R. F. L. Evans

Transient ferromagnetic-like state mediating ultrafast reversal of antiferromagnetically coupled spins

Ferromagnetic or antiferromagnetic spin ordering is governed by the exchange interaction, the strongest force in magnetism. Understanding spin dynamics in magnetic materials is an issue of crucial importance for progress in information processing and recording technology. Usually the dynamics are studied by observing the collective response of exchange-coupled spins, that is, spin resonances, after an external perturbation by a pulse of magnetic field, current or light. The periods of the corresponding resonances range from one nanosecond for ferromagnets down to one picosecond for antiferromagnets. However, virtually nothing is known about the behaviour of spins in a magnetic material after being excited on a timescale faster than that corresponding to the exchange interaction (10–100 fs), that is, in a non-adiabatic way. Here we use the element-specific technique X-ray magnetic circular dichroism to study spin reversal in GdFeCo that is optically excited on a timescale pertinent to the characteristic time of the exchange interaction between Gd and Fe spins. We unexpectedly find that the ultrafast spin reversal in this material, where spins are coupled antiferromagnetically, occurs by way of a transient ferromagnetic-like state. Following the optical excitation, the net magnetizations of the Gd and Fe sublattices rapidly collapse, switch their direction and rebuild their net magnetic moments at substantially different timescales; the net magnetic moment of the Gd sublattice is found to reverse within 1.5 picoseconds, which is substantially slower than the Fe reversal time of 300 femtoseconds. Consequently, a transient state characterized by a temporary parallel alignment of the net Gd and Fe moments emerges, despite their ground-state antiferromagnetic coupling. These surprising observations, supported by atomistic simulations, provide a concept for the possibility of manipulating magnetic order on the timescale of the exchange interaction.

Ultrafast heating as a sufficient stimulus for magnetization reversal in a ferrimagnet

The question of how, and how fast, magnetization can be reversed is a topic of great practical interest for the manipulation and storage of magnetic information. It is generally accepted that magnetization reversal should be driven by a stimulus represented by time-non-invariant vectors such as a magnetic field, spin-polarized electric current, or cross-product of two oscillating electric fields. However, until now it has been generally assumed that heating alone, not represented as a vector at all, cannot result in a deterministic reversal of magnetization, although it may assist this process. Here we show numerically and demonstrate experimentally a novel mechanism of deterministic magnetization reversal in a ferrimagnet driven by an ultrafast heating of the medium resulting from the absorption of a sub-picosecond laser pulse without the presence of a magnetic field.

Atomistic spin model simulations of magnetic nanomaterials

Atomistic modelling of magnetic materials provides unprecedented detail about the underlying physical processes that govern their macroscopic properties, and allows the simulation of complex effects such as surface anisotropy, ultrafast laser-induced spin dynamics, exchange bias, and microstructural effects. Here we present the key methods used in atomistic spin models which are then applied to a range of magnetic problems. We detail the parallelization strategies used which enable the routine simulation of extended systems with full atomistic resolution.

The Curie temperature distribution of FePt granular magnetic recording media

We present atomistic calculations of the magnetic phase transition behavior in an L10 FePt system to study the effect of grain size distribution on the Curie temperature (Tc) dispersion with relevance to heat assisted magnetic recording. Identifying the relation between the size and Tc of a grain by means of finite size scaling analysis of the differentiated magnetization versus T data allows to show that a lognormal size distribution transforms into a lognormal Tc distribution with moments dependent on the critical exponents. We also address the question of the universality class of FePt.

Thermally induced error: Density limit for magnetic data storage

Magnetic data storage is pervasive in the preservation of digital information, and the rapid pace of computer development requires ever more capacity. Increasing the storage density for magnetic hard disk drives requires a reduced bit size, previously thought to be limited by the thermal stability of the constituent magnetic grains. The limiting storage density in magnetic recording is investigated treating the writing of bits as a thermodynamic process. A “thermal writability” factor is introduced and it is shown that storage densities will be limited to 15 to 20 TBit/in2 unless technology can move beyond the currently available write field magnitudes.

The thermodynamic limits of magnetic recording

Thermal stability of the recorded information is generally thought to set the limit of the maximum possible density in magnetic recording. It is shown that basic thermodynamics always cause the probability of success of the write process to be less than 100%. This leads to a thermally induced error rate, which eventually limits the maximum possible density beyond that given by the traditional thermal stability limit. While the thermally induced error rate is negligible for recording of simple single domain particles, it rapidly increases in the presence of a write assist, in particular if the write assist is accomplished by an increased recording temperature. For the ultimate recording system that combines thermally assisted writing with a recording scheme that uses one grain per bit, the upper bound for the maximum achievable density is 20 Tbit/inch2 for a bit error rate target of 10−2.

Constrained Monte Carlo method and calculation of the temperature dependence of magnetic anisotropy

We introduce a constrained Monte Carlo method which allows us to traverse the phase space of a classical spin system while fixing the magnetization direction. Subsequently we show the method’s capability to model the temperature dependence of magnetic anisotropy, and for bulk uniaxial and cubic anisotropies we recover the low-temperature Callen-Callen power laws in M. We also calculate the temperature scaling of the two-ion anisotropy in L10 FePt, and recover the experimentally observed M 2.1 scaling. The method is newly applied to evaluate the temperature-dependent effective anisotropy in the presence of the Neel surface anisotropy in thin films with different easy-axis configurations. In systems having different surface and bulk easy axes, we show the capability to model the temperature-induced reorientation transition. The intrinsic surface anisotropy is found to follow a linear temperature behavior in a large range of temperatures.

Ultrafast thermally induced magnetic switching in synthetic ferrimagnets

Synthetic ferrimagnets are composite magnetic structures formed from two or more anti-ferromagnetically coupled magnetic sublattices with different magnetic moments. Here, we report on atomistic spin simulations of the laser-induced magnetization dynamics on such synthetic ferrimagnets and demonstrate that the application of ultrashort laser pulses leads to sub-picosecond magnetization dynamics and all-optical switching in a similar manner as in ferrimagnetic alloys. Moreover, we present the essential material properties for successful laser-induced switching, demonstrating the feasibility of using a synthetic ferrimagnet as a high density magnetic storage element without the need of a write field.

Energy losses in interacting fine-particle magnetic composites

The coercivity and energy losses in superparamagnetic (SPM) magnetite and FePt nanoparticle composites subjected to an external, alternating magnetic field have been calculated as a function of the mean particle size and packing density. The effect of interactions has been investigated by fitting the Sharrock law to the coercivity results as a function of the field cycle frequency of the magnetic field. This fitting leads to effective parameters for the anisotropy field and βeff = KV/kBT, which are themselves dependent on the interaction strength. The increase or decrease in the coercivity with interactions depends upon the relative change of and βeff, thus demonstrating the complex effect that interactions have in these nanoparticle composites. The interparticle interactions have a non-trivial effect on the energy loss per cycle. The energy loss is reduced for systems with larger particles since the reduction in coercivity together with a corresponding reduction in the remanence dominates. For small particle sizes, the energy loss is increased. The primary mechanism here seems to be an enhancement of the energy barrier due to interactions, which changes the nature of the particles from SPM to being thermally stable.

Atomistic spin model simulation of magnetic reversal modes near the Curie point

In order for the current increase in magnetic storage density to continue, one must overcome the so-called magnetic recording trilemma; namely, that smaller grains are required for higher data densities and to ensure their thermal stability, materials with a high anisotropy are required. The higher coercive field that this produces also becomes a limiting factor as the maximum field produced by the recording head is constrained by the saturation magnetization of the pole. One proposed solution to the trilemma is the use of heat assisted magnetic recording HAMR, which utilizes the temperature dependence of the anisotropy to enable writing of materials with a high coercivity. For the highest anisotropy media, this will require heating to the Curie temperature TC of the material. Close to TC, longitudinal fluctuations in the magnetization can have a significant impact on the expected energy barriers and therefore the relaxation time of the magnetization. These effects become especially important when attempting to minimize the time to reverse the magnetization state of the media that will be important at higher storage densities.