Björn Alling
Linköping University, Sweden
The interplay of spin and lattice in the high-temperature paramagnetic state: Consequences for phase stability and properties of steel and other magnetic materials
Most magnetic materials are synthesized in their high-temperature paramagnetic (PM) state. Many of them are also used in their paramagnetic state, like austenite steels and many ceramic compounds with magnetic components. Thus, for an ab-initio description of magnetic materials allowing for design, the description of paramagnetism Is crucial.
However, the description of the PM state is complex due to the simultaneous presence of magnetic, vibrational, structural and electronic disorder, in this high-temperature state. As well as due to coupling mechanisms between these degrees of freedom. Thus, quantitative ab-initio modeling of magnetic materials, in particular regarding phase stability at high temperature, has been lagging behind the corresponding work on their non-magnetic counterparts.
In this work I demonstrate the method development we have done on Atomistic Spin Dynamics coupled to ab-initio Molecular Dynamics (ASD-AIMD) [1] for treating magnetic materials at high temperature, including temperature dependent magnetic short-range-order, effects of magnetic disorder on interatomic forces, longitudinal spin-fluctuations [2], as well as direct dynamical spin-lattice coupling.
Applying ASD-AIMD simulations to CrN, I show how this method can be used together with phonon lifetime analysis to explain the non-intuitive increase in thermal conductivity with temperature in PM CrN. Moving further, together with thermodynamic integration schemes, both over temperature and stress-strain integration, we demonstrate how one can obtain the free energy differences between competing phases of magnetic materials.
In pure Fe, our ab-initio method derives free energy differences between bcc and fcc phases as a function of temperature that are within a few meV/atom of experiment-based CALPHAD derivations, and reproduce the alpha-gamma-delta transition series well. In particular, the gamma-delta transition is predicted within 50 K of the experimental value of 1667 K. [3]
In order to explore the generality of these accurate results, the gamma-delta transition in Fe1-xCrx and Fe1-xMnx alloys are derived using the same methodology. We demonstrate that the experimentally observed decrease in gamma-delta transition temperature for Fe1-xCrx and increase in Fe1-xMnx is quantitatively well reproduced.
These results shows promise for ASD-AIMD based simulations to be the long-sought framework allowing for ab-initio design of steels.