Chirality-dependent growth life on solid catalysts
The mechanisms underlying the chirality selection of SWCNTs are of great interest to the scientific community with a recent proliferation of theoretical and numerical contributions describing varying aspects of chirality selection and growth. Newly developed solid catalysts particularly require further investigations to determine the dominant mechanism of selective growth, given their outstanding performance. Recent useful contributions include (i) interface energy–driven nucleation thermodynamics (9–11), (ii) interface configurational entropy–driven nucleation thermodynamics (27), (iii) cap nucleation kinetics with Zeldovich factor (28), (iv) catalyst epitaxy with symmetry matching (5, 6, 8), (v) CNT growth kinetics (10), and (vi) etching agent–dependent growth kinetics (29), among others. These contributions provide a useful set of potential theoretical mechanisms and physical parameters that must be measured and tested to refine the varied contributions into a cohesive and useful theory that can guide the development of models, experiments, and production of chirality-controlled CNTs.
The production and precise size control of solid catalysts for CNT growth provide a means to examine and refine theories regarding the abundance of CNT chiralities by limiting multivariant factors such as variable chemistry, interface energy, and catalyst size. By limiting the degree of freedom, investigations can also focus on variations in chirality led by growth environment and time differences.
Our present investigations allow us to examine several existing mechanisms. It is apparent that the ability for a catalyst NP to remain solid and to maintain its shape/diameter during growth is the key feature for chirality enrichment around the (2n,n) line. This trend has been observed for three different catalyst compositions having different crystal symmetries (point groups). Catalyst symmetry (iv) seems to have no discernible influence on the CNT chirality abundance. Previous reports (9–11, 27) argue that tight matching between tubes and catalysts would constrain the growth rate of tubes and may limit instead of enriching the abundance.
The enriched chiralities near (2n,n) support the CNT growth kinetics mechanism (v), but with an even stronger trend when comparing fig. S9E with Fig. 5G. Comparing the ED results with those from Raman mapping, the appearance of near-zigzag tubes shows general agreement with interface energy–driven nucleation thermodynamics (i) despite lacking precise inputs for critical values, such as catalyst interface energy (fig. S9D). The slightly broader chirality distribution from ED compared with Raman mapping follows the recently reported A|Z segregation on the interface of solid catalyst tubes, which broadens the CNT nucleation probability distribution. By comparing the results from three different catalysts, the subtle variations can likely be attributed to differences in tube-catalyst interface energy and the consequently different nucleation (i) and growth of CNTs (v).
To explain the additional enrichment of near-(2n,n) chiralities compared with the prediction based on growth kinetics mechanism, we extend the theory (9–11) with chirality-dependent growth time (Fig. 5I; additional details are in section S4.). In addition to being preferred by kinetics, these near-(2n,n) chiralities also prolong their growth time before being halted from catalyst poisoning by an abundance of carbon (catalysts frequently found to be encapsulated by a layer of carbon; Fig. 3, E and F, and fig. S8 indicated by arrows). Thus, the length of a CNT would not only be constrained by growth speed but also be determined by the chirality-dependent growth time. In addition, the size of catalyst determines the region of possible chiralities, thus affecting final chirality abundance.
Although the inclusion of catalyst poisoning within existing theoretical framework improves theoretical prediction of results concurrent with our experiments, the current theories are still not robust in explaining all observed phenomena. More experimental studies—particularly the in situ TEM observation of single tube growth based on size-selected solid catalysts—are needed to further test and refine existing models, e.g., time-dependent growth. Our results show a need for inclusion of catalyst deactivation as a parameter that influences CNT chiral abundance. Also, we demonstrate that currently unmeasured or simulated constants, such as interface energy, must be examined to advance our understanding of the chirality selection and growth process.
To date, methods to produce solid catalysts are batch processes that rely on either specialized, expensive precursors or specific substrate compositions. Gas-phase synthesis of solid or liquid catalysts allows for continuous production of physically selected NPs that can be directly deposited onto any substrate material. The resultant high throughput also increases the application potential for high-throughput screening, automated catalyst design, etc.
The solid state of the reported catalyst during reduction and growth steps suppresses Ostwald ripening and contributes to a reliable and predictable catalyst size as selected. In contrast, because of Ostwald ripening and flexible morphology of traditional liquid catalysts, chiralities grown are inherently diverse.
Beyond the three metals studied here, many others may also be viable candidates for solid catalysts, including other refractory metals such as V, Nb, Ru, Rh, and Os. The extensive list of compositions can offer different interface energies at the catalyst-tube interface, providing the enrichment of other chiralities in the final products with our compatible experimental setup.
Extension of these techniques can achieve further reduced catalyst size (<1 nm) and narrower size distributions, which would further prohibit perpendicular growth of SWCNTs. Combined with the difference in chirality-dependent growth times, semiconducting chiralities such as (8,4), (9,4), (10,3), and (11,3) are predicted to dominate the final products (fig. S9i) to achieve high-purity semiconducting SWCNTs product with a simple direct growth method.
