Pressure induced isostructural transitions showing a homothetic volume contraction are extremely unusual. In fact, the usual pressure-induced phase transitions in crystalline solids are first order type, having a displacive nature associated with a change of structure. What are the necessary elements for a homotethic volume collapse of isostructural nature? First of all, the homothety constraint makes cubic structures the only good candidates for such an isostructural transition. But as the atomic structure is not varied, the only way to modify the properties of the crystal to allow for a volume collapse is "to modify the atoms". This is in fact what has been proposed for the few cases where pressure-induced isostructural volume collapse of cubic structures have been observed (Sm chalcogenides, Ce). This original idea was first proposed by E. Fermi for the Cs case, in which pressure gives rise to an inversion between the s and d atomic orbitals of the atom. Finally, for this particular case, recent work has shown the non-isostructural character of the transition.

The structure of the Ba8Si46 type-I silicon clathrate can be viewed as the arrangement of face sharing "super-atoms" made of nano-cages of 20 or 24 silicon atoms each containing one guest Ba at the centre (see Figure 109). It constitutes an open form of sp3 silicon allowing for endohedral intercalation. The application of pressure modifies the size of the nano-cages. This opens the way for the existence of a critical size by which the nature of the "super-atom" can be varied due to a modification of the hybridisation between the guest atom and the host cage.

 

 

Fig. 109: The cubic type-I clathrate structure of Ba8Si46.

 

We have performed high-pressure X-ray absorption and X-ray diffraction studies on the Ba8Si46 clathrate. The experiments were performed at beamlines ID24 (Ba L-III edge), BM29 (Ba K-edge) and ID30 (angular dispersive X-ray diffraction). The main conclusion has been the unambiguous determination of the homothetic character of the volume collapse transition [1] taking place at 11.5-14.0 GPa (Figure 110). The comparison of our results with high-pressure Raman experiments [2], allowed us to explore the details of the driving mechanism of such an original phase transformation. A change of hybridisation in the Ba@Si20 "super-atom" between Si-p and Ba-d orbitals appears as the most probable mechanism.

 

 

Fig. 110: Lattice parameter variation of the clathrate structure of Ba8Si46 obtained by X-ray diffraction and EXAFS at the Ba K-edge assuming that the Ba atom is at the centre of the "super-atoms". Inset: observed behaviour with pressure of the silicon atomic positions, showing the homothetic evolution of the silicon cages through the isostructural volume collapse.

 

A further low-pressure anomaly observed in Raman experiments [2], was also visible at the Ba L-III edge, showing that the disappearance of the Ba@Si24 quasi-vibronic type excitations at 5-7 GPa is coupled with a change of hybridisation of the Ba atom within the larger cage. This transformation, also of isostructural nature, does not introduce any observable effect in the X-ray diffraction patterns.

The effect of barium intercalation stabilises the clathrate structure up to pressures three times higher than the ones observed for the empty silicon clathrate structures and allows tetrahedral silicon with record interatomic distances as low as 2.13 Å to be obtained. At 49 GPa, the highest pressure studied, the structure becomes irreversibly amorphous.

References
[1] A. San Miguel, P. Mélinon, D. Connétable, X. Blase, F Tournus, E Reny, S Yamanaka and J.P. Itié, Phys Rev B 65, 054109 (2002).
[2] T. Kume, H. Fukuoka, T. Koda, S. Sasaki, H. Shimizu and S. Yamanaka, Phys. Rev. Lett. 90, 155503 (2003).

Principal Publications and Authors
A. San Miguel (a), P. Toulemonde (a), A. Merlen (a), T. Kume (a), S. Le Floch (a), A. Aouizerat (a), G. Aquilanti (b), S. Pascarelli (b), J.P. Itié (c), S. Yamanaka (d) Europhys. Lett., 69(4), (15 Feb. 2005).
(a) Laboratoire de Physique de la Matière Condensée et Nanostrucutres, University of Lyon and CNRS (France)
(b) ESRF
(c) Soleil, Saclay (France)
(d) University of Hiroshima (Japan)