Superhard W_0.5Ta_0.5B nanowires prepared at ambient pressure Free

The search for new superhard materials has predominately focused on intrinsically hard materials. Nature has only provided the covalent crystal diamond, and since then, the superhard materials field has blossomed with the development of new compounds (such as cubic boron nitride) that mimic diamond's crystal structure.¹ More recently, by adding short covalent bonds into electron-dense metals, a new class of superhard metals, which can be synthesized at ambient pressure, has been developed including ReB₂,² WB₄,³ and CrB₄.⁴ Crystal structure engineering has led to a deeper understanding of the bonding motifs responsible for this new class of superhard metals with exciting potential for practical applications. Recently, tungsten monoboride was induced to be superhard through the amelioration of a slip plane. By selectively substituting tungsten atoms with tantalum, the (020) plane is solid-solution hardened, thus removing the weakest link and promoting the W_1-xTa_xB (x = 0.0–0.5) system into the superhard regime.⁵ 

To date, however, little emphasis has been placed on examining extrinsic hardening effects in this new class of materials. As superhard materials are at the limit of material strength, it is completely unknown how conventional extrinsic mechanisms such as the Hall-Petch effect, morphological control, and nanostructuring will affect the high hardness possessed by these materials. The preliminary works on diamond,⁶ cubic boron nitride,⁷ and tungsten tetraboride⁸ have shown that nanostructuring can indeed increase hardness, but morphological control of crystallite and grain shape are yet to be demonstrated.

Here, we demonstrate that W_0.5Ta_0.5B can be synthesized in an aluminum flux at ambient pressures to yield nanowires. By controlling the flux ratio, the aspect ratio of the nanowires can be controlled. This material in bulk has been previously shown to be superhard (H_v = 42.3 ± 2.6 GPa at a load of 0.49 N) and ultraincompressible (K_o = 337 ± 3.0 GPa), and to date, this is the only reported synthesis of a superhard nanostructure under ambient pressure.

Tungsten (Strem, 99.95%), tantalum (Roc/Ric, 99.9%), and boron (Materion, 99%) powders were stoichiometrically ground in a ratio of W_0.5Ta_0.5B in an agate mortar and pestle. The mixed elemental powders were then loaded into an alumina boat with aluminum shavings (Strem Chemicals 99+%) and heated in a tube furnace under flowing argon. The heating profile was as follows: ramp up to 1050 °C at 1.71 °C/min, dwell for 12 h, cool down to 700 °C at 2.92 °C/min, and cool down to 25 °C at 11.25 °C/min. The molar ratio of aluminum to W_0.5Ta_0.5B varied from (250:1) to (50:1) to (10:1). The aluminum was then etched away with 6.0 M NaOH. The resulting powders were characterized by powder X-ray diffraction (PXRD), scanning electron microscopy (SEM), and transmission electron microscopy (TEM). PXRD was carried out using a Bruker D8 Discover Powder X-ray Diffractometer (Bruker Corporation, Germany) with Cu_Kα X-ray radiation (λ = 1.5418 Å). The collected patterns were cross-referenced against the patterns found in the database of the Joint Committee on Powder Diffraction Standards (JCPDS) to identify the phases present. The SEM images were acquired using a JEOL JSM-67 Field Emission Scanning Electron Microscope, and the TEM images were acquired using an FEI TF20.

To prepare highly crystalline nanowires with few defects, we modified the experimental conditions used for single crystal growth. Aluminum was chosen as it is the solvent of choice for borides.⁹ The heating profile closely mirrors the method used in the synthesis of single crystal ReB₂.¹⁰ This method had previously been used to synthesize highly crystalline metal borides, and thus, the ramp rates were used with minor modifications. According to nucleation theory, there are two factors that dominate final crystallite size: cooling rate and nucleation site density. As the heating rates were preserved to maximize crystallinity, the chief mechanism for controlling crystallite size becomes flux to monoboride loading ratio. A low flux to monoboride ratio should lead to a higher density of nucleation sites as upon cooling the monoboride will quickly precipitate en masse out of the molten flux due to saturation. This high density of nucleation sites will result in smaller nanocrystals. On the other hand, a high flux to monoboride ratio should lead to a lower density of nucleation sites and larger crystallites. Thus, the following loading ratios of Al to W_0.5Ta_0.5B were tested: 250:1, 50:1 and 10:1.

As can be seen from the PXRD (Figure 1), the resulting black powders are highly crystalline and index nicely to the anticipated W_0.5Ta_0.5B pattern, with some minor impurities peaks present from the trapped oxidized flux (Al₂O₃). Noticeably, the PXRD peaks broaden as the flux to monoboride ratio drops, confirming that by varying this ratio, the resulting crystallite size can be controlled. Indeed, the 10:1 sample shows far more broadening than the 250:1 or the 50:1 samples. According to the Scherrer equation analysis of the (021) peak, the crystallite size of the nanocrystals prepared from the 10:1 flux is approximately 37.8 nm, confirming that the prepared samples are indeed nanostructured. Analysis of samples prepared from the flux ratios of 250:1 and 50:1 showed negligible broadening.

Not only can the crystallite size of the nanowires be controlled by varying the flux ratio, but the morphology of the superhard nanocrystallites can also be controlled by altering the flux to monoboride loading ratio, as observed through SEM analysis. At a high flux to monoboride ratio of 250:1, the low nucleation density results in large crystallites (Figure 2(a)). As the loading ratio is lowered to 50:1, the resulting product grows into nanowires (Figure 2(b)), with the width of the nanowires as low as 500 nm. Finally, as the loading ratio is further lowered to 10:1, the resulting product grows into nanorods (Figure 2(c)), with the width of the nanorods roughly 100 nm. This morphology results from the low loading ratio and is consistent with classical nucleation theory and nucleation site density.

To elucidate the growth direction of these nanowires, we performed high resolution TEM. Unfortunately, W_0.5Ta_0.5B (14.97 g/cm³) is even denser than lead (11.34 g/cm³), and we were only able to obtain a few selected area electron diffraction images on the thinnest nanowires. To account for the low transmission of electrons, we performed convergent beam electron diffraction (CBED). The CBED patterns allowed us to gauge the orientation of the nanowires similar in thickness to the ones observed in SEM. From the TEM images (Figure 3), we were able to determine that the nanowires grow along the (002) direction, which is parallel to the boron-boron chains (Figure 4). It is fortuitous that these nanowires grow along the boron-boron chains, as the directionality of the covalent bonds in turn prevents shear.¹ 

It is interesting to speculate on the growth mechanism of these superhard nanowires. W_0.5Ta_0.5B crystallizes in the same crystal structure as high temperature orthorhombic WB. High temperature tungsten monoboride can be best described as tungsten metal bilayers separated by parallel boron chains. As this is a highly anisotropic structure, one would expect to find a defined crystal habit. One can draw an analogy to the crystal habits found in silicates. Sheet silicates (such as mica) possess a crystal structure comprising [Si_2nO_5n]^2n- layers, while silicates that grow as fibers (such as asbestos) possess a crystal structure comprising linear chains of [Si_nO_3n]^2n-. This is a result of periodic bond chain theory, where growth is favored along directions where bonding is the strongest. With tungsten-tantalum monoborides, the crystal habit will prefer an elongated fibrous structure along the strong covalent boron-boron chains (Figure 4).

Indeed, flux growth by Okada et al.¹¹ of low temperature tetragonal tungsten monoboride, where the boron chains alternate layers in a perpendicular fashion, yielded a crystal habit of flattened squares that confirms that the growth of the boron chains is crucial to the resulting crystal shape. Interestingly, there seems to be some self-assembly of the nanowires (Figure 5) found in the 50:1 sample. We note that several of the nanowires begin to fuse to form 2-D plates that are indicative of hierarchical assembly. Indeed, the assembly of these nanowires resembles that of polymer self-assembly.¹² From the SEM images, there is clear evidence for 1-D nanowires, oriented 1-D nanowires, and 2-D sheets. We anticipate that if the growth were allowed to continue, we would expect to see morphologies similar to that found in the 250:1 flux to monoboride sample.

As crystallite size is controlled through the flux to monoboride ratio, the hardness of consolidated compacts should only improve because the high density of grain boundaries prevents the propagation of dislocations, and this is known as the Hall-Petch effect. Furthermore, as these superhard nanowires grow along the boron chain, the nanowires should gain some pliability from the polymeric boron. It is anticipated that consolidated compacts should possess an additional degree of ductility and achieve increased toughness.

Superhard metal nanowires have been grown via flux growth in bulk. Careful control of the flux to monoboride ratio yields highly crystalline W_0.5Ta_0.5B nanostructures. The aspect ratio and size of these nanowires can be controlled by modifying the flux to metal boride ratio. This falls in line with classical nucleation theory, where more concentrated solutions lead to more nucleation sites (and thus, smaller nanowires approaching nanorods). From the CBED pattern, the nanowires are found to grow along the c-axis (along the boron chains), as expected from periodic bond chain theory. Interestingly, these nanowires exhibit hierarchical assembly, where some nanowires are observed to fuse into 2-D plates. These superhard nanowires may find exciting applications in composites where both electrical conductivity and mechanical robustness are required.

This research was financially supported by the National Science Foundation Grant No. DMR-1506860 (R.B.K.), a VCU startup Grant No. 137422 (R.M.), and the NSF IGERT: Materials Creation Training Program (MCTP)–DGE-0654431 Fellowship (M.T.Y. and D.J.K.). The authors would like to thank Lingxuan Pang for the helpful TEM discussions.