The use of microgravity in our studies of layered microporous tin(IV) sulphides was stimulated by the sensitivity of their crystal growth to the uniformity of the solution environment and the extreme difficulty of growing diffraction-quality single crystals. It was anticipated that a comparative study of the morphology, surface topology and microstructure of crystals prepared in space and on Earth would provide an insight into the nucleation and growth steps that are affected by gravity-driven convective mass transport and Stokes sedimentation effects. This knowledge could lead to methods for producing better crystals in Earth-based processes.

A number of layered microporous tin(IV) sulphides have recently been synthesized and structurally characterized (Fig. 1top)4,5,6,7,8,9,10,11. The current understanding of their mode of formation is shown in Fig. 1(bottom)9,10. In brief, solution-phase (Sn2S64−) building units ‘snap’ together, as donor–acceptor pairs, to create Sn3S4 broken-cube clusters. These assemble around organic template cations R+(where R+ is TMA+(N(CH3)4+) or TBA (N(C4H9)4+)) to form porous anionic [SnnS2n+12−] sheets. The extra sulphur is eliminated as sulphide. The templates direct the sheets into poorly organized stacks which move into registry in the R2SnnS2n+1 product.

Figure 1: Top, thestructureofthemicroporoussheetsofTM2Sn3S7 (a) and TBA2Sn4S9 (b).
figure 1

The template cations are not shown. Bottom, current understanding of the model of formation of layered microporous tin(IV) sulphides, showing precursors, intermediates and products; modular construction from dimer building-blocks to layer registry is shown (here T stands for template)9,10,11.

To study the effect of microgravity on this process, a total of 19 syntheses were selected based on the results of two years of ground-based study involving variations of reaction stoichiometry, choice of precursors, heating profile, co-solvents and ageing. Synthesis of the orthorhombic TBA2Sn4S9 (space group Pbcn) and TMA2Sn3S7 (P212121) were optimized using the precursors xROH : Sn : yS : zSnS2 : pH2O : qTEG, where R+ is the organic template cation and TEG is the co-solvent tetraethyleneglycol. Each microgravity preparation was run in duplicate for a total of 38 reaction mixtures to check for the reproducibility of the results. Every one of these was repeated, under identical conditions, for the Earth-based study. The reaction mixtures were aged in their crucibles for a period of 79 days during which time they were transferred to the Kennedy Space Center, installed in Endeavour, lifted off into space at 5:30 on 19 May 1996 and heated to reaction temperature on 22 May 1996. The temperature of each tube was independently controlled with a computer which recorded the temperature profile. Samples were heated to 150 °C over 10 hours, maintained at this temperature for 13 hours 23 minutes, and cooled to 25 °C over 24 hours. The space samples were aged for 42 days during which time they were transferred from the Space Shuttle to our laboratory. A special lathe removed the steel caps from the containers for the crucible recovery process. All space products were filtered, washed with water and acetone, air dried and stored under argon in the dark. The reaction conditions and ageing steps were reproduced to the best of our ability for the Earth samples.

Representative scanning electron microscopy (SEM) images of space- and Earth-grown samples of TBA2Sn4S9 are shown in Fig. 2left. The corresponding powder X-ray diffraction (PXRD) patterns are displayed in Fig. 2right. Crystals grown in space have well-formed rectangular plate-like habits. This is the expected morphology for the orthorhombic Pbcn space group and layered structure of the material. The dimensions of the {102} face of the crystals fall in the range 30–40 μm. The SEM images of the Earth-grown crystals reveal significant differences in morphology and size. They have submicrometre dimensions and consist of spherical agglomerates. The PXRD pattern for the microgravity sample displays better-resolved reflections which indicate a much higher degree of layer registry and overall crystallinity. Similar effects are observed in the SEM images and PXRD patterns of TMA2Sn3S7 grown in space and on Earth. The full-width at half-maximum values of the 0k0 reflections that correspond to the interlayer spacing are consistently smaller for the space-grown crystals, demonstrating that they have better layer registry than the Earth analogues. The spacing between the layers in these materials is 9 Å for TMA2Sn3S7 and 14 Å for TBA2Sn4S9 (ref. 11). The anionic layers are held together by weak electrostatic, hydrogen-bonding and van der Waals interactions with the template cations. Organization of the layers into registered stacks is expected to be a low-activation-energy process and sensitive to the uniformity of the solution environment. By contrast, the forces governing ordering within the layers involve strong covalent bonds which are less likely to be perturbed by the growth environment (see below). Note that co-existing spherical particles seen in the SEM images of the Earth-grown samples are unreacted tin (see also PXRD) and are less prevalent in the space samples. Tin particles tend to sediment in a 1g environment and passivation by sulphur can occur. In microgravity, tin particles will instead levitate in solution with less opportunity to passivate and leave unreacted tin behind.

Figure 2
figure 2

Above, scanning electron microscopy images of TBA2Sn4S9 grown in space (left; scale bar 50 μm) and on Earth (right; scale bar 30 μm); far right, powder X-ray diffraction patterns of TBA2Sn4S9 grown in space (top trace) and on Earth (bottom trace). Reflections around 30–32° and 44–46°, 2θ depict elemental tin.

Representative electron diffraction (ED) patterns of space- and Earth-grown crystals of TMA2Sn3S7 are shown in Fig. 3. The beam direction was arranged in both cases to be orthogonal to the dominant {010} face of the hexagonal-shaped crystals. Under similar recording conditions (that is, size and thickness of crystals, beam location on crystals, aperture size and exposure time) the space-grown crystals consistently show higher-intensity ED patterns with no streaking of spots compared with the Earth analogues. The ED pattern of the space crystal corresponds to that expected for the orthorhombic P212121 space group of TMA2Sn3S7. This ED pattern is to be contrasted with the one observed for the Earth-grown crystal where the streaking between ED spots probably originates from interlayer stacking faults (planar defects)12. This difference implies that gravity contributes to the creation of interlayer disorder. Overall, the ED results show better crystallinity and registry of the layers for the space-grown crystals.

Figure 3
figure 3

Electron diffraction patterns of orthorhombic P212121 TMA2Sn3S7 grown in space (left) and on Earth (right).

The adsorption of CO2 at 77 K by TMA2Sn3S7 shows a type I isotherm (diagnostic of a microporous material) for both space and Earth samples. A Langmuir plot of the data shows that the accessible void volume determined for the space sample (0.045 cm3 g−1) is 60% greater than that of the Earth analogue. Better-aligned porous layers are expected to provide less restricted channels running between the sheets. As a result, the CO2 adsorbed into the space-grown material experiences less tortuous diffusion pathways into the structure and has a larger available void volume than the Earth sample.

Atomic-force microscopy (AFM) has been used to quantify the roughness of crystal surfaces over an area of 1× 1 μm2. Averaging has been conducted for five crystal samples with ten area measurements for each. The space-grown crystals consistently have smoother faces (roughness of 5–30 Å) than the Earth-grown analogues (25–100 Å). By contrast, there is little difference between the degree and spatial extent of surface microstructural order in these samples. The surface parallel to the {010} face of space- and Earth-grown TMA2Sn3S7 crystals shows a well resolved orthorhombic pattern with high-contrast features separated by a distance of 13 ± 1.0 Å. This corresponds to the spacing between capping-sulphur atoms of alternate Sn3S4 broken-cube clusters arranged in an up–down configuration around the 24-atom pore (Fig. 1). Similarly, AFM images of the dominant surface parallel to the {102} face of space and Earth TBA2Sn4S9 crystals reveal a well-defined orthorhombic pattern. High-contrast features with a separation of 16.0 ± 1.0 Å and 19.2 ± 1.2 Å correspond to the spacings between the capping-sulphur atoms of alternate Sn3S4 broken-cube clusters that are arranged in an up–down configuration around the 32-atom pore. From these AFM results it can be deduced that the less turbulent solution environment in microgravity enhances the mesoscopic smoothness of the dominant crystal face that contains the layer planes in TMA2Sn3S7 and TBA2Sn4S9. However, the microscopic order within the layers is about the same for the space- and Earth-grown materials (see below).

Polarizing optical microscopy images for TMA2Sn3S7 space- and Earth-grown crystal samples have been recorded. The uniform and complete extinction of white light observed between crossed polarizers for the space-grown samples confirms that the space samples are single crystals. By contrast, the Earth-grown samples display mosaic patterns and birefringence interference colours in their images; these patterns and colours are diagnostic of crystals consisting of multicrystalline agglomerations (this has been confirmed by laser scanning confocal optical microscopy). The difference could originate from increased populations of defects in the Earth-grown samples which serve as secondary nucleation sites for the formation of crystal agglomerates. Photoluminescence studies show that the space samples do indeed contain fewer defects.

Optical reflectance spectroscopy has revealed a pronounced enhancement of intensity on the optical absorption edge but without a shift of the bandgap for Earth-grown compared with space-grown TMA2Sn3S7 samples. This difference probably has its origin in a relaxation of the selection rules for S(−II) to Sn(IV) interband electronic transitions for the Earth samples because of poorer registry of the layers.

The techniques of 119Sn magic-angle-spinning NMR, 119Sn Mössbauer spectroscopy and Fourier-transform Raman spectroscopy proved to be effective probes of local electronic and geometrical structure around the tin(IV) sites in TMA2Sn3S7 and TBA2Sn4S9. No difference was found in the following parameters between space- and Earth-grown samples: NMR isotropic chemical shifts, Mössbauer isomer shifts and quadrupole splittings, Raman vibrational frequencies, intensities and spectral line shapes. The electrical response of these materials to the adsorption of guest molecules has been investigated by an in situ conductivity study. The resistance of TMA2Sn3S7 drops drastically, up to 4–5 orders of magnitude, when the sample in vacuo is exposed to water at 20 °C. The electrical conductivities as a function of partial pressure of the adsorbate for the space-grown and Earth-grown samples are, however, virtually indistinguishable. Because the conductivity of TMA2Sn3S7 is expected to be highly anisotropic with greater intralayer than interlayer charge transport, the similarity of the conductivity isotherms of the space- and Earth-grown samples again implies that intralayer order is not significantly influenced by gravity. Thus different analytical techniques can reveal the steps in the crystallization of a self-assembling microporous layered material that are under the influence of gravity.

Although the results of these experiments have not established that gravity alone is the controlling factor in these experiments, there is evidence that the observed differences in interlayer registry and population of defects between space- and Earth-grown crystals probably originate in a microgravity effect rather than a change in the kinetics of the process. Full details will appear elsewhere13.