Hollandites(OMS-2) are an intriguing class of energy storage materials with a tunnel structure permitting one-dimensional insertion and deinsertion of ions and small molecules along the c direction.  Previous research demonstrated the ability to control crystal size of Ag_xMn₈O₁₆ by manipulation of Ag content over a range of 1.0 ≤ x ≤ 1.8 where lower Ag content was accompanied by small crystallite size and higher silver content demonstrated large crystallite size (Figure 1a-c) Notably, we recently observed a seven-fold increase in delivered capacity for Li/Ag_xMn₈O₁₆ electrochemical cells (160 versus 23 mAh/g) upon a seemingly small change in silver content (x ~1.1(L-Ag-OMS-2) and 1.6(H-Ag-OMS-2)) (Figure 1f).  This led us to characterize the structure and defects of the silver hollandite nanorods through the combined use of local (atomic imaging, nanodiffraction, electron energy-loss spectroscopy) and bulk (synchrotron based x-ray diffraction, thermogravimetric analysis) techniques.  Selected area diffraction and high resolution transmission electron microscopy show a structure consistent with that refined by synchrotron based XRD (NSLS II), however the Ag occupancy varies significantly even within neighboring channels.  Both local (EELS) and bulk measurements (TGA) indicate a greater quantity of oxygen vacancies in L-Ag-OMS-2, resulting in lower average Mn valence relative to H-Ag-OMS-2.  Electron energy loss spectroscopy shows a lower Mn oxidation state on the surface relative to the interior of the nanorods (Figure 1d, e).  Considering Ag occupancy and oxygen vacancies, the average Mn valence can be estimated to be about Mn^3.7+ for H-Ag-OMS-2 and Mn^3.5+ for L-Ag-OMS-2 nanorods, respectively.  The significance of the oxygen vacancies can be best understood through consideration of the crystallographic structure of silver hollandite.  Silver hollandite has a body-centered tetragonal structure where the Ag is surrounded by eight MnO₆ octahedra, forming a tunnel with a dimension of ~0.51 nm in the ab plane which extends along the [001] direction.

Though there is a small gap (less than 0.06 nm) in the a or b direction, in the ab plane all directions are basically closed by MnO₆ octahedra. Therefore, if the nanorod has no defects, the diffusion of Li ions in ab plane would be limited. The oxygen vacancies and MnO₆ octahedra distortion may open the MnO₆ octahedra walls, facilitating Li ion diffusion in the ab plane (Figure 1g, h). Thus, the higher capacity of the L-Ag-OMS-2 samples may be due to not only to the smaller crystallite size but also to the presence of more oxygen vacancies as compared to the H-Ag-OMS-2 samples.  Thus the oxygen vacancies and MnO₆ octahedra distortion are assumed to open the MnO₆ octahedra walls, facilitating Li diffusion in the ab plane.  In-situ transmission electron microscopy was used to probe lithium transport and phase evolution in silver hollandite, Ag_1.63Mn₈O₁₆, nanowires by nanoscale imaging, diffraction and EELS spectroscopy.  Additionally, the findings were further rationalized through density functional theory (DFT).  These results indicate both crystallite size and surface defects are significant factors affecting battery performance.

Figure 1 caption:   Typical TEM bright field images from (a) H-Ag-OMS-2 and (b) L-Ag-OMS-2.  (c) HRTEM image of L-Ag-OMS-2 nanorod. (d) Relative composition map calculated from EELS spectrum across L-Ag-OMS-2 nanorod. (e) HAADF of H-Ag-OMS-2 along [001], Ag (red) and Mn (green) indicated. (f) Discharge curves of Li/Ag_xMn₈O₁₆ (L-Ag-OMS-2, H-Ag-OMS-2) batteries.  (g) Conceptual perspective along [001], with Ag (red), Mn (green), O (yellow), (h) with oxygen vacancy facilitating ab plane diffusion.