GLASS FORMING SPIN LIQUID DYNAMICS IN THE FRUSTRATED PYROCHLORE MAGNETS Ho2Ti2O7 AND Dy2Ti2O7
Pyrochlore oxides with a rare earth ion in the A site have been subjects of intense research for the last two decades. In this crystal structure, the magnetic rare earth ions create a corner sharing tetrahedral network and the magnetic state of the system is determined by the combination of multitude of different interactions. Ho2Ti2O7 and Dy2Ti2O7, frequently called spin ices, have interesting configuration of spins at very low temperatures. Due to the geometrical frustration created by the tetrahedral lattice, the strong crystal field interactions and significant dipolar interactions between the spins, the ground state in these materials is not known to be ordered at very low temperatures. Theoretical models have proposed that the lowest energy state has two spins pointing into each tetrahedron and two spins pointing out, and the spin flips from this ground state is equated to magnetic monopole like quasi particles that should propagate through the system by flipping more spins. We created a boundary free, superconducting toroidal experiment to detect the effects of monopole motion, or more generally the magnetization dynamics in these two systems. Our design allows as us to measure the AC susceptibility very precisely and probe the time domain relaxation of magnetization to temperatures as low as 0.6 K. The boundary free geometry removes the ambiguity of the data that comes from demagnetization effects at sample boundaries, which have to be considered when studying rod shaped sample. The time domain magnetization relaxation experiments show Kohlrausch- Williams-Watts (KWW) stretched exponential decay of the rate of change of magnetization with time. This result disagrees with the predictions of the simple monopole model, which expects a regular exponential, however the data from all measured temperatures and for both materials show clearly show this behavior. The AC measurements performed in the temperature range 0.9-2 K and frequency range 10 Hz-100 kHz show that the AC susceptibility of both materials can be explained by the Havriliak-Negami (HN) functional form. The unified analysis of the temperature dependence of the relaxation time from time domain and frequency domain show that at low temperatures T < 2K the relaxation time increases faster than Arrhenius law, and is well described by the Vogel-Tammann-Fulcher (VTF) form. These forms, KWWstretched exponential relaxation, HN susceptibility and VTF trajectory of relaxation time are dynamical identifiers of the glass forming liquids. The liquids cooled below freezing temperature without long range crystalline ordering stay as a metastable liquid, and the functional forms described above are used to characterize such glass forming liquids. Our discoveries show that the magnetic state in Ho2Ti2O7 and Dy2Ti2O7 is the first example of glass forming liquid state in the magnetic systems. This is different from a common frustrated magnetic state, the spin glass, as the latter is usually associated with sharp change in AC susceptibility at some critical freezing temperature and the position of this temperature is sensitive to experimental parameters like the external field and frequency. The magnetic state in the non doped Ho2Ti2O7 and Dy2Ti2O7 does not show characteristics of a spin glass, instead it is better described as the glass forming liquid that one usually associated with liquid silica. Both materials have similar VTF temperature, T0 200mK. We expect that at such low temperature, the magnetic state will completely fall out of equilibrium and the system cannot reach equilibrium regardless of the wait times of the experiment. The Kramers vs non Kramers nature of spins in these materials affects the quantum mechanical transition probability between different ground state configurations. Dy2Ti2O7 which has only Kramers doublets has lower attempt frequency than Ho2Ti2O7 which has both singlets and doublets. The dipolar interactions which act as transverse fields in these materials are able to induce transition more efficiently (proportional to square of the field) in Ho2Ti2O7 when compared to Dy2Ti2O7(proportional to cube of the field). Additionally, the abundance of the monopoles in Dy2Ti2O7 compared to Ho2Ti2O7 leads to segmentation of the subsystems of spins and makes the magnetic liquid in Dy2Ti2O7 more heterogeneous and more fragile. These intriguing discoveries in systems expected to show magnetic monopole dynamics lead to a profound question, are these glass forming liquid properties the result of the monopole dynamics? Perhaps a more detailed monopole theory of the dynamics in these systems is necessary to understand such a state in magnetic pyrochlores. This is something that has to be studied both theoretically and experimentally; in the latter by going to different experimental systems, other magnetic pyrochlore oxides, to see the universality of glass forming liquids properties and by doing measurements that can directly probe the disordered state like tests of two step relaxation.