Crystal Growth of Spin-Frustrated Ba4Nb0.8Ir3.2O12: A Possible Spin Liquid Material

Polycrystalline Ba4NbIr3O12 has recently been shown to be a promising spin liquid candidate. We report an easy and reliable method to grow millimeter-sized single crystals of this trimer based spin liquid candidate material with the actual stoichiometry of Ba4Nb0.8Ir3.2O12. The growth of large crystals is achieved using BaCl2 as flux. The crystals show hexagonal plate-like habit with edges up to 3 mm in length. The structure is confirmed by single crystal X-ray diffraction and is found to be the same as of previously reported phase Ba12Nb2.4Ir9.6O36 [Ba4Nb0.8Ir3.2O12], indeed with a mixed occupancy of Nb/Ir at 3a site. The magnetic and calorimetric study on the individual single crystals confirms the possibility of a spin liquid state consistent with a recent report on a polycrystalline sample


Introduction
Quantum spin liquid (QSL) materials have been a topic of intense research very recently owing to their very rich and interesting magnetic properties like long-range entanglement and fractional quantum excitations without any spontaneous symmetry breaking of the crystal lattice or spins. In simple terms, such materials are characterized by fluctuating spins entangle over long distances without showing magnetic order in the zero-temperature limit. These materials are rare and in fact, the spin liquid state has only been realized experimentally in a handful of compounds mostly with triangular lattice structures. [1][2][3][4][5][6][7] All of these materials possess various degrees of spin frustration. Very recently an Ir-based oxide containing Ir3O12 trimers has been proposed as a candidate material based on the magnetic and calorimetric studies on a polycrystalline sample. 8 Magnetic and calorimetric measurements on this sample suggested spin frustration and no magnetic ordering. Although this compound had been first reported as microcrystalline material with crystals only suitable for diffraction studies, no other properties were reported. 9 High-quality single crystals are thus essential to confirm the spin liquid state by various thermodynamic and microscopic measurement techniques such as specific heat, dynamic susceptibility, neutron scattering, etc at low temperatures. Single crystals are also required to ascertain that there are no other underlying magnetic impurity phases, which may contribute to paramagnetism and give rise to other artifacts in low-temperature measurements.
Also, it is important to eliminate the possibility of a disorder due to non-stoichiometry or unintentional doping during the sample growth, because in frustrated magnets only a tiny disorder can destroy the spin-liquid state and lead to a spin glass state.
In this work, we report the growth of large millimeter-sized single crystals of the title compound and corroborate its properties with the reported work on the polycrystalline sample.
The structure is found to be the same as reported earlier by Müller-Bushbaum, in fact with a significant disorder at the Nb site with Ir occupying as much as 20% of the Nb sites. 9 This disorder was overlooked in the recent study reported by Nguyen et al. 8 However, despite the disorder, the magnetic and specific heat data still strongly suggest a spin liquid state might exist in this compound. 3

Experimental:
Synthesis. Crystal growth was carried out following two methods, both using BaCl2 as flux. In first, a phase pure polycrystalline material was synthesized by heating an appropriate mixture of BaO2, Nb2O5 and IrO2 with a target composition of Ba4NbIr3O12 in an evacuated quartz tube at 1100 o C for 48 h. The obtained polycrystalline sample was then mixed with an excess of BaCl2•2H2O, placed in an alumina crucible, heated to 1100 o C for 24 h, and slowly cooled to 950 o C at various rates (x o C/h) after which the furnace was allowed to cool naturally. Single crystals could also be grown by directly heating BaCO3, IrO2, and Nb2O5 with an excess of BaCl2·2H2O under the same heating conditions. The crystalline product was obtained after dissolving the flux in distilled water and then sonicating in ethanol to remove any polycrystalline matrix or flux.
Characterization: Phase purity and identification of the obtained samples were verified by the powder X-ray diffraction (XRD) technique. The powder patterns were collected in the 2θ range from 3.5 to 100 o at room temperature on a HUBER G670 imaging plate Guinier camera with Cu-Kα1 radiation (λ = 1.5406 Å). Rietveld fitting was performed with the TOPAS-4.2.0.2 (AXS) program. 10 The structure was solved by the Charge-Flipping algorithm 11 and refined with Jana2006 12 based on single-crystal X-ray data. Crystallographic data have been deposited with Fachinformationszentrum Karlsruhe, D-76344 Eggenstein-Leopoldshafen (Germany) and can be obtained on quoting the depository number CSD 1983872.
Compositional analyses of the single crystals from different batches of samples were carried out using SEM-EDS and wet chemical analysis (ICP) techniques.
The crystallographic orientation of the crystal was determined from the Laue diffraction patterns. The plane of the crystal was found to be the crystallographic ab-plane.
Physical property measurements: DC magnetic susceptibility measurements in several fields (μ0H) and temperatures (T) were performed in a SQUID magnetometer (Quantum Design). The magnetization of a single crystal was measured along two different crystallographic orientations with the field ranging from 0.05 to 3.5 T applied parallel and perpendicular to the c-axis (Hab and Hc) and in the temperature between 2 and 320 K in field-cooled protocols (FC)). Fielddependent-magnetization was measured at 2 K in the field range of ± 7 T.
Specific heat (Cp) in the temperature range of 2-20 K was measured using the relaxation method. Electrical resistivity () was measured in the ab-plane in the temperature range of 90-400 K by a four-probe method in a Quantum Design PPMS instrument. When the reaction is carried out starting with a mixture of BaCO3, IrO2, and Nb2O5, the crystal size is generally small and Ir metal is found on the edges or surface of the crystals. Using a polycrystalline powder as charge, this problem is largely eliminated and crystals of larger size are obtained. Larger crystals also tend to grow at a slower cooling rate in combination with the use of polycrystalline material. The quality of crystalline products obtained in each batch was adjudged from the powder X-ray ( Figure S1 in SI) and Laue diffraction patterns ( Figure S2 in SI). Sharp diffraction spots in Laue patterns confirm that the crystals are of high quality. The flat surface of the crystals is found to be the crystallographic ab-plane.   Table S1 in SI.

Results and discussion:
Rietveld refinement of the powder diffraction data was carried out to confirm the structure,  perovskites. 14    To gather more information on the absence of magnetic ordering at low temperatures (down to 2 K) we performed temperature-dependent specific-heat measurements ( figure 5). A smooth dependence of specific heat on temperature and the absence of any anomaly eliminates the possibility of magnetic ordering, strengthening the results of susceptibility measurements. The inset of figure 5 shows that there is a tendency of linearity in the Cp (T) data below ~5 K: a