Magnetically doped ZnO for spintronics applications
J.M.D. Coey's lemma: "Conventional electronics has ignored the spin of the electron"
Spintronics or spin electronics is a new branch in fundamental research for the development of new, ultra fast computers. In common computational devices the data is stored on magnetic hard discs by means of a recording head switching the local magnetization, while the data processing is based on the electric charge, e.g. in p-n-junctions, transistors etc. In order to overcome this mismatch, the spin of the electron must be processed additionally to its charge. One way to utilize both properties is the use of diluted magnetic semiconductors (DMS) where the charge carriers, i.e. holes and electrons are magnetically polarised due to codoping with transition or rare earth metals. Choosing the right doping material and doping conditions is the current goal of intense worldwide research.
Magnetic semiconductors are already well known. The materials EuX (X=O, S, Se and B6), CdCr2X4 (X=S, Se), XCr2S4 (X=Mn, Fe, Co) exhibit semiconducting and ferro- resp. ferrimagnetic properties but low Curie temperatures. On the other hand, half-metallic materials with low resistivity but high Curie temperatures do exist, e.g. Heusler alloys (NiMnSb, PtMnSb, Co2MnSi), simple oxides (CrO2, Fe3O4) or complex oxides (Sr2FeMoO6, LaXMnO3 with X=Ca, Sr, Ba). However, materials with well known semiconducting properties like Si, GaAs, GaN or ZnO are desired, that also exhibit the half metallic state. Such a material was found a few years ago, e.g. GaMnAs exhibiting a Curie temperature around 110 K. Now the race for the creation of a room temperature DMS meeting the following requirements is on:
- real DMS, e.g. no magnetic secondary phases
- 100% ferromagnetic polarization of the conducting electrons (half metallic state)
- Curie temperature well above 300 K
- semiconducting transport behavior (resistivity between metallic and insulating)
- p- and n-type doping capability
Up to now these material requirements are not fulfilled. Partial success has been obtained in developing room temperature diluted ferromagnetic ZnO and GaN, but often secondary phases are responsible for the ferromagnetic properties.
In the division "Nanofunctional Films" ZnO single crystals are implant doped with different transition or rare earth metals like Fe, Co, Gd an Tb. While other methods of doping like solid state reaction require high temperatures, implant doping can be performed at any temperature with any concentration of the dopant material. Thus processing far from thermal equilibrium prevents the system from secondary phase formation but introduces ion collision caused defects. Such defects can suppress or even enhance the ferromagnetic coupling between the dopant ions diluted within the host matrix. They can be removed via furnace or rapid thermal annealing. Figure 1 shows superconducting quantum interference device magnetometry (SQUID) for ZnO implanted at different Fe-fluences and temperatures. Structural properties have been analyzed using transmission electron microscopy (TEM) and X-ray diffraction (XRD). While for high implantation temperatures strong diffusion prevents the Fe ions from ferromagnetic coupling but produces Fe nanoparticles, low fluence implanted samples show DMS behavior. The introduced defects must play a key role for the ferromagnetic coupling.
Hysteresis loops upon magnetization reversal of Fe implanted ZnO measured using SQUID magnetometry. The ion implantation profile has a Gaussian shape with maximal atomic concentrations indicated. The implantation temperature Timp was varied between -20 °C and 350 °C. While for both samples room temperature ferromagnetism was observed, its origin is different. For the sample with the high Fe concentration and the high Timp small Fe nanoparticles are formed and identified using XRD and TEM. For the sample with the low Fe concentration implanted at -20 °C a diluted situation is established. The Fe ions are coupled indirectly via the ZnO host lattice.
Hysteresis loops upon magnetization reversal of Gd implanted ZnO measured using SQUID magnetometry. After furnace annealing at 550°C a pronounced hysteresis loop can be observed. Secondary phases can be excluded due to XRD and their expected low Curie temperatures.
In Figure 2, SQUID hysteresis loops of Gd implanted ZnO single crystals are shown. It has been found, that furnace annealing of the samples leads to a ferromagnetic coupling of the Gd ions within the ZnO host matrix. This effect must be related to the decrease of the sheet resistivity of the sample due to annealing. A structural analysis has been performed using XRD but no secondary phases have been found. Possible secondary phases like Gd, ZnGd or Gd2O3 do not exhibit room temperature ferromagnetism.