Longitudinal unzipping of 2D transition metal dichalcogenides

Unzipping of the basal plane offers a valuable pathway to uniquely control the material chemistry of 2D structures. Nonetheless, reliable unzipping has been reported only for graphene and phosphorene thus far. The single elemental nature of those materials allows a straightforward understanding of the chemical reaction and property modulation involved with such geometric transformations. Here we report spontaneous linear ordered unzipping of bi-elemental 2D MX2 transition metal chalcogenides as a general route to synthesize 1D nanoribbon structures. The strained metallic phase (1T′) of MX2 undergoes highly specific longitudinal unzipping owing to the self-linearized oxygenation at chalcogenides. Stable dispersions of 1T′ MoS2 nanoribbons with widths of 10–120 nm and lengths up to ~4 µm are produced in water. Edge abundant 1T′ MoS2 nanoribbons reveal the hidden potential of idealized electrocatalysis for hydrogen evolution reactions at a competitive level with the precious Pt catalyst.

). Afterwards, we kept 0.48 M as the optimal concentration of n-BuLi and investigated the effect of reaction temperature ( Supplementary Fig. 13B). Interestingly, we found that the composition of 1T′ phase in MoS2 increased up to 90 °C and a further increase in temperature was found to decrease the composition of 1T′ phase. For more insight on this behaviour, we have studied the effect of intercalation time at different temperatures. As shown in Supplementary Fig. 13C, while intercalation time did not have a significant effect at 20 °C, the composition of 1T′ phase increased steadily at 90 °C and reached a maximum of ~99% at 48 h. A further increase of intercalation time was found to adversely affect the evolution of 1T′ phase. From the results, we can assume that both high temperature and long exposure may cause the restoration of 2H phase due to probable desorption of intercalated Li ions. XPS spectra of Mo3d displays the gradual evolution of 1T′ phase ( Supplementary Fig. 13D). Note that, the optimization parameters described here for achieving 1T′ phase may greatly depend on the flake size and crystal quality of bulk MoS2.
Supplementary Note 2. Synthesis of 1T′ WSe2. Typical wet chemical and chemical vapor deposition (CVD) synthetic methods for 2D WSe2 have been reported to form 2H phase thus far 4,5 .
Colloidal synthesis of 1T′ WSe2 flower-like 3D nanostructures with an average diameter of 200 nm has been reported 6 . However, to the best of our knowledge, there is no report for the direct 3 production of 2D 1T′ WSe2 sheets yet. Chemical exfoliation using organolithium intercalation treatment of bulk layered WSe2 followed by exfoliation in water result in predominantly 2H phase WSe2 sheets with a very small fraction of 1T′ 7 . It is noteworthy that we have achieved 99.4% 1T′ phase of WSe2 sheets by optimizing exfoliation parameters in this work. Supplementary Fig. 13E provides the result of initial optimization experiments carried out to investigate the intercalation assisted chemical exfoliation of bulk WSe2 (2H phase) using different n-BuLi concentrations ranging from 0.064 M to 0.64 M. The composition of 1T′ phase in the exfoliated WSe2 was found to increase up to 47% when the concentration of n-BuLi was 0.32 M. No significant improvement in the composition of 1T′ phase has been observed with a further increase of n-BuLi concentration after the 24 h of intercalation reaction. Li intercalation reaction at 40 °C was found to yield the highest composition of 1T′ phase in the exfoliated WSe2 (82.6%) after the 24 h of reaction ( Supplementary Fig. 13F). Kinetic experiments with different intercalation times showed that an increase of reaction time adversely affects the 1T′ phase at 40 °C, which yielded the maximum 1T′ phase of 98.7% after 12 h of reaction ( Supplementary Fig. 13G), revealing the significance of reaction time, temperature and concentration in the phase change of WSe2. XPS spectra of displaying the gradual evolution of 1T′ phase is provided in Supplementary Fig. 13H. Note that, the optimization parameters described here for achieving of 1T′ phase may greatly depend on the flake size and crystal quality of bulk WSe2.

Supplementary Note 3. Synthesis of 1T′ MoSe2 and MoTe2. Chemical exfoliation of
MoSe2 nanosheets through n-BuLi treatment has been reported to form a mixture of 1T′ and 2H phases. The maximum 1T′-phase concentration achieved thus far is ~65% 2 . By contrast, chemical exfoliation method for MoTe2 has not been reported yet. The synthesis of 1T′-MoTe2 is mainly relying on chemical vapor deposition (CVD) or colloidal synthesis 8,9 . We have carefully optimized the reaction condition for chemical exfoliation in order to achieve maximized concentrations of 1T′ phase in MoSe2 and MoTe2. As shown in Supplementary Fig. 14A, the maximum yield of 1T′ η =b log j + c (S1) where b is the Tafel slope and j is the current density.
The exchange current density (j0) was determined from the following equation (S2): where b and c are the Tafel slope and intercept, respectively, in equation (S1).

Supplementary Note 5. Computational Details.
Spin-polarized density functional theory (DFT) computations were performed using the Vienna Ab-initio Simulation Package (VASP) with the projector-augmented-wave (PAW) method to account for core-valence interactions [10][11][12] . The exchange-correlation interactions were described by generalized gradient approximation (GGA) with Perdew-Burke-Ernzerhof (PBE) functional 13 . Moreover, van der Waals correction was also 5 incorporated using an empirical scheme by Grimme as implemented within the framework of VASP for DFT-D3 method 14 . A polarized continuum model of Hennig and co-workers implemented in VAPS as VASPsol was used to simulate the effect of a water solvent environment with the dielectric constant set to 78.4 15 . For a modeling of 2D surfaces, we considered (3x3) and (4x2) supercells for 2H and T′ phases of MX2, respectively, with a vacuum of 20 Å along zdirection to avoid undesired interactions due to the periodic images ( Supplementary Fig. 15A,B).
A cutoff energy of 500 eV was used to generate plane wave basis set and the Brillouin zone was sampled using 3x3x1 Monkhorst-Pack k-points mesh 16 . All the structures were relaxed until the convergence criterion of 10 -5 eV and 0.05 eV Å -1 were met for self-consistent field (SCF) energies and forces on each atom, respectively. In order to test the HER activity on the edge sites of 1T′ MX2, we modeled a hydrogen passivated nanoribbon (Supplementary Fig. 15F) structure with a vacuum of 20 Å in the y-direction. Two kinds of edge structures are created by the above procedure, including the one that ends with the exposed metal atoms (Mo/W) and the other ending with chalcogen only (S/Se/Te).
All the dangling bonds on these edges are capped by the H atoms.
We considered efficient Volmer-Heyrovsky reaction mechanism for the evaluation of theoretical overpotential for HER at the surface and edges of MX2 systems 17  Changes in the free energies given in Fig. 5A (main text) is that of the exposed metal edge sites, which are the most active sites for HER in the MX2 NR structures.  Supplementary Note 11. Local density of states (LDOS) analysis: LDOS diagram for X (S/Se/Te) atoms in 1T′ MX2 (Fig. 2C and Supplementary Fig. 15C-E) show the existence of different degrees of 3p projected density of states for X atoms. The X atoms in S1 sites have a higher density of states at Fermi level, which reflects its strong binding nature towards O atom compared to the X atom at S2 site. This tendency has been consistent in all the tested MX2 (MoS2, MoSe2, MoTe2, and WSe2). The ECSA of the electrode was determined by measuring the electrochemical double layer capacitance (Cdl) obtained from the CV curves from 10 to 200 mV/s in the non-faradic region, as shown in supplementary Figure 19. The half of cathodic and anodic current density difference (∆j/2, ∆j=ja-jc) is plotted against the scan rate. The absolute value of slope gives Cdl.
The ESCA is directly related to Cdl by the following equation 27 : where Cs is the specific capacitance, which is generally considered to be in the range of 20-60 µF/cm 2 for a flat surface. The average of cathodic and anodic current density as a function of scan rate.