profile extracted.
In one bilayer system, ferromagnetic Permalloy on antiferromagnetic CoO, which displays exchange bias when cooled through the Neel temperature of the CoO, resonant x-ray reflectivity has shown that, in the unbiased state, at the interface some of the Co spins are not completely compensated, giving rise to a thin layer of Co magnetization that follows the external field - S. Roy et al., Physical Review B 75, 014442 (2007). The observation of exchange bias appears to depend very strongly on how the uncompensated spins at the interface react on entering the biased state, with most current models relying on pinned spins.
In the biased state, we have observed (E. Blackburn et al., Physical Review B 78, 180408 (2008)) that thin layer of uncompensated Co spins shows two different types of behaviour. The majority of the Co spins follow the external field. However, 10% of them are pinned antiparallel to the cooling field used to bias the sample. These pinned spins are antiferromagnetically coupled to the ferromagnetic layer and are responsible for the formation of exchange bias.
The X-rays programme is led by Dr. Elizabeth Blackburn