Magneto-structural distortions in the noncollinear hexagonal antiferromagnet UPdSn

Abstract:

The intermetallic compound UPdSn has been studied by means of high-resolution neutron powder diffraction at low temperature, and both orthorhombic and monoclinic structural distortions have been observed. In addition to confirming the previously published noncollinear antiferromagnetic structures with these symmetries, full Rietveld refinements including the magnetic cross sections have been made on the new data, jointly with older low-resolution data. The structure is monoclinic below 25 K with space group P21 and magnetic symmetry Pc21, orthorhombic between 25 and 40 K with structural space group Cmc21 and magnetic space group Pcm'c21, and paramagnetic above 40 K with space group P63mc. The temperature variations of the orthorhombicity (b/Vr3- a) and monoclinicity (3' -90) parameters have been extracted and the monoclinicity is linearly coupled to the x-component of the uranium magnetic moment /.Lx, the 'magnetic monoclinicity' order parameter. In contrast, the orthorhombicity seems to be coupled to the total uranium magnetic moment. The sign of the coupling between /~x and 3'- 90 has also been determined. It is positive, that is the projected moments prefer to point across the shorter diagonal of the monoclinic basal plane, rather than the longer diagonal.

To download the article click on the link below:

http://www.sciencedirect.com.sci-hub.cc/science/article/pii/0304885393908694

UPdSn forms in the non-centric hexagonal space group P63mc in the GaGeLi structure type and undergoes two magnetic phase transitions at approximately 40 and 25 K. The uranium atoms lie on a simple hexagonal lattice, with two uranium layers per crystallographic unit cell. Our previous low-resolution neutron powder diffraction work [3] showed that the diffraction data can be analysed in terms of a canted antiferromagnet with orthorhombic symmetry (phase I) between 
25 and 40 K, and as a canted antiferromagnet with monoclinic symmetry (phase II) below 25 K.

The  observed  magnetic  reflections  can  also  be indexed  in  a  larger  hexagonal  unit  cell, with  a'  =  2a  and  with  four  times  the  volume  of  the  crystallographic  cell,  in  which  case  the  magnetic  structure would  have  hexagonal  symmetry.  In  this  case, the  magnetostrictive  effects  of  the  ordered  magnetism  would  give  no  change  in  crystal  symmetry as  one  goes  from  the  paramagnetic  phase  to phase  I  and  then  to  phase  II.  But,  if  our  model for the magnetic structures  is correct,  there  should in  principle  be  an  orthorhombic  distortion  in phase  I  and  a  further  monoclinic  distortion  in phase  II.  These  changes  of  crystallographic  symmetry  will  be  manifested  by  splittings  of  the hexagonal  reflections.  It  is  the  purpose  of  this paper  to  report  such  splittings,  and  thereby  establish  definitively  that  our  magnetic  cell  is  to  be preferred  over  a  larger  hexagonal  one.