A new Li-Mn-Ge-O phase solved by combining 3D electron diffraction methods

Pyroxene compounds are a common form of natural minerals and have been studied as such for a long time. More recently the research on quasi-one-dimensional magnetic and multiferroïc materials has renewed the interest in pyroxenes of the stoichiometry AMX2O6 (A = alkali metal, M = transition metal, X = Si or Ge), since the magnetic M3+ ions form chains. Chemical substitution on the A and M sites can change the magnetic coupling along these chains making this system a rich field for the exploration of new phases of interesting magnetic properties [1]. In this work we report the discovery of a new phase in the Li-Mn-Ge-O system. A HP-HT solid state reaction was performed on a mixture of nominal stoichiometry LiMnGe2O6 during 1 h at a temperature of 850⁰C and a pressure of 3 GPa in a belt press. Powder X-ray diffraction yielded a diffractogram that could not be indexed by known phases of this system. An electron diffraction study in a transmission electron microscope was conducted in order to identify any unknown phases. In the case of structures that promise interesting properties a more targeted synthesis can then be undertaken. For the purpose of this work, we studied one of several unknown phases in the powder in more detail. From standard selected area electron diffraction the unit cell was determined to be triclinic with cell parameters a = 2.51 nm, b = 1.30 nm, c = 1.30 nm, α = 96.0⁰, β = 98.8⁰ and γ = 80.8⁰. No comparable unit cell could be found in the databases neither in this system nor with different A, M or X ions. Intensities were recorded by in-zone axis precession electron diffraction and by electron diffraction tomography. Combining the data from both methods yielded the first model of the structure which we will present here.

Identifying almost identical phases by 3D electron diffraction

Magnetically frustrated materials have been the subject of many studies over the last decades. In search for a 3-dimensional quantum spin liquid, where quantum-mechanical fluctuations prevent magnetic order, different phases of stoichiometry Ba3NiSb2O9 have recently [1] been synthesized some of them at high pressure. Two of these phases are hexagonal. The hexagonal phases (space groups P63/mmc and P63mc, respectively) have different structures but cell parameters that differ by less than 1%. Similar phases have been obtained with Cu [2] or Co [3]. These phases are well distinguished by powder X-ray diffraction when they appear in sufficient quantity in a newly synthesized powder. When these phases are present only in minor quantities, which is a common situation when synthesizing new materials, only transmission electron microscopy can give structural information on a very local scale. However, the accuracy of unit cell parameter determination by electron diffraction (usually 1% or worse) and the identical extinction conditions for the 2 space groups don't permit to distinguish between the two phases. Convergent beam electron diffraction could show the difference between the centrosymmetric and non-centrosymmetric space groups provided a suitably oriented particle can be found. In this work we propose a different method of distinguishing structures in such complicated cases by actually solving the structure. Sufficient in-zone axis precession electron diffraction and/or electron diffraction tomography data can be obtained from any crystal regardless of its orientation. In the subsequent structure solution we have tested both space groups. The quality (or absence thereof) of the structure solutions obtained clearly makes it possible to distinguish between the two hexagonal structures.