Tungsten–niobium–sulfur composite nanotubes

Abstract

Novel W–Nb–S composite inorganic nanotubes have been generated by replacing W in the S–W–S sandwich layers of WS₂ by Nb. The tubes are uniform, with the Nb concentration 10 atom%. Raman spectral features of the tubes are described.

---

Inorganic fullerene (IF) nanotubes, e.g. MX₂ (M = W, Mo; X = S, Se), first reported by Tenne et al.,¹ have engendered considerable scientific interest because they are good semiconductors and display excellent wear-resistant properties. To date, various approaches to the synthesis of these materials have been developed, such as chemical transport,² gas–solid reactions³ and in situ heating,⁴ leading to pure MX₂ nanoparticles, short and long tubes, bundles, and even microtubes. The structural features, lubricating properties and electronic behaviour of the IF nanotubes have been widely investigated.^5–8 For example, a single WS₂ nanotube has been used successfully as a microscope probe.^3,9 Recently, Levy and coworkers showed that small amounts of noble metals (Au and Ag) could be intercalated between WS₂ tube walls.^10,11 Akin to their bulk crystals, IF nanotube band gaps are dependent upon tube diameter.¹² It is therefore interesting to prepare doped nanotubes and to assess their electronic and mechanical performance. Here, we describe the generation of a novel W_xNb_yS₂ IF nanotube containing ca. 10 atom% Nb.

We prepared needle-like WO_x (mainly W₁₈O₄₉) nanotubes by the previously described procedure.¹³ The needles were then covered with a suspension of Nb₂O₅ (Aldrich, 99.9% pure, particle size < 2 μm) in acetone, and the acetone was allowed to evaporate leaving Nb₂O₅ uniformly distributed over the needles. The sample was then reheated to ca. 1100 °C under 200 Torr H₂S. Using a scalpel, the grey needle-like deposit was removed from the quartz surface, subjected to high resolution transmission electron microscopy (HRTEM, JEOL-4000) and to energy dispersive X-ray (EDX, element > B) and Raman spectroscopic (Renishaw system 1000, λ₀ = 514.5 and 632.8 nm) analysis.

Previously, we had refined the oxide–sulfide conversion method⁷ and successfully created very long pure WS₂ nanotubes by forming WO_x nanorods prior to oxide–sulfide conversion.⁴ We have now found that the constituent atoms of Nb₂O₅-coated WO_x nanorods easily diffuse at high temperature (ca. 1100 °C), particularly through materials possessing substantial vacancies. Thus Nb probably occupies some of the W positions within W₁₈O₄₉, leading to W_xNb_yO_z. This replacement process is reminiscent of the behaviour of the bulk oxides, but in this case the nanorod templates and the suboxide features were retained. The presence of gaseous H₂S in the chamber facilitated oxide-to-sulfide conversion in an identical manner to that found for the synthesis of pure WS₂ or MoS₂ nanotubes,^1,3,4 resulting in homogeneous W–Nb–S composite nanotubes. These tubes are straight, ca. 5–10 μm long and ca. 20–100 nm in diameter. HRTEM images of the tubes are not different from those of WS₂ nanotubes. The majority have hollow cores, less than 30% of them being fully or partly filled with elongated WO_x crystals. Concentric nanotubes (layer separation ca. 0.62 nm) are observed. The outer layers are smooth and appear to be free of lattice defects. All the tubes are closed, ca. 20% by flat caps, forming a ca. 90-degree connection with the cylinders. Dislocations at the cap–cylinder junctions are apparent [Fig. 1(a), arrowed]. EDX revealed the presence of Nb [Fig. 1(b)] within the nanotubes. Measurements on ten isolated hollow nanotubes revealed a ca. 9∶1 W to Nb atomic ratio, the metal∶sulfur ratio being ca. 1∶2. The overall stoichiometry of the nanotubes is W_xNb_yS₂ (x = 0.9, y = 0.1, 5% uncertainty). The flat tip feature (ca. 90-degree angle with the body) of the nanorods is determined by their intrinsic crystalline nature.¹³ Hence, the flat nanotube tips probably form as a result of oxide–sulfide template conversion. This conjecture is supported by supplementary experiments, which showed that NbS₂ nanotubes could not be prepared directly from Nb₂O₅ powder alone. These results confirm that the suboxide plays a key role in the IF nanotube formation, and the process can be written in terms of eqns. (1) and (2):

WO_{x (nanorod)} + Nb₂O_{5 (particle)} → W_0.9Nb_0.1O_{x (nanorod)}  (1) (1)

W_0.9Nb_0.1O_{x (nanorod)} + H₂S _(gas)→ W_0.9Nb_0.1S_{2 (nanotube)} + H₂O _(gas)  (2) (2)

In order to ascertain whether or not Nb was present as an intercalate within the WS₂ tube walls, or as a dopant in the WS₂ structure, we undertook additional EDX analyses. The results showed that the W∶Nb∶S ratio remained nearly constant, independent of nanotube diameter and position. This behaviour is consistent with the EDX mapping, and implies that the nanotubes are fairly uniform. When intercalated Au and Ag are present, the nanotubes are heterogeneous, irrespective of their diameter and length.^10,11 Heterogeneous intercalation is likely to lead to internal stresses owing to the different atomic diameters of W and Nb, resulting in significant defects. In fact, our HRTEM observations show that the distance between layer fringes is ca. 0.62 nm [Fig. 1(a)], identical with that in pure WS₂; only a few defects are present. Furthermore, expanded layer separations, i.e. >0.62 nm, would be expected if Nb were to be intercalated between the WS₂ layers. However, very few crystals were found with expanded layer separations. Thus our nanotubes are Nb-doped and not intercalated.

Fig. 1 (a) HRTEM image showing a ca. 90-degree flat cap, layer separation = 0.62 nm. Dislocations at the cap-cylinder junctions are apparent (arrowed); (b) EDX of an isolated hollow tube, revealing the presence of W and Nb. S signals are omitted from this profile.

Additional experiments have revealed that the W∶Nb ratio can be altered by applying various amounts of Nb₂O₅ coating to the WO_x nanorods, prior to passage of H₂S during heating. The Nb concentration (ca. 10% atom%) can be doubled in the nanotubes; however, our results showed that Nb-dominated W–Nb–S tubes are difficult to construct.

The Raman spectrum of W_0.9Nb_0.1S₂ nanotubes taken with 632.8 nm excitation is identical to that of WS₂ nanotubes. The 514.5 nm spectrum of the Nb-doped tubes, however, does show some differences from that of WS₂ (Fig. 2), notably the appearance of a medium intensity band at 313 cm⁻¹, a shoulder on the low wavenumber side of the e_2g¹ band at 356 cm⁻¹ and an increase in the relative intensity of the 192 cm⁻¹ band. The intensity of the 192 cm⁻¹ band is ca. half that of the disorder-induced zone-edge phonon [LA(M)] band at 173 cm⁻¹ for WS₂, whereas its intensity for the Nb-doped tubes is ca. 5 times that of the 173 cm⁻¹ band. The 192 cm⁻¹ band has been observed in the Raman spectra of 2H-bulk and nanotubes of WS₂, but has not been assigned.¹⁴ Its increase in relative intensity on going from pure WS₂ nanotubes to Nb-doped WS₂ nanotubes implies that it is caused by scattering from a disorder-induced zone-edge phonon. The band at 313 cm⁻¹ has not previously been seen in the Raman spectra of any form of WS₂, and is also probably due to the increased disorder in the lattice owing to the presence of Nb. The shoulder at 351 cm⁻¹ is the first overtone of the LA(M) band, with its apparent intensity enhanced by its proximity to the strong band at 365 cm⁻¹. The Raman bands and their assignments are summarised in Table 1 of ESI.†

Fig. 2 Raman spectra (λ₀ = 514.5 nm) of W_0.9Nb_0.1S₂ (upper) and WS₂ (lower) nanotubes.

The formation of W_0.9Nb_0.1S₂ nanotubes can be accounted for by considering the layered structure of the W_0.9Nb_0.1S₂ crystals. According to our EDX analysis, only ca. 10% of the W atoms are replaced by Nb. Therefore, we suggest that the Nb-doped material remains as the 2H_b-type WS₂ structure, rather than becoming the 2H_a-type NbS₂ structure. We propose a model for layered W_0.9Nb_0.1S₂ (Fig. 3). In this structure, the doped Nb atoms (medium circles) should not cause severe lattice defects and can easily be stabilised by neighbouring atoms. This suggestion is supported by our Raman spectra; the spectra of the Nb-doped and pure WS₂ tubes are identical using 632.8 nm excitation, whereas new bands attributable to increased lattice disorder are seen using 514.5 nm excitation. The crystal lattice spacing, particularly c/2, remains almost unchanged when compared with that of pure WS₂ layers. If a nanotube is to be formed by rolling up such composite layers, one would not expect a significant c/2 change. This conjecture is also supported by the facts that the composition of our nanotubes is uniform and that we did not observe severe defects on the tube walls.

Fig. 3 Proposed model for the layered W_0.9Nb_0.1S₂ structure.

Acknowledgements

We thank the Royal Society, the JFCC, the EPSRC and the ULIRS, for financial support. We are grateful to J. Thorpe and D. Randall (Sussex) for assistance with TEM and SEM facilities. We thank M. Mayne and I. Maurin for helpful discussions.

Notes and references

1. R. Tenne, L. Margulis, M. Genut and G. Hodes, Nature, 1992, 360, 444.
2. M. Remskar, Z. Skraba, M. Regula, C. Ballif, R. Sanjines and F. Levy, Adv. Mater., 1998, 10, 246.
3. M. Homyonfer, B. Alperson, Y. Rosenberg, L. Sapir, S. R. Cohen, G. Hodes and R. Tenne, J. Am. Chem. Soc., 1997, 119, 2693.
4. Y. Q. Zhu, W. K. Hsu, N. Grobert, B. H. Chang, M. Terrones, H. Terrones, H. W. Kroto, D. R. M. Walton and B. Q. Wei, Chem. Mater., 2000, 12, 1190.
5. J. Sloan, J. H. Hutchison, R. Tenne, Y. Felfman, T. Tsirlina and M. Homyonfer, J. Solid State Chem., 1999, 144, 100.
6. M. Remskar, Z. Skraba, R. Sanjines and F. Levy, Appl. Phys. Lett., 1999, 74, 3633.
7. L. Rapoport, Y. Bilik, Y. Feldman, M. Homyonfer, S. R. Cohen and R. Tenne, Nature, 1999, 387, 791.
8. G. Seifert, H. Terrones, M. Terrones, G. Jungnickel and T. Frauenheim, Solid State Commun., 2000, 114, 245.
9. A. Rothschild, S. R. Cohen and R. Tenne, Appl. Phys. Lett., 1999, 75, 4025.
10. M. Remskar, Z. Skraba, C. Ballif, R. Sanjines and F. Levy, Surf. Rev. Lett., 1999, 6, 1283.
11. M. Remskar, Z. Skraba, P. Stadelmann and F. Levy, Adv. Mater., 2000, 12, 814.
12. G. Seifert, H. Terrones, M. Terrones, G. Jungnickel and T. Frauenheim, Phys. Rev. Lett., 2000, 85, 146.
13. Y. Q. Zhu, W. B. Hu, W. K. Hsu, M. Terrones, N. Grobert, H. Terrones, H. W. Kroto and D. R. M. Walton, Chem. Phys. Lett., 1999, 309, 327.
14. G. L. Frey, R. Tenne, M. J. Matthews, M.S. Dresselhaus and G. Dresselhaus, J. Mater. Res., 1998, 13, 2412.

---

Footnote

† Electronic supplementary information (ESI) available: HRTEM image of a typical 10-layered W_0.9Nb_0.1S₂ nanotube, layer separation = 0.62 nm and Raman bands of W_0.9Nb_0.1S₂ and WS₂ nanotubes and their assignments. See http://www.rsc.org/suppdata/cc/b0/b007074m/

---