*2.1. Scanning and Transmission Electron Microscopy Analysis*

Typical scanning and transmission electron microscopy images of a pristine (untreated) multiwall WS2 ("mother") nanotube are presented in Figure 2a,b, respectively. The majority of the predecessor INT was 5–20 microns in length and 30–120 nm in diameter.

The HRTEM images in Figure 3a,b display the range of daughter nanotubes obtained by plasma-treating of the multiwall WS2 nanotubes at 600 W for 40 min: tiny daughter nanotubes adjacent to the outer surface (Figure 3a) of the mother nanotube and a few isolated daughter nanotubes (Figure 3b). The amount of such daughter nanotubes increased with the treatment time from 10 to 40 min at 400 W plasma power. The extension of the plasma treatment time to 80 min did not reveal any additional improvement; however, the increase of plasma power from 400 to 600 W resulted in a sharp increase in the amount of the daughter nanotubes.

At 600 W and 40 min of treatment, a rough statistical estimate shows that daughter nanotubes were attached to about 80% of the plasma treated multiwall nanotubes. In comparison, only 10–20% of the multiwall WS2 nanotubes were covered with daughter nanotubes by a 400-W plasma treatment. Some nanostructures could be better described as nanoscrolls. However, the majority of the daughter nanostructures are nanotubes, having at least one perfectly closed layer. Future work will be focused on devising this technique to increase the yield of a single- to a few-layer nanotubes of WS2 or other INT, as well. Indeed, by irradiating MoS2 powder with a focused solar beam, single-wall MoS2 was rarely observed in the processed powder [21], which confirms that highly exergonic conditions produced by focused solar (laser) ablation may lead to the production of single-wall nanotubes of this kind.

**Figure 2.** (**a**) SEM and (**b**) TEM micrograph of a pristine multiwall WS2 nanotube.

**Figure 3.** TEM images of daughter WS2 nanotubes obtained by plasma ablation of multiwall inorganic nanotube (INT)-WS2 at 600 W for 40 min: (**a**) A large number of daughter nanotubes next to a treated multiwall nanotube; (**b**) A group of daughter nanotubes isolated from plasma-treated multiwall WS2 nanotubes by sonication.

Moreover, many daughter nanotubes were found attached and being tilted at *ca.* 30 or 60° with respect to the mother nanotube axis (see the white arrows in Figure 4a,b). However, some daughter nanotubes were found to be attached, having a common growth axis with the multiwall nanotube (see the black arrow in Figure 4b).

**Figure 4.** (**a**) TEM images of the daughter nanotubes tethered to the surface and tilted at approximately 30 or 60° with respect to the mother nanotube growth axis (white arrows); (**b**) TEM image of a daughter nanotube with growth axis parallel to the mother nanotube (black arrows).

These observations suggest that the small nanotubes were exfoliated by unzipping the outer walls of the mother nanotubes along specific crystallographic directions. In addition to nanotubes/nanoscrolls, a few layers-thick WS2 nanoplatelets of typical sizes in the range of the nanotubes' length, *i.e.*, 50–100 nm, were also observed.

In an attempt to separate the daughter nanotubes from the mother nanotubes, the plasma-ablated WS2 nanotube powder was ultrasonically treated in ethanol for 10 min. The high-resolution TEM (HRTEM) images in Figure 5 (see also Figure 3b) clearly depict the daughter nanotubes being more easily observed after detachment from the predecessor nanotubes.

Rough statistical analysis revealed that the interlayer distance in most daughter nanotubes varied between 6.3–6.5 Å (see Figure 5a), which is larger than the interlayer spacing of 2H-WS2 (6.23 Å/2-Theta = 14.32°) and multiwall nanotubes (6.31 Å/2-Theta = 14.13°) [2]. This observation suggests that the daughter nanotubes were not fully relaxed during the growth process and that the annealing of the sample could possibly lead to further structural relaxation. The Fourier (FFT) analysis (see the inset in Figure 5b) of the area framed by the square shows that the nanotube is chiral with a helical angle of six degrees. Once daughter nanotubes are observed in larger yields, techniques like ultracentrifugation could be used to separate them according to the number of layers and the length.

**Figure 5.** (**a**) HRTEM images of a two–three-layer nanotube with a non-uniform diameter after detachment from the large WS2 multiwall nanotube; (**b**) Another three-layer daughter nanotube after detachment. The Fourier (FFT) analysis (see the inset) of the area framed by the square shows that the nanotube is chiral with a helical angle of six degrees.

Energy dispersive X-ray analysis (EDS) within the TEM (not shown) confirmed that the nanotubes are made solely of tungsten and sulfur. Negligible traces of oxygen were found, which could be mainly attributed to surface impurities.

In a separate series of experiments, several powders of different microcrystalline layered materials, including 2H-WS2, 2H-MoS2 and 2H-NbS2, and also the respective diselenides, received a similar plasma treatment. No daughter nanotubes were found in these treated samples, whatsoever. A few layers-thick WS2 nanoplatelets with typical sizes in the range of the nanotubes (50–100 nm) were nevertheless abundant in plasma-treated 2H-WS2 powder. Furthermore, the NbSe2 powder turned out to be unstable under the plasma treatment conditions. Plasma treatment (400 W) of fullerene-like WS2 nanoparticles with hollow cage structure (inorganic fullerene-like (IF-WS2) nanoparticles) resulted in a few layers exfoliation and a few small-sized ("daughter") fullerene-like nanoparticles or nanotubes (see Figure 6). The daughter IF-WS2 nanoparticles are reminiscent of the arc-discharge produced IF-MoS2 nanoparticles [20]. In concluding this large series of experiments, it is possible to state that only plasma irradiation of multiwall WS2 nanotubes yielded daughter nanotubes in a reproducible fashion.

**Figure 6.** TEM image of the attached daughter single wall fullerene-like nanoparticles generated by plasma treatment of multiwall fullerene-like WS2 nanoparticles.

Single-wall carbon nanotubes can be obtained in large quantities, e.g., via the arc-discharge technique [22]. Given the interlayer distance of 3.4 Å in graphite, the monoatomic graphene plane can be closed into nanotubes of a diameter smaller than 0.5 nm [23,24]. On the other hand, the WS2 (MoS2) layer consists of six-fold bonded tungsten (molybdenum) atoms sandwiched between two sub-layers of three-fold bonded sulfur atoms. This makes the WS2 layers pretty rigid, with interlayer spacing of 6.23 Å; it is no wonder that the elastic energy for WS2 (MoS2) nanotube formation is appreciably larger than that of graphitic carbon. If one takes the elastic energy threshold for folding to be 0.05 eV/atom, the calculated diameter of a single-wall carbon nanotube is between one and 1.2 nm [25], and that of a single-wall MoS2 should be 6.2 nm [9,26]. Therefore, the diameters of the daughter WS2 nanotubes observed in the current series of experiments reconcile very well with the previous calculations.

#### *2.2. Growth Mechanism*

It is hypothesized that the formation of the daughter nanotube occurs through a strong interaction of the highly energetic plasma, used in this work, with a point or line defect on the outer surface of the mother nanotube, leading to rapid unzipping and exfoliation of 1–3 layers-thick WS2 fragments. One way for the exfoliated nanosheets to release the large elastic strain and fold into a nanotube is through an "inverted umbrella" reaction, which is the manifestation of the "Walden inversion" typical of a nucleophilic attack of a stereoisomer by an electron-rich moiety [27].

In an effort to understand the mechanism of formation of these daughter nanotubes, additional HRSEM analysis of the plasma-treated nanotubes was undertaken. The HRSEM in Figure 7a reveals a reversely revolved nanoscroll of ~20 nm in diameter attached to the surface of the mother nanotube. Furthermore, a clearly observed step defect or dark contrast on the mother nanotube beneath the daughter nanoscroll is reminiscent of the exfoliation process of the WS2 patch. Unfortunately, the resolution of the SEM did not permit viewing the smaller (3–7 nm) daughter nanotubes.

**Figure 7.** (**a**) HRSEM micrograph of a daughter nanoscroll attached to a large nanotube; (**b**) (I) HRTEM of a two-layer daughter nanoscroll viewed head-on along its axis and attached to a large multiwall nanotube; (II) an initial scrolling stage of an exfoliated single layer and three (III) layers before the formation of the nanotube; (**c**) HRSEM image of a daughter nanotube attached to a mother INT; and (**d**) HRTEM images revealing the folding of WS2 nanosheets to form a daughter nanotube.

Nonetheless, this analysis suggests very strongly that the elastic strain of the exfoliated WS2 sheet produces oppositely revolved daughter nanotubes. Moreover, at better resolution, the HRTEM image in Figure 7b depicts a two-layer daughter nanoscroll viewed head-on along its axis marked by "I". Furthermore, an initial scrolling stage of an exfoliated single layer ("II") and three layers ("III") before the formation of the nanotube was also observed. In addition, Figure 7c,d shows an HRSEM image of a daughter nanotube attached to a mother nanotube and HRTEM images of a WS2 nanosheet in the process of folding to form a daughter nanotube.

A schematic model for the growth mechanism of the daughter nanotubes is depicted in Figure 8. This growth mechanism proposes that the fragments of the outermost (1–3) layers of the predecessor nanotubes were unzipped by the plasma treatment, exfoliated and folded into daughter nanotubes. The large excitation energies of the plasma together with the mechanical strain lead to a nanoscopic "Walden-type inversion" [27]. The reactive edges of the inverted layers induce further folding into daughter nanotubes with a smaller radius of curvature than the predecessor (mother) multiwall WS2

nanotube. Detachment of the 1–3 layers from the mother nanotube may also be followed by rotation and inclination, in this case, the axes of the mother and daughter nanotubes do not necessarily coincide or form a specific angle between them. In other cases the rapid quenching of the excess energy of the nanosheets does not permit them to fully close, which leads to nanoscrolls or to a nanotube with one closed wall and the others remaining unclosed. Nanoscrolls may also occur due to steric hindrance, where the plasma-induced exfoliated nanosheets released their energy without being able to undergo timely inversion.

**Figure 8.** Schematics of the proposed growth mechanism of the daughter nanotubes by plasma treatment of the multiwall mother nanotubes.

In a few cases, WS*x* nanoclusters were observed adjacent to the daughter nanotubes (see Figure 9). These non-stoichiometric nanoclusters could be obtained by the condensation of tungsten and sulfur atoms or WS2 molecules from the vapor phase. In turn, the condensation of clusters onto the tube edges could lead to further elongation or even the growth of an extra layer on the daughter nanotube surface. Another plausible event is the condensation of the vapors into separate nanosheets, which, upon quenching, form isolated nanotubes.

The proposed mechanism is consistent with the data presented in this work. In order to shed light on the detailed growth mechanism of the (daughter) nanotubes and to control their length, diameter and the number of layers, future experiments will focus on the variation of the plasma treatment process, including the substrate temperature, pressure in the chamber, *etc.* 

Since highly excited clusters of WS2 (MoS2) can be formed using arc-discharge and a variety of other techniques, high-power plasma ablation would possibly allow synthesizing a few-wall nanotubes under controlled conditions in higher yields.

**Figure 9.** TEM images of nanoclusters surrounding daughter nanotubes. Presumably, the clusters were generated by the condensation of tungsten and sulfur atoms or WS2 molecular clusters from the vapor phase created by the plasma treatment of the multiwall WS2 nanotubes.

#### **3. Experimental Section**
