Dual-Surface Synthesis and Characterization of Osmium Nanostructures
April 14, 2026 – Researchers from the United Kingdom (University of Nottingham, University of Oxford and Diamond Light Source) and Germany (Ulm University) have demonstrated how the geometric duality of single-walled carbon nanotubes (SWNTs) can be harnessed as a platform for the targeted synthesis of inorganic nanomaterials. Using local-probe electron microscopy in combination with ensemble-level spectroscopy – including the SALVE Cc/Cs-corrected TEM operated at 60 kV[27] – the team shows that the convex outer surface of the nanotubes favours accessible, bulk-like nanostructures, while the concave inner channel acts as a confined template for quasi-one-dimensional materials. The findings open new routes to metastable osmium phases and underscore the role of carbon nanotubes as a dual-interface laboratory for the design of next-generation functional nanomaterials.
Single-walled carbon nanotubes (SWNTs) offer a remarkable platform for the controlled synthesis of nanomaterials, largely because their surfaces have a structural duality. The convex outer wall promotes the growth of larger, accessible nanostructures well suited for catalysis and sensing.[1] By contrast, the concave inner channel acts as a spatially confined template for quasi-one-dimensional materials.[2],[3] Such nanoscopic confinement stabilizes exotic structures — including linear chalcogen chains of sulfur, selenium and tellurium[4],[5],[6] — that cannot exist in bulk form. Organometallic complexes such as triosmium dodecacarbonyl are particularly effective precursors for growing well-defined osmium clusters within the nanotube architecture. This strategy builds on the known versatility of SWNTs as templates for striking nanostructures, among them coordination-reduced TbCl3 polyhedra[7] and magnetic nanoribbons based on cobalt phthalocyanine.[8]
SWNTs also serve as ideal platforms for in situ structural analysis, permitting the real-time observation of nanomaterial growth by transmission electron microscopy (TEM). Molecular precursors inside or on the nanotubes can be transformed by thermal or electron-beam stimuli,[9],[10] typically yielding products of reduced dimensionality.[3],[11] Recent studies have shown the power of stepwise synthesis, which makes it possible to insert elements in controlled sequences and ratios. Previous work has applied this approach to a range of metal iodides and sulfides — including MoS2, WS2 and PtS2.[9],[12],[13],[14]
In the osmium chemistry reported here, the researchers employ the inner and outer surfaces of single-walled carbon nanotubes as a dual-functional analytical scaffold. Unlike conventional studies that examine each surface in isolation, the strategy presented by the team allows a direct comparison of how spatial constraints shape material growth under identical experimental conditions. By combining local-probe electron microscopy (HRTEM, EDX) with macroscopic analysis (XPS, resonance Raman spectroscopy) (Scheme 1), the authors demonstrate that the nanoscopic environment plays a decisive role in the structural evolution of inorganic nanostructures.
Organometallic complexes such as triosmium dodecacarbonyl [Os3(CO)12] and [Os3(CO)10(NCMe)2] are well-established precursors for growing osmium clusters inside single-walled (SWNTs)[15] and multi-walled carbon nanotubes (MWNTs).[16] Osmium species have also been used to functionalize the SWNT sidewall: the cycloaddition of osmium tetroxide (OsO4) under photoirradiation, for example, tunes the electronic properties by increasing nanotube resistance.[17] Encapsulation of exohedral fullerenes such as [Os3(CO)11(η2-C60)][18] has further shown that the strong van der Waals interaction between the fullerene cage and the concave nanotube wall can efficiently pull osmium clusters into the tube interior.[19]
Based on these considerations, the team selected Os3(CO)12 as the primary precursor, since it converts cleanly into a range of osmium species without leaving residual ligand impurities.[19] As a first step, Os3(CO)12 was sublimed in vacuum onto SWNTs, ensuring a uniform distribution across both the inner and outer surfaces of the nanotubes.
To preserve the fragile metal–carbonyl complexes, the researchers performed HRTEM imaging at 100 kV with a low electron dose (5.1 × 104 e−/nm2). Under these conditions, Os3(CO)12 molecules became visible as high-contrast elliptical features (Fig. 1a) — overcoming earlier difficulties in which electron-beam irradiation rapidly ruptured the metal–ligand bonds, as previously seen in the formation of "naked" rhenium nanoparticles from rhenium carbonyl precursors.[20]
At the optimized low dose, discrete molecular species roughly 0.65 nm in diameter were observed translating and rotating inside the SWNTs. Image analysis confirmed that Os3(CO)12 fits the 1.2 nm diameter host tubes well: the molecules are large enough to maintain strong van der Waals contact with the nanotube wall, yet retain enough freedom to adopt the various orientations resolved in the HRTEM projections (Fig. 1b, inset). Although EDX analysis confirmed the presence of C, O and Os (Fig. 1b), intrinsic signals from the nanotubes and residual oxygen prevented a precise quantification.
Prolonged electron-beam exposure (5 min), however, cleaved the carbonyl ligands and transformed the molecules into metallic osmium clusters. These clusters then promoted etching of the nanotube walls by electron-beam-induced ejection (EBIE).[21] Os3(CO)12 species adsorbed on the outer surface could not be imaged directly — most likely because of their high mobility and sublimation in the TEM vacuum — but their presence was unequivocally established by their later reactivity.
Subsequent thermal treatment, however, produces a marked shift in the Os 4f 7/2 binding energy to 52.4 eV (Fig. 1c) — higher than the value expected for metallic osmium (50.8 eV) and than previously reported for free Os3(CO)12 molecules (51.6 eV).[22] This value, together with the iodine 3d 5/2 signal at 619.3 eV, confirms the formation of metal iodide species consistent with Os(I) environments.[23] The same transition is reflected in the Raman spectra (Fig. 1d), which indicate only negligible charge transfer between the precursor and the host. The G-band profile shifts toward a more Lorentzian shape, consistent with a change from metallic to semiconducting character.[24] Subtle shifts in the radial breathing modes (RBMs, 3.2 cm−1) are further characteristic of SWNT filling.[25] Although the signal of the encapsulated species overlaps substantially with the RBM of the SWNT host, its presence is clearly revealed by the second, third and fourth harmonics.[26] Intact carbonyl ligands are corroborated by IR spectroscopy (Fig. 1e), which resolves all characteristic Os3(CO)12 vibrational modes (1985–2068 cm−1). Combined with the Os 4f 7/2 binding energy of 52.4 eV, these data confirm that the precursor remains chemically intact upon adsorption.
Thermal treatment produced nanoparticles on the outside of the SWNT bundles (Fig. 2a); a size distribution analysis yielded a mean diameter of 5.1 ± 0.9 nm (range: 2–8 nm). AC-HRTEM imaging of these nanoparticles resting on the convex nanotube surface confirmed their crystalline character, revealing a characteristic d-spacing of 0.26 nm (Fig. 2b). This lattice spacing appeared reproducibly in multiple projections, including regions where overlapping nanoparticles generated a distinct hexagonal interference pattern.
Local EDX analysis of individual nanoparticles yielded an osmium-to-iodine ratio of 1:1.2, confirming the formation of osmium mono-iodide on the outer SWNT surfaces. Together with the AC-HRTEM lattice data, this EDX quantification is consistent with a NaCl-type cubic structure, in which both osmium and iodine occupy six-coordinate environments (Fig. 2c). Taken together with the resonance Raman observations above, the findings suggest that pentaiodide anions (I5−) act as the primary oxidants for the osmium precursors during the thermal activation step.
To characterize the species formed inside the nanotubes during thermal treatment, the authors used the SALVE (Sub-Angstrom Low Voltage Electron microscopy) Cc/Cs-corrected TEM operated at 60 kV.[27] The images revealed four parallel rows of atoms in close proximity within a 1.71 nm SWNT (Fig. 3a). These features correspond to two independent, two-atom-thick nanowires extending along the entire tube length and displaying autonomous rotational and translational motion (Fig. 3b). Rotation of the nanowires around the tube axis enabled a detailed structural analysis across several projections: a 45° rotation from the face-on view, for example, brings the atomic rows into overlap, producing the noticeably higher contrast in the central rows of Fig. 3a.
Further time-series imaging captured the nanowires in various orientations, one of them appearing as two parallel lines due to a tilt in the nanowire axis (Fig. 3c). The continuous rotation around the tube axis provides access to a detailed structural analysis from multiple viewpoints. Throughout these orientations, the atomic contrast remains remarkably uniform — indicating that each feature arises from the overlap of an osmium and an iodine atom in the projection.[14] The measured distance between these lines (0.28 nm) is strikingly close to the d-spacing observed in the exterior nanoparticles. Based on interatomic distances of 0.31 and 0.36 nm, the team developed a structural model representing a distorted cubic NaCl-type phase (Fig. 3d). In this low-coordinate 1D structure, each iodine atom bonds to only three osmium atoms (Fig. 3e). AC-HRTEM simulations using these models match the experimental projections across the entire series (Fig. 3a–e).
Finally, EDX analysis of filled SWNT bundles (Fig. 3f) confirmed an Os:I ratio of approximately 1:1, signalling the formation of osmium mono-iodide (OsI) inside the nanotubes, analogous to the exterior nanoparticles.[13] The gap between each OsI nanowire and the concave nanotube surface was 0.31 nm — consistent with a typical van der Waals distance in graphitic materials.[28] Although bulk OsI is reported to be amorphous,[29] the findings underscore the pivotal role of SWNTs in templating and stabilizing low-coordination compounds. Crucially, the dual-surface approach shows that spatial confinement dictates the coordination environment: osmium centres in the 1D nanowires are three-coordinate, whereas those on the unconfined convex surface are six-coordinate. The concave SWNT wall therefore stabilizes the high surface area and low-coordinate centres that would otherwise be unstable in the bulk phase.
TEM imaging of the nanoparticles on the SWNT surfaces revealed both compositionally uniform particles and core–shell architectures, the latter carrying between one and three distinct shells (Fig. 4a–c, inset). The mean nanoparticle diameter was 5.3 ± 1.4 nm, commensurate with the size of the osmium iodide precursors from the previous step. Local EDX analysis on individual core–shell structures identified osmium and sulfur as the constituent elements, yielding an Os:S atomic ratio of 1:1.6 (Fig. 4c).
AC-HRTEM imaging of the core–shell nanoparticles identified a d-spacing of 0.23 nm in the core, matching the (211) plane of cubic pyrite-type OsS2 (Fig. 5a).[30] The atomic architecture of the shell, clearly resolved in digitally magnified images (Fig. 5b), showed a distinct structural motif. In these edge-on projections, the shell resembles hexagonal transition-metal dichalcogenide (h-TMD) nanoribbons, with a central row of high-contrast metal atoms flanked by lower-contrast sulfur atoms. Within the central atomic row, the interatomic distance of 0.27 nm points to the formation of an h-OsS2 phase (Fig. 5c) isomorphic to h-WS2.
In regions where the cubic core was only partially encapsulated, the hexagonal in-plane structure of the shell became apparent (Fig. 5d, e). The images reveal alternating contrast intensities and a lattice spacing of 0.24 nm between neighbouring high-contrast features. HRTEM simulations of a hexagonal OsS2 monolayer (Fig. 5f) agree closely with the experimental data (Fig. 5g), confirming that the nanoparticles are a biphase composite — denoted here as c-OsS2@h-OsS2.
Hexagonal disulfide phases are common for molybdenum and tungsten, but have not previously been reported for osmium. Their formation here is probably driven by the stabilization of dangling bonds on the cubic OsS2 core, which the convex SWNT surface — curving away from the particle — cannot saturate. The arrangement is reminiscent of core–shell carbon architectures such as nanodiamond@graphene, where a hybridization shift in the shell stabilizes the reactive core.[31],[32] For osmium disulfide, however, the transition also involves a shift in oxidation states: while c-OsS2 (pyrite-type) consists of Os2+ and S22−, the isostructural h-OsS2 phase is expected to comprise Os4+ and S2−.
To resolve the structure of OsS2 encapsulated within the SWNTs, the authors used high-frame-rate HRTEM (10 fps), which proved essential for capturing snapshots of the nanowires in on-axis orientations despite their high translational and rotational mobility (Fig. 6a).[33] Three-atom-thick nanowires with a diameter of 0.6 nm extend along the full length of the nanotubes. The nanowires also show helical twists, most likely to maximize van der Waals interactions with the concave host surface and relieve internal strain; simulations of these twisted segments remain consistent with the cubic model across multiple projections (Fig. 6b, c).
A detailed image analysis along the nanowire axis identified alternating contrast intensities (Fig. 6e, f). Given the large atomic-number difference between sulfur and osmium, individual S atoms are hard to resolve; the observed features therefore correspond to columns containing varying numbers of osmium atoms.[14] This assignment is consistent with a face-on projection of the pyrite-type cubic unit cell, in which corners and the centre of the square face host two overlapping Os atoms, while the edge midpoints contain only one (Fig. 6g).
The interstitial space between the SWNT wall and the nanowire face appears to accommodate additional sulfur or disulfide species (Fig. 6d, black arrows), which explains why the edges of the cubic nanowire are effectively stabilized by the surrounding SWNT wall — removing the need for the hexagonal phase transition seen in the exterior nanoparticles. This sulfur-rich environment is confirmed by EDX analysis of filled SWNT bundles, which shows the complete displacement of iodine and an approximate Os:S atomic ratio of 1:3 (Fig. 6h).
The discovery of novel hexagonal and low-coordinate osmium phases highlights the potential of carbon nanotubes as a "nano-laboratory" for expanding the library of nanomaterials. The dual-surface approach opens the way to rational design of hybrid structures whose properties can be tailored through precise spatial and strain engineering. Beyond osmium, transferring the strategy to other transition metals could unlock a broad range of metastable phases inaccessible to traditional bulk synthesis. Future work is expected to focus on the electronic and catalytic properties of these structures, in particular exploiting the semiconducting character of the h-OsS2 shell and the under-coordinated centres of the 1D nanowires for site-specific catalysis and quantum devices.
Resource:
Norman, L. T., Cull, W. J., Stoppiello, C. T. et al. (2026).
Inorganic hybrid materials integrating carbon nanotubes with osmium compounds: formation mechanisms, structure and reactivity.
Dalton Trans. 2026; Advance Article.
https://doi.org/10.1039/D6DT00379F
-
Ju, S., Lee, J. M., Jung, Y., Lee, E., Lee, W., & Kim, S. J. (2010). Highly sensitive hydrogen gas sensors using single-walled carbon nanotubes grafted with Pd nanoparticles. Sensors and Actuators B: Chemical, 146(1), 122–128.
-
Khlobystov, A. N. (2011). Carbon nanotubes: From nano test tube to nano-reactor. ACS Nano, 5(12), 9306–9312. https://doi.org/10.1021/nn204596p
-
Miners, S. A., Rance, G. A., & Khlobystov, A. N. (2016). Chemical reactions confined within carbon nanotubes. Chemical Society Reviews, 45(17), 4727–4746. https://doi.org/10.1039/C6CS00090H
-
Tonkikh, A. A., Rybkovskiy, D. V., Enyashin, A. N., et al. (2025). The probing of local charge with sulfur encapsulated in single-walled carbon nanotubes. Carbon, 232, 119806. https://doi.org/10.1016/j.carbon.2024.119806
-
Cull, W. J., Ramasse, Q. M., Biskupek, J., et al. (2025). Flexible selenium nanowires with tuneable electronic bandgaps. Advanced Materials, 37(32), 2501821. https://doi.org/10.1002/adma.202501821
-
Medeiros, P. V., Marks, S., Wynn, J. M., et al. (2017). Single-atom scale structural selectivity in Te nanowires encapsulated inside ultranarrow, single-walled carbon nanotubes. ACS Nano, 11(6), 6178–6185. https://doi.org/10.1021/acsnano.7b02225
-
Xu, C., Sloan, J., Brown, G., et al. (2000). 1D lanthanide halide crystals inserted into single-walled carbon nanotubes. Chemical Communications, (24), 2427–2428. https://doi.org/10.1039/b006817i
-
Tonkikh, A. A., Rybkovskiy, D. V., & Obraztsova, E. D. (2023). Charge-induced structure variations of 1D-iodine inside thin SWCNTs. The Journal of Physical Chemistry C, 127(6), 3005–3012. https://doi.org/10.1021/acs.jpcc.2c06920
-
Wang, Z., Li, H., Liu, Z., et al. (2010). Mixed low-dimensional nanomaterial: 2D ultranarrow MoS2 inorganic nanoribbons encapsulated in quasi-1D carbon nanotubes. Journal of the American Chemical Society, 132(39), 13840–13847. https://doi.org/10.1021/ja1058026
-
Chamberlain, T. W., Biskupek, J., Rance, G. A., et al. (2012). Size, structure, and helical twist of graphene nanoribbons controlled by confinement in carbon nanotubes. ACS Nano, 6(5), 3943–3953. https://doi.org/10.1021/nn300137j
-
Miners, S. A., Rance, G. A., & Khlobystov, A. N. (2013). Regioselective control of aromatic halogenation reactions in carbon nanotube nanoreactors. Chemical Communications, 49(49), 5586–5588. https://doi.org/10.1039/c3cc42414f
-
Botos, A., Biskupek, J., Chamberlain, T. W., et al. (2016). Carbon nanotubes as electrically active nanoreactors for multi-step inorganic synthesis: sequential transformations of molecules to nanoclusters and nanoclusters to nanoribbons. Journal of the American Chemical Society, 138(26), 8175–8183. https://doi.org/10.1021/jacs.6b03633
-
Wang, Z., Zhao, K., Li, H., et al. (2011). Ultra-narrow WS2 nanoribbons encapsulated in carbon nanotubes. Journal of Materials Chemistry, 21(1), 171–180. https://doi.org/10.1039/c0jm02821e
-
Stoppiello, C. T., Biskupek, J., Li, Z. Y., et al. (2017). A one-pot-one-reactant synthesis of platinum compounds at the nanoscale. Nanoscale, 9(38), 14385–14394. https://doi.org/10.1039/c7nr05976k
-
Cao, K., Zoberbier, T., Biskupek, J., et al. (2018). Comparison of atomic scale dynamics for the middle and late transition metal nanocatalysts. Nature Communications, 9(1), 3382. https://doi.org/10.1038/s41467-018-05831-z
-
Yung, K. F., & Wong, W. T. (2007). Synthesis and catalytic studies of uniform Os & Os–Pd nanoparticles supported on MWNTs. Journal of Cluster Science, 18(1), 51–65. https://doi.org/10.1007/s10876-006-0079-4
-
Cui, J., Burghard, M., & Kern, K. (2003). Reversible sidewall osmylation of individual carbon nanotubes. Nano Letters, 3(5), 613–615. https://doi.org/10.1021/nl034135p
-
Chamberlain, T. W., Champness, N. R., Schröder, M., & Khlobystov, A. N. (2011). A piggyback ride for transition metals: Encapsulation of exohedral metallofullerenes in carbon nanotubes. Chemistry – A European Journal, 17(2), 668–674.
-
Zoberbier, T., Chamberlain, T. W., Biskupek, J., et al. (2012). Interactions and reactions of transition metal clusters with the interior of single-walled carbon nanotubes imaged at the atomic scale. Journal of the American Chemical Society, 134(6), 3073–3079. https://doi.org/10.1021/ja208746z
-
Chamberlain, T. W., Zoberbier, T., Biskupek, J., et al. (2012). Formation of uncapped nanometre-sized metal particles by decomposition of metal carbonyls in carbon nanotubes. Chemical Science, 3(6), 1919–1924. https://doi.org/10.1039/c2sc01026g
-
Zoberbier, T., Chamberlain, T. W., Biskupek, J., et al. (2016). Investigation of the interactions and bonding between carbon and group VIII metals at the atomic scale. Small, 12(12), 1649–1657. https://doi.org/10.1002/smll.201502210
-
Zanoni, R., Carinci, V., Abu-Samn, R. H., Psaro, R., & Dossi, C. (1985). Osmium carbonyl clusters: XPS characterization of some catalyst precursors. Journal of Molecular Structure, 131(3–4), 363–369. https://doi.org/10.1016/0022-2860(85)87038-1
-
Moulder, J. F., Stickle, W. F., Sobol, P. E., & Bomben, K. D. (1992). Handbook of X-ray photoelectron spectroscopy: A reference book of standard spectra for identification and interpretation of XPS data. Physical Electronics Division of Perkin-Elmer Corporation, p. 261.
-
Kharlamova, M. V. (2016). Electronic properties of single-walled carbon nanotubes filled with manganese halogenides. Applied Physics A, 122(9), 791. https://doi.org/10.1007/s00339-016-0335-x
-
Kharlamova, M. V., Kramberger, C., & Pichler, T. (2016). Semiconducting response in single-walled carbon nanotubes filled with cadmium chloride. Physica Status Solidi (B), 253(12), 2433–2439.
-
Grigorian, L., Williams, K. A., Fang, S., et al. (1998). Reversible intercalation of charged iodine chains into carbon nanotube ropes. Physical Review Letters, 80(25), 5560. https://doi.org/10.1103/PhysRevLett.80.5560
-
Linck, M., Hartel, P., Uhlemann, S., et al. (2016). Chromatic aberration correction for atomic resolution TEM imaging from 20 to 80 kV. Physical Review Letters, 117(7), 076101. https://doi.org/10.1103/PhysRevLett.117.076101
-
Khlobystov, A. N., Britz, D. A., & Briggs, G. A. D. (2005). Molecules in carbon nanotubes. Accounts of Chemical Research, 38(12), 901–909. https://doi.org/10.1021/ar040287v
-
Fergusson, J. E., Robinson, B. H., & Roper, W. R. (1962). Iodides of osmium and rhenium. Journal of the Chemical Society, 2113–2115. https://doi.org/10.1039/jr9620002113
-
Stingl, T., Müller, B., & Lutz, H. D. (1992). Crystal structure refinement of osmium(II) disulfide, OsS2. Zeitschrift für Kristallographie – Crystalline Materials, 202(1–4), 161–162. https://doi.org/10.1524/zkri.1992.202.1-2.161
-
Lan, G., Qiu, Y., Fan, J., et al. (2019). Defective graphene@diamond hybrid nanocarbon material as an effective and stable metal-free catalyst for acetylene hydrochlorination. Chemical Communications, 55(10), 1430–1433. https://doi.org/10.1039/c8cc09361j
-
Liang, B., Zhang, L., Wang, G., et al. (2019). Nanodiamond core/onion-like carbon shell materials with excellent visible light photocatalytic activity. Materials Research Express, 6(4), 045609. https://doi.org/10.1088/2053-1591/aafa23
-
Cull, W. J., Skowron, S. T., Hayter, C. T., et al. (2023). Subnanometer-wide indium selenide nanoribbons. ACS Nano, 17(6), 6062–6072. https://doi.org/10.1021/acsnano.3c00670






