RESUMO
Magnetic, structural and energetic properties of bulk Fe and Cr were studied using first-principles calculations within density functional theory (DFT). We aimed to identify the dependence of these properties on key approximations of DFT, namely the exchange-correlation functional, the pseudopotential and the basis set. We found a smaller effect of pseudopotentials (PPs) on Fe than on Cr. For instance, the local magnetism of Cr was shown to be particularly sensitive to the potentials representing the core electrons, i.e. projector augmented wave and Vanderbilt ultrasoft PPs predict similar results, whereas standard norm-conserving PPs tend to overestimate the local magnetic moments of Cr in bcc Cr and in dilute bcc FeCr alloys. This drawback is suggested to be closely correlated to the overestimation of Cr solution energy in the latter system. On the other hand, we point out that DFT methods with very reduced localized basis sets (LCAO: linear combination of atomic orbitals) give satisfactory results compared with more robust plane-wave approaches. A minimal-basis representation of '3d' electrons comes to be sufficient to describe non-trivial magnetic phases including spin spirals in both fcc Fe and bcc Cr, as well as the experimental magnetic ground state of bcc Cr showing a spin density wave (SDW) state. In addition, a magnetic 'spd' tight binding model within the Stoner formalism was proposed and validated for Fe and Cr. The respective Stoner parameters were obtained by fitting to DFT data. This efficient semiempirical approach was shown to be accurate enough for studying various collinear and non-collinear phases of bulk Fe and Cr. It also enabled a detailed investigation of different polarization states of SDW in bcc Cr, where the longitudinal state was suggested to be the ground state, consistent with existing experimental data.
RESUMO
Crude death rates show great discrepancies in diabetic mortality between industrialized and developing countries, the latter being less affected by diabetes mellitus. Nevertheless, when age is controlled, the gap becomes much narrower. For instance, the crude death rates in males by diabetes mellitus rise to 20.6 per 100,000 population in Belgium, and only 5.8 in Costa Rica (difference: 14.8). But after controlling age, the gap becomes only 2.8 (13.4 in Belgium and 10.6 in Costa Rica). As a matter of fact, the main difference between industrialized countries and those in the Third World is that diabetic mortality occurs at older ages in the former and at younger ages in the latter. Generally speaking, diabetic mortality is increasing in the developing countries, due most probably to improved data collection. In France, diabetic mortality was growing in the "départements" of the Nord Est and the Languedoc, and many rural areas were also experiencing a rising mortality rate. On the other hand, the rate was decreasing in Brittany and the Paris area.
