CaTiOSiO4, A POSSIBLE NON-LINEAR OPTICAL MATERIAL?
Martin Kunz
Labor für
Kristallographie, ETH Zentrum, Sonneggstr. 5, CH-8092 Zürich
CaTiOSiO4, (titanite) has a structural topology, which is very similar to the one of the non-linear optical material KTiOPO4 (KTP). Both structure types are characterized by corner linked octahedral chains, which in turn are mutually connected through PO4 and SiO4 tetrahedra, respectively. The large cavities of these two frameworks are occupied by K+ and Ca2+ cations, respectively. The most striking structural difference between these two compounds is the fact, that in KTP the octahedral chains are built up in a ...-trans-cis-trans-cis-... way, while the Ti-octahedra within the titanite chains are all arranged in trans configuration. A more subtle difference is revealed by comparing the correlation of the out-of-center displacements of the Ti4+ cations within their octahedral surrounding. Such a displacive distortion is characteristic for octahedrally coordinated d0 transition metals1).
The displacement vectors are all sub-parallel in KTP, while in titanite a cooperative distortion is only observed within individual chains. Neighbouring octahedral chains in titanite have their Ti-displacement vectors oriented in an anti-parallel way. It is this subtle difference, which causes an interesting difference in physical behaviour between these two compounds. The parallel displacement vectors in KTP impose an acentric symmetry on its structure, which favours the occurrence of a strong non-linear optical effect. The anti-parallel orientation of the displacement vectors in titanite, on the other hand, prevents this structure to crystallize in an acentric structure, thus inhibiting any technically interesting optical properties.
In order to explore any possibilities to tune the titanite structure into a technically interesting acentric structure, we have to understand the mechanisms, which control the correlation between the Ti4+ out-of-center displacements. Such an understanding is linked to the understanding of the driving forces of structural phase transitions, which in turn are closely related to changes in the Ti-displacement.
Phase transitions in KTP are known both at high pressure2) and high temperature3). The high temperature transition at 980°C is most probably caused by a disordering of the out-of-center displacement vectors. It involves a symmetry change from the acentric space group Pn21a to centric Pnna. The high pressure transition at 5.8 GPa, on the other hand, is an isostructural first order transition which is mainly revealed by a large shift of the K+ cation and interpolyhedral torsions of the TiO6-PO4 framework.
Titanite has a well studied phase transition at 490 K4). This transition is - similar to the high temperature transition in KTP - caused by a disordering of the orientation of the Ti-displacement vectors3). At higher temperatures, a further phase transition is described to occur at 850 K, which probably is connected to a dynamic disorder between the two possible distortion vectors for each individual Ti4+-cation5). Both, static and dynamic disorder of the out-of-center distortion cause a symmetry change from P21/a to A2/a.
In order to compare the structural behaviour of titanite with the one of KTP, we performed a series of experiments.
We first examined the effect of a dilution of the density of distortion vectors on the coherence of the out-of-center distortion in titanite, by gradually substituting Ti4+ by Sn4+.6) The change of the average symmetry from P21/a to A2/a even on a TEM scale indicates a loss of the coherence of the out-of-center distortion at a Sn concentration of as little as 10%. The low Sn-concentration necessary to disturb the displacement correlation suggests an only weak and short ranging interaction between the individual displacement dipoles.
In a second experiment, we compared the high-pressure behaviour of titanite with the one of KTP7). Similar to KTP, titanite undergoes a pressure driven phase transition at about 3.5 GPa. In contrast to KTP, however, the high-pressure phase of titanite involves a symmetry change, which is identical to the one at high-temperature, i.e. from spacegroup P21/a to A2/a. A closer crystal chemical analysis reveals a difference between the titanite high-temperature phase and its high pressure phase, in that high pressure seems to induce a suppression of the out-of-center distortion rather than a static or dynamic disorder as it is observed at high temperature.
A third experiment investigated the connections between the high-temperature and the high-pressure phase. It revealed a negative P-T slope of the high-pressure phase transition of about -175 KGPa-1. This extrapolates to a zero-pressure transition of about 900 K, which suggests a link between the high pressure phase-transition and the 850 K transition. The behaviour of the 490 K transition with increasing pressure is unclear as yet and remains to be explored. Our data did not indicate any change in slope of the P-T-boundary. A possible isostructural phase transition at high-pressure and high-temperature as observed for example in anorthite is therefore not supported by these data.
An intriguing question emerging from this set
of experiments is, whether the existence of an undistorted phase
at high-pressure and high-temperature could be exploited to tune
the CaTiOSiO4 compound into an acentric structure,
e.g. by applying an external static electric field to the
undistorted phase and decompressing or cooling it into the
distorted state. If this indeed induces an acentric structure,
CaTiOSiO4 is a candidate for a new non-linear optical
material.