This work presents study of magnetocaloric properties of compounds exhibiting unusual magnetic characteristics. Several systems were studied: TbNiAl, (Tb,Y)NiAl, TbNi(Al,In), TbFeAl, GdFeAl, Ni2MnGa based compounds and amorphous metallic alloy Gd-Co-Al-Y. Influence of magneto-crystalline anisotropy on magnetocaloric effect was studied on singlecrystalline sample of TbNiAl compound. The maximum of entropy change was measured with orientation of magnetic field along c axis and compared to polycrystalline sample a 100% increase was observed. Substitutions of Y and In in TbNiAl led to a change of magnetic ordering. Both substitution led to increase of RCP values of studied material. TbFeAl and GdFeAl compounds, characterized by partial disorder of Fe and Al atoms, showed magnetocaloric effect in wide temperature region which led to interesting values of RCP~350 J kg-1 . Very broad region of magnetocaloric effect was also observed on amorphous metallic material of Gd-Co-Al-Y. The effect of Er substitution in Ni2MnGa based compound on its magneto-structural transition, connected with inverse magnetocaloric effect, was studied. The direct measurement method of adiabatic temperature change is described and instrument for such measurement, developed in collaboration with FZU AVČR, v.v.i., is presented.

Ever since the discovery of exotic phenomena like high temperature (Tc) superconductivity in the cuprates and colossal magnetoresistance in the manganites, strongly correlated electron systems have become the center of attention in the field of condensed matter physics research. This renewed interest has been further kindled by the rapid development of sophisticated experimental techniques and tremendous computational power. Computation plays a pivotal role in the theoretical investigation of these systems, because one cannot explain their complicated phase diagrams by simple, exactly solvable models. Among the plethora of experimental techniques, various kinds of high energy electron spectroscopies are fast gaining importance due to the multitude of physical properties and phenomena which they can access. However the physical processes involved and the interpretation of the spectra obtained from these spectroscopies are extremely complex and require extensive theoretical modelling. This thesis is concerned with the theoretical modelling of a certain class of high energy electron spectroscopies, viz. the core-level electron spectroscopies, for strongly correlated systems of various kinds. The spectroscopies covered are Auger electron spectroscopy (AES), core-level photoemission spectroscopy (core-level PES) and X-ray absorption spec- troscopy (XAS), which provide non-magnetic information, and also X-ray magnetic circular and linear dichroism (XMCD and XMLD), which provide magnetic information. .