WorksPhysics and Astronomy
What is a tunnel? A comment on frequent imprecision when speaking about tunnels, at the example of two representative structures in the octahedral molecular sieve group
Abstract
A paper by Su et al (2013)[1] is focused on the presentation of \b{eta}-MnO2 (pyrolusite) as a new high-capacity electrode for Na batteries, arguing that there is plenty of place in the tunnels to host sodium, which is clearly contrary to general knowledge. Similar imprecisions are found in other work so that it might be useful to shortly address the issue of cavities and tunnels in a general way. Here, two structures of the octahedral molecular sieve group (\b{eta}-MnO2 pyrolusite and α-MnO2 hollandite) are taken as an example and analysed at some depth regarding their tunnel structure using advanced software. Tunnels and their occupation are of fundamental importance for the application of such materials as electrodes or in strongly correlated electron systems.
“The excellent electrochemical performance of -MnO nanorods could be ascribed to the compact and dense () tunnel-structure in -MnO crystals. The radius of Na ion ( Å) is much smaller than the size of the () tunnel ( Å Å)30,31. Therefore, Na ions can facilely insert and extract along the () tunnels in -MnO nanorods.”
The preceding string is cited from Su et al. [1][1], for the references see original paper).
Confusion is made with the use of radius and diameter and, more importantly, between interatomic distances and purely geometric measures, including projections, which serve to describe a cavity or bottleneck. In fact, the above cited distances of Å Å are normally[2, 3] used to give an idea of the dimensions of the tunnel and refer to the width of the octahedral chain projected on the (001) plane (see Figure 1c). This width is derived from an O-O distance measured between the centres (nuclei) of two oxygen atoms, so conceptually, there is no way to compare this with ionic radii or diameters used in space-filling considerations.
Figure 1. K-hollandite (a) with and tunnels evidenced, and (b) with the and the small perpendicular tunnels evidenced; manganese octahedra sky blue, K red (half-filled), oxygen yellow, polyhedral bonds green. (c) Pyrolusite with the small perpendicular tunnels evidenced between the tunnels. There are no cations filling the tunnels in either hollandite nor pyrolusite, the green spheres just indicate the centres of the respective tunnels. Graphics prepared using VESTA[21].
In the following, I will try to re-assess some terminology and give a detailed description of the tunnel structure in the octahedral molecular sieve frameworks of pyrolusite and hollandite, with the help of advanced software. Materials in this family are presently much studied for their use as insertion electrodes in new batteries and, due to mixed valence Mn/Mn and pronounced structural anisotropy, they are also interesting as strongly correlated electron systems[4]; for both applications, knowledge of the structure and occupation of the tunnels is fundamental.
Octahedral molecular sieves or tunnel oxides are a well known structural family to which many manganese oxides belong. All of these structures are formed by edge-sharing octahedral chains running along [5]5. In -MnO (pyrolusite, Figure 1c), single chains share their corners to form tunnels, and in -MnO (usually cited as hollandite framework), there are double chains forming tunnels (Figure 1a and Figure 1b). To date, six further (large) tunnel sizes are known in the family (, , , , and ). For all structures with large tunnels, many guests have been reported to occupy the cavities but for pyrolusite, cavities in the tunnels are very small and no stable structures with guests are known with the exception of MnOOH manganite[6] where H occupies a site between two oxygen atoms facing the tunnel at 2.60 Å, and even for the ramsdellite structure ( tunnel), only Li is known to occupy the cavity, in a distorted tetrahedron with Å[7].
What are tunnels?
A pore system is generally defined by cavities and bottlenecks between adjacent cavities. The cavities are lined with ligands like oxygen which serve to hold the guest atoms somewhere in their interior, related with the stability of the compound which is seen in ion exchange isotherms, BET nitrogen adsorption measures, phase transitions and the like. The bottlenecks, on the other hand, are the windows defining the width of the tunnel along which intracrystalline diffusion of the guest may take place and which is related with an activation energy obtained from diffusion or ionic conductivity experiments. Tunnels may form as a linear 1D sequence of cavities, but there are also zig-zag tunnels and crossing tunnel systems.
To describe a tunnel, we consider its windows and use either topology, such as the ring sizes given above (e.g. for a ring made up by four corner sharing octahedra), or a metrical measure which gives the dimensions in Ångström.
In hollandite, e.g., the window has topology and its dimensions are usually[2, 3] cited as Ų, the projected width of the double chain. The cavity corresponding to the tunnel is a cube shortened along with ų (Figure 1a), coordinating the centre at Å which is optimal for potassium in coordination VIII ( Å, exactly the radius given by Shannon[8]). The window is a square with diagonal Å which leaves an open radius of Å, a little small for potassium with Å, but open for sodium and silver or smaller ions.
The window in pyrolusite, which is also present in hollandite, is usually cited to have the (projected) width Ų. The cavity in this tunnel is a tetrahedron flattened along with two edges of Å () and four edges of Å (Figure 1a), and with a radius of Å which leaves an open radius of Å, much smaller than the Shannon[8] radius of sodium, Å, or even lithium, Å. Sodium is therefore never expected to be found in the pyrolusite tunnels: from DFT calculations[9], the cavity has a eV higher energy than the cavity. For Li, the energy difference is much lower ( eV) and indeed, there is experimental evidence[10, 11, 12] that Li might be reversibly inserted in the cavities under very special circumstances, as well as there exist rutile structures based on octahedral M ions with larger radii than Mn (- instead of Å[13, 14]).
The bottleneck (window) of the tunnel is defined by the triangular face of the tetrahedron (, and Å) which leaves a free radius of Å, resulting in a zig-zag tunnel along . Such complicate tunnel structures can be described with appropriate software[15, 16, 17, 18, 19, 20] (for others see reference[20]) which give the exact measures of cavities and bottlenecks along with a graphical representation (e.g. Figure 2).
Figure 2. Representation of tunnels in -MnO, showing a and a tunnel. The tunnel is the big black hole in the large ring at left, and the tunnel is actually reduced to a small, almost invisible, zig-zag tunnel parallel with lateral pockets. Graphics prepared using Jmol[22], as specified in the Appendix (isosurface at Å from O(Mn), respectively).
Figure 2 compares the tunnels along in the two cases. The tunnel is linear and easily described by the cavities and bottlenecks, whereas the small and contorted tunnel is difficult to visualize and not well separated from alternative tunnels running perpendicular to (there are tetrahedral cavities, evidenced in Figure 1b and Figure 1c, only little smaller with instead of Å, the radius defined by the tunnels). There is an interesting case of stuffed derivative of the pyrolusite-rutile structure, [23], where the tetrahedra fill a much enlarged2 tunnel with Cu occupying approximately the pockets seen in Figure 2. From the picture of the tunnel in Figure 2, this compound might have been predicted.
In the preceding, the two octahedral molecular sieve structures of hollandite and pyrolusite were used to demonstrate new ways for an exhaustive description of their tunnel systems, with the help of modern software most of which is freely available. It is hoped that such considerations will become standard in the future, similar to common practice in the tunnel structures of zeolites[24].
Acknowledgements
I thank Ángel Herráez and Robert Hanson for their kind help in the use of program Jmol.
Appendix
For convenience, the Jmol code to produce the cavity image in Figure 2 is included here.
$ load myhollanditeframework.cif {2,1,2}; set window 800 800
$ connect delete; connect 2.2 (_Mn) (_O); connect single radius 0.10
$ {_Mn}.radius=0.18; {_O}.radius=0.46
$ {_Mn}.vdWradius=0.53; {_O}.vdWradius=1.37
$ isosurface resolution 10 pocket cavity 0.31; isosurface subset [2, 5]
$ write IMAGE 2400 2400 JPG 85 "2x2+1x1pocket.31"References
- [1]1 D. Su, H.-J. Ahn and G. Wang, Hydrothermal synthesis of α-MnO₂ and β-MnO₂ nanorods as high capacity cathode materials for sodium ion batteries, J. Mater. Chem. A, 2013, 1, 4845–4850, doi.org/10.1039/c3ta00031aDOI
- [2]2 S. L. Brock, N. Duan, Z. R. Tian, O. Giraldo, H. Zhou and S. L. Suib, A review of porous manganese oxide materials, Chem. Mater., 1998, 10, 619–2628, doi.org/10.1021/cm980227hDOI
- [3]3 O. Ghodbane, J. L. Pascal and F. Favier, Microstructural effects on charge-storage properties in MnO₂-based electrochemical supercapacitors, ACS Appl. Mater. Interfaces, 2009, 1, 1130–1139, doi.org/10.1021/am900094e
- [4]4 J. Matsuno, J. Fujioka, T. Okuda, K. Ueno, T. Mizokawa and T. Katsufuji, Strongly correlated oxides for energy harvesting. Sci. Technol. Adv. Mater., 2018, 19. 899–908, doi.org/10.1080/14686996.2018.1529524DOI
- [5]5 M. Pasero, A short outline of the tunnel oxides, Rev. Mineral. Geochem., 2005, 57, 91–305, doi.org/10.2138/rmg.2005.57.9DOI
- [6]6 T. Kohler, T. Armbruster and F. Libowitzky, Hydrogen bonding and Jahn–Teller distortion in groutite, α-MnOOH, and manganite, γ-MnOOH, and their relations to the manganese dioxides ramsdellite and pyrolusite, J. Solid State Chem., 1997, 133, 486–500, doi.org/10.1006/jssc.1997.7516
- [7]7 J. Akimoto, Y. Gotoh, M. Sohma, K. Kawaguchi, Y. Osawa and H. Takei, Synthesis and crystal structure of ramsdellite-type Li₀.₅TiO₂, J. Solid State Chem., 1994, 110, 150–155, doi.org/10.1006/jssc.1994.1149DOI
- [8]8 R. D. Shannon, Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides, Acta Crystallogr. A, 1976, 32, 751–767, doi.org/10.1107/S0567739476001551DOI
- [9]D. A. Tompsett and M. S. Islam, Electrochemistry of hollandite α-MnO₂: Li-ion and Na-ion insertion and Li₂O incorporation, Chem. Mater., 2013, 25, 2515–2526, doi.org/10.1021/cm400864n
- [10]W. Tang, X. Yang, Z. Liu and K. Ooi, Preparation of β-MnO₂ nanocrystal/acetylene black composites for lithium batteries, J. Mater. Chem., 2003, 13, 2989–2995, doi.org/10.1039/B306780GDOI
- [11]F. Jiao and P. G. Bruce, Mesoporous crystalline β-MnO₂ — a reversible positive electrode for rechargeable lithium batteries, Adv. Mater., 2007, 19, 657–660, doi.org/10.1002/adma.200602499
- [12]T. X. T. Sayle, R. R. Maphanga, P. E. Ngoepe and D. C. Sayle, Predicting the electrochemical properties of MnO₂ nanomaterials used in rechargeable Li batteries: simulating nanostructure at the atomistic level, J. Am. Chem. Soc., 2009, 131, 6161–6173, doi.org/10.1021/ja8082335DOI
- [13]D. W. Murphy, F. J. Di Salvo, J. N. Carides and J. V. Waszczak, Topochemical reactions of rutile related structures with lithium, Mater. Res. Bulletin, 1978, 13, 1395–1402, doi.org/10.1016/0025-5408(78)90131-9
- [14]I. J. Davidson and J. E. Greedan, Neutron and X-ray powder diffraction study of LiₓRuO₂ and LiₓIrO₂: the crystal structure of Li₀.₉RuO₂, J. Solid State Chem., 1984, 51, 104–117, doi.org/10.1016/0022-4596(84)90321-9DOI
- [15]R. Basso and A. Della Giusta, A computer program for the determination of cavity shapes and dimensions in crystal structures, J. Appl. Cryst., 1977, 10, 496, doi.org/10.1107/S0021889877014095DOI
- [16]J. Rius, Eine auf Fouriermethoden basierende Lückensuchfunktion, J. Appl. Cryst., 1991, 24, 187–189, doi.org/10.1107/S0021889890012304DOI
- [17]D. Y. Naumov and E. V. Boldyreva, Software for localization and visualization of cavities in crystal structures and supramolecular assemblies, J. Structural Chem., 1999, 40, 86–93, doi.org/10.1007/BF02700783DOI
- [18]A. Herráez, How to use Jmol to study and present molecular structures, Vol. 1, Lulu Enterprises, Morrisville, NC, USA, 2007, 144p., ISBN 978-1-84799-259-8
- [19]R. M. Hanson, Jmol – a paradigm shift in crystallographic visualization, J. Appl. Cryst., 2010, 43, 1250–1260, doi.org/10.1107/S0021889810030256DOI
- [20]J. B. Maglic and R. Lavendomme, MoloVol: an easy-to-use program for analyzing cavities, volumes and surface areas of chemical structures, J. Appl. Cryst., 2022, 55, 1033–1044, doi.org/10.1107/S1600576722004988DOI
- [21]K. Momma and F. Izumi, VESTA 3 for three-dimensional visualization of crystal, volumetric and morphology data, J. Appl. Cryst., 2011, 44, 1272–1276, doi.org/10.1107/S0021889811038970DOI
- [22]Jmol: an open-source Java viewer for chemical structures in 3D, http://www.jmol.org (version 16.1.5).
- [23]A. Mewis, Eine mit tetraederketten aufgefüllte rutilstruktur: die verbindung CaCu₄P₂ ≅ CaP₂(Cu₄), Z. anorg. allg. Chem., 1987, 546, 43–46, doi.org/10.1002/zaac.19875450205DOI
- [24]International Zeolite Association, Database of Zeolite Structures, www.iza-structure.org