HRTEM INVESTIGATION OF TOBELITE-2M2: IMPLICATIONS FOR CRYSTAL STRUCTURE REFINEMENT

The crystal structure of tobelite 2M2, a dioctahedral mica with ideal composition (NH4,K)Al2(Si3Al)O10(OH)2, has been refined for the first time via single crystal XRD by Mesto et al. (2006). Those authors refined two single crystals of tobelite in the C2/c S.G., to R-values 6.0% and 7.2%. The average cell parameters were: a = 9.026, b = 5.203, c = 20.721, beta = 100.05°. However, residues in the Fourier density maps pointed to some kind of structure disorder. In the present work a HRTEM investigation was undertaken in order to elucidate the nature of the disorder. This turned out to be quite a difficult task due to the high sensitivity of the samples to electron beam damage (see also Nieto 2002, Fig. 5). After few seconds of exposure to the electron beam, even at very low magnification (i.e., low beam dose), “buttonhole” shaped features develop parallel to (001). They look like inflated pockets, probably due to NH4 + diffusion and accumulation, before volatilization, within the interlayer. In spite of these difficulties, numerous stacking faults parallel to (001) were detected at low magnification and some of them were characterized by high resolution imaging. Other kinds of inhomogeneities, such as intergrowths with other micas have not been observed in the studied samples at the present state of the investigation. Full stacking sequence characterization of micas by HRTEM requires structural images along two not equivalent direction of the [UW0] zone type of the same faulted area (Kogure 2002). For the samples considered here, this turned out as very demanding, because of the aforementioned instability of tobelite under prolonged exposure. This notwithstanding, assuming that stacking faults belonging to subfamily A are absent – a reasonable assumption since diffraction patterns taken along [-130] do not show any streaks or spot splitting along [31l]* rows – most of the near atomic resolution images of stacking faults taken along [010] can be interpreted as interruptions of the alternating ± 60° rotations of the 2M2 stacking sequence resulting from the occurrence of 180° rotations – i.e. nanotwins given by 180°-rotations along c*. These findings can explain the presence of residues in the Fourier density maps in single crystal structure refinements and hopefully would help to improve the disorder modelling for future structural refinements