Investigation of the spin-Peierls system TiOCl under
pressure and with alkali doping
While in the studies of complex magnetic materials,
the extraction of model Hamiltonian parameters was central to the
studies and the work on the structural properties had only a
preparatory role in the investigation, this activity deals with a Mott
insulator (TiOCl) where the focus is on the relationship between
structural and electronic properties and where thus the careful
prediction of structures is crucial.
The oxyhalide TiOCl has been intensively studied recently since it is
the third known inorganic spin-Peierls system. It is of special
interest because it shows a very large spin-Peierls distortion (η
= 0.18 Å) and because the room temperature paramagnetic Mott
insulator is separated from the low temperature nonmagnetic dimerized
phase by an incommensurate phase for
66 K < T < 92 K [1].
a) Underlying Hamiltonian
In a first study we determine the parameters of the two-dimensional
spin-Peierls model for TiOCl. This involves the determination not only
of three exchange coupling parameters but also of two spin phonon
couplings and an elastic constant. The latter are determined by
distorting the structure along the directions of the spin Peierls
phonon (the phonon mode that has been shown to lead to the
spin-Peierls dimerization) and calculating total energies. We perform
mean field and exact diagonalization calculations for the model we
obtain, and we find the parameters to be fully consistent with the
experimentally observed spin-Peierls distortion. Our investigation
also allows us to identify the phonons as adiabatic (they behave
classically). This study is published in Ref. [ZJV08a].
b) High pressure structure and properties
In a second study, we address the controversy about the
insulator-to-metal transition that was observed under
pressure [2,3]. We perform constant pressure Car
Parrinello molecular dynamics
calculations [4,5,6] in order to determine
the high pressure structures of TiOCl. We find that at a critical
pressure, a first order phase transition from the orthorhombic
structure at ambient pressure to a dimerized monoclinic structure at
high pressure occurs. This structural change is accompanied by a
rearrangement of the Ti t2g orbitals and a closing of the gap,
leading to an anisotropic metallic state. At even higher pressures,
there is a second phase transition back to a high symmetry metallic
phase. The structural transition to a dimerized phase at high
pressures has since been confirmed by experiment [3].
This study is published in Ref. [ZJV08b]. In this study, the
treatment of the strong correlations with an LDA+U functional is
essential, and the success of the prediction of a high pressure
structure makes the approach useful for many other problems. As the correlation
strength can be measured in terms of the ratio U/t with site
interaction strength U and hopping parameter t, increasing
pressure leads to increasing t and is therefore a way of tuning the
strength of the correlations. Thus, the methods that were tested on
TiOCl have significant future potential.
c) Can TiOCl be metallized via doping?
In a third study, we investigate the effect of alkali doping on TiOCl.
Doping carriers into low dimensional Mott insulators like the layered
TiOCl is of fundamental interest; the most spectacular effect of such
doping is the high temperature superconductivity found when electrons
or holes are doped into the CuO2 planes of cuprates. The advantage
of studying carrier doping in a system with a small unit cell like
TiOCl is that the structure upon doping can be determined by first
principles molecular dynamics, and the mechanism of the electronic
structure modification can be studied directly. In the case of TiOCl,
carrier doping by the introduction of alkali ions into the structure
fails to metallize the material because the dopant atoms enter
octahedral cages of halogen atoms and thereby strongly distort the
titanium oxide planes. Consequently, the dopant atoms give their
charge only to the closest Ti atoms whose electronic structure is
strongly modified. The resulting state with localized impurity
potentials remains insulating, in agreement with experiment.
Rückamp, J. Baier, M. Kriener, M. W. Haverkort,
T. Lorenz, G. S. Uhrig, L. Jongen, A. Möller, G. Meyer, M.
Grüninger, Phys. Rev. Lett. 95, 097203 (2005).
S. Blanco-Canosa, F. Rivadulla, A. Pineiro, V.
Pardo, D. Baldomir, D. I. Khomskii, M. M. Abd-Elmeguid, M. A.
Lopez-Quintela, J. Rivas, arXiv:0806.0230v1 (unpublished).
