Magnetoelastic Options to mcphas

-doeps	If the program mcphas is started with option -doeps and it finds elastic constants in the input file mcphas.j (note, that the elastic constants in the input file are normalised to the primitive crystallographic unit cell, units are meV / primitive crystallographic unit cell), it will use these and determine selfconsistently the strain $\epsilon$ by solving equations (112) and (110). Elastic energy and strain tensor are stored in results/mcphas.fum. In this way it is be possible to model Jahn Teller transitions, phase diagrams, magnetostriction, thermal expansion (magnetic part) and dynamics consistently based only on point charges and Born von Karman springs. If mcphas is used with option -doeps and it finds files mcphas.djdx, mcphas.djdy, mcphas.djdz with derivatives of the two ion interaction parameters with respect to spatial coordinates $x,y,z$ respectively, it performs a calculation of the corresponding correlation functions and computes the strain using equation (116) at each iteration. Subsequently, making use of the Taylor expansion of the interaction constants (114) the computed strain is used in the next iteration for the the mean field Hamiltonian (mean field theory is used by McPhase to solve the Hamiltonian (113) and (110)). At the end of the iteration loop mean fields, moments $\langle I \rangle$ and a selfconsistent strain tensor $\epsilon$ is obtained.
-linepscf	The option -linepscf together with -doeps will trigger a calculation where equation (112) (and in case of exchange striction terms (116)) is used to calculate the strain $\epsilon$, however the last two terms in (110) are always evaluated for zero strain, i.e. the crystal field striction is not calculated selfconsistently, but only in the linear approximation assuming that the strain leads only to a negligible perturbation of the crystal field Hamiltonian.
-linepsjj	The option -linepsjj together with -doeps will trigger a calculation where equation (116) is used to calculate the strain $\epsilon$, yet the modification of the two ion interaction by the strain is assumed to be small and two ion interaction parameters (114) are used in the mean field loop always for zero strain, i.e. the exchange striction is not calculated selfconsistently, but only in the linear approximation assuming that the strain leads only to a negligible perturbation of the two ion interaction Hamiltonian.

The length change $\Delta L/L$ of a sample in a dilatometer experiment can be calculated from the strain tensor components using ¹³

  $$\frac{\Delta L}{L}=\sum_{\alpha\beta} \epsilon_{\alpha\beta} \hat L_{\alpha} \hat L_{\beta}$$ (111)

where $\hat \mathbf L$ denotes the unit vector in the direction of measurement.