Concept of the Project and Challenges
Information storage technology is essentially based on nanostructured magnetic materials. Considerable research effort is aimed at increasing the density of stored information and this generally requires increasingly sophisticated media design to engineer the desired combination of low write field and thermal stability of recording information [1,2]. An alternative approach is Heat Assisted Magnetic Recording  in which a laser is used to heat the medium to a sufficiently high temperature to assure writability using currently available write head fields. Also a new, highly promising, development is that of spin electronics in which the spin of the electron rather than merely the charge forms the basis of the device operation. This holds the prospect of allowing technology to develop beyond the limits of miniaturisation of standard electronics and may yield the solution of the increasing power requirements for conventional electronic devices . Common to all these technologies is the need for fast manipulation of the carriers of the information of the devices. Conventional magnetic elements, such as magnetic write heads, are driven by current-generated magnetic fields. This, combined with the relatively slow speed of precessional reversal processes (around 100ps) imposes a serious limitation on their speed.
However, optical spin manipulation and magnetisation processes have been shown experimentally to be intrinsically much faster than that of precessional switching. Beaurepaire et al  first demonstrated sub-picosecond magnetisation changes in Ni using ultrafast laser pulses. This was followed by further investigations [6,7] confirming complete demagnetisation on the timescale of hundreds of fs. Evidently the speed of magnetisation processes is not limited by precession, which presents exciting opportunities for future device applications. In a further important development, members of the consortium have demonstrated experimentally for the first time all-optical reversal with a 40fs laser pulse using circularly polarised light [8,9] resulting also in a patent . In a further highly successful collaboration between theory and experiment within the consortium we have shown  that this laser induced reversal proceeds via a novel ‘linear’ reversal mechanism where the reversal path is through a strongly non-equilibrium state of zero magnetisation. Consequently, the speed of reversal is governed by the longitudinal relaxation time, which is around 2 orders of magnitude faster than the precessional switching used in conventional devices, allowing the engineering of new devices with unprecedentedly fast switching and manipulation of spin states. An important recent discovery by members of the consortium  is the existence of a transient ferromagnetic state during the reversal of GdFeCo which drives the reversal process and leads to switching times at the sub-picosecond level. This indicates the thermodynamic complexity of ultrafast magnetism and the necessity for models beyond the state-of-the-art.
The standard formalism used for materials and device simulations (micromagnetics) is a continuum formalism dealing with the spin precession timescale of 10ps upward. It is important to note that progress beyond precessional switching speeds requires models based in the timescales associated with exchange, spin-orbit coupling and spin-flip scattering.
Femtosecond spin manipulation relies on the efficiency of intrinsic scattering mechanisms giving rise to spin-flip processes. Understanding the timescales of photons, electrons, phonons and spin interactions in different materials is thus an important part of the material design. This can be done only on the basis of time-dependent density function theory. On the other hand, realistic comparison with experiments can be done only on the basis of mesoscopic continuum models. Recently, members of the consortium have been working on the development of multi-scale modelling schemes [13,14] capable to pass the information from DFT to mesoscopic model via atomistic spin models to include thermal fluctuations. One of the important developments of our multiscale approach is the use of the Landau-Lishitz-Bloch (LLB) equation of motion . In comparison to the standard Landau-Lifshitz-Gilbert equation the LLB equation does not conserve the magnitude of the magnetisation, which is an essential requirement for the description of ultra-fast demagnetisation at the mesoscopic level.
This figure illustrates that the problem of comparison with experiment requires a multiscale process in timescale as well as in lengthscale. Specifically, it shows the effect of circularly polarized 100fs optical pulses on the magnetic structure of a GdFeCo domain. Note that after 1ps the domain image for both polarizations are the same and indicate a nearly demagnetized state (albeit, importantly with a small reversed magnetisation); however the final state is clearly determined by the pump polarization, though this excitation was only of 100fs duration. This shows that the models must span the fs to 100ps timescale in order to allow quantitative comparison with experiments. Equally, it is important to develop experimental techniques with fs time resolution that allows spin, orbital and chemical specificity, while in addition the experiments need higher spatial resolution in order to investigate the microscopic details (nucleation) between for example the two images at 1ps. Consequently the partnership between the theoretical modelling and the advanced experimental measurements offers exciting possibilities for improved understanding of the physical basis of ultrafast magnetic spin manipulation with the ultimate goal to design new and better materials for ultrafast devices.
Femtosecond spin manipulation by optical pulses is clearly an emerging scientific area that may lead to the development of new materials for the technology of tomorrow. However, before we can incorporate these new scientific ideas into actual technology there are many open challenges which need to be addressed, in particular:
- At the light-spin interaction level
- What are the fundamental interactions between spin, conduction electrons and the laser light?
- What is the origin of the laser-induced “optomagnetic” field?
- What is the relevance of the specific spin ordering (ferro, ferri, antiferro - magnetic) in the materials of interest.
- At the electronic structure level
- How to model the non-equilibrium spin-scattering channels that cause an ultrafast demagnetization.
- How to treat the interaction of matter with intense, fs laser fields theoretically.
- At the atomistic length scale
- How can these mechanisms be introduced into the atomistic lengthscale?
- How to include energy transfer on the fs timescale.
- At the mesoscopic level
- To develop macrospin models of ferrimagnetic and antiferromagnetic materials.
- To develop quantum-based macrospin models.
These challenges can only be met by developing a multiscale approach that spans all relevant time and length scale from the time dependent density functional theory to include the fundamental spin and electronic interactions, via atomistic models to include thermal fluctuations to the mesoscopic level to compare with realistic nanostructured materials.
References: - D. Weller et al., IEEE Trans. Magn. 35, 4423 (1999) (read more).  - D. Suess et al., Appl. Phys. Lett. 87, 012504 (2005) (read more).  - F. Hendrik et al. Appl. Phys. Lett. 84, 810 (2004) (read more).  - C. Chappert et al. Nat. Mater. 6, 813-823 (2007) (read more).  - E. Beaurepaire et al. Phys. Rev. Lett. 76, 4250 (1996) (read more).  - B. Koopmans et al. Phys. Rev. Lett. 85, 844 (2000) (read more).  - J. Hohlfeld et al. Phys. Rev. B. 65, 012413 (2001) (read more).  - A. Kimel et al. Nature 435, 655-657 (2005) (read more).  - C. Stanciu et al. Phys. Rev. Lett. 99, 047601 (read more).  - C. Stanciu et al. All-optical magnetic switching, Patent PCT/NL2006/000264.  - K. Vahaplar et al. Phys. Rev. Lett. 103, 117201 (2009) (read more).  - I. Radu et al. Nature 472, 205-208 (2011) (read more).  - N. Kazantseva et al. Europhys. Lett. 81, 27004 (2008) (read more).  - L. Szunyogh et al. Phys. Rev. B 79, 020403(R) (2009) (read more).  - U. Atxitia et al. Appl. Phys. Lett. 91, 232507 (2007) (read more).