High elasticity and strength of ultra-thin metallic transition metal dichalcogenides

Mechanical properties of transition metal dichalcogenides (TMDCs) are relevant to their prospective applications in flexible electronics. So far, the focus has been on the semiconducting TMDCs, mostly MoX2 and WX2 (X = S, Se) due to their potential in optoelectronics. A comprehensive understanding of the elastic properties of metallic TMDCs is needed to complement the semiconducting TMDCs in flexible optoelectronics. Thus, mechanical testing of metallic TMDCs is pertinent to the realization of the applications. Here, we report on the atomic force microscopy-based nano-indentation measurements on ultra-thin 2H-TaS2 crystals to elucidate the stretching and breaking of the metallic TMDCs. We explored the elastic properties of 2H-TaS2 at different thicknesses ranging from 3.5 nm to 12.6 nm and find that the Young's modulus is independent of the thickness at a value of 85.9 ± 10.6 GPa, which is lower than the semiconducting TMDCs reported so far. We determined the breaking strength as 5.07 ± 0.10 GPa which is 6% of the Young's modulus. This value is comparable to that of other TMDCs. We used ab initio calculations to provide an insight into the high elasticity measured in 2H-TaS2. We also performed measurements on a small number of 1T-TaTe2, 3R-NbS2 and 1T-NbTe2 samples and extended our ab initio calculations to these materials to gain a deeper understanding on the elastic and breaking properties of metallic TMDCs. This work illustrates that the studied metallic TMDCs are suitable candidates to be used as additives in composites as functional and structural elements and for flexible conductive electronic devices.


Experimental Methods
We obtained bulk crystals from HQ Graphene and exfoliated using Nitto blue tape over polydimethylsiloxane (PDMS) stamps for the transfer. Then, using a deterministic dry transfer method, thin crystals are transferred over the holes drilled on SiO2 substrate following Castellanos-Gomez et al. 1 . 3R-NbS2 crystals are synthesized on sapphire substrate via saltassisted chemical vapor deposition in a split-tube furnace. The growth recipe is provided in the following section. MFP-3D Asylum Research AFM is used for the indentation experiments. We used Tap300Al-G and Tap300DLC diamond-like coated tips from the Budget Sensors with force constant of 40 N/m. As mentioned in the main text, a dummy sample is scanned for about an hour before we conduct the indentation experiments on the real sample to limit the exposure of our samples to the ambient. Further scans are performed over the actual sample till the signs of drift disappears. During the indentation, displacement speed was controlled at 100 nm/s to get hysteresis free loading and unloading curves. We have chosen to wrinkle free flakes with large surface area, uniformly suspended over the holes to obviate inaccuracy, hysteresis, and slippage of flakes.
Deflection ( ) of the suspended flake can be found from the deflection of AFM cantilever (∆ ) and the scanning piezotube displacement (∆ ) through 2 : Force applied to suspended 2D flakes can be calculated using Hook`s law = ∆ here is deflection of AFM cantilever (40 Nm -1 ) that we calibrated using GetReal calibration in Asylum AFM, which first performs the thermal noise spectrum measurement followed by the Sader's method calibration. Finally, the thermal noise method is used one more time to obtain the inverse optical lever sensitivity. This method is slightly better than the hard strike in preserving the cantilever tip.

Computational Methods
First-principles calculations based on density functional theory (DFT) 3,4 as implemented in the Vienna Ab Initio Simulation Package (VASP) 5 were performed. The Kohn-Sham equations were solved using the projector augmented-wave method 6 and the exchange-correction functional was described within the generalized gradient approximation (GGA) 7 . The kinetic-energy cutoff for plane wave basis set taken to be 550 eV and 15x15x1, 15x15x11, 9x15x11 Gamma-centered kpoint mesh was used for the numerical integrations over the Brillouin zone for 2H-, 3R-, 1T-Electronic Supplementary Material (ESI) for Nanoscale Advances. This journal is © The Royal Society of Chemistry 2021 crystals, respectively. The lattice constants and atomic positions were optimized until the total energy and force convergence were below 10 -6 eV and 0.01 eV Å -1 , respectively. A vacuum space of 20 Å was taken along non-period direction to avoid the interactions between the neighboring images for monolayer crystals. The charge transfer analysis was performed by using Bader technique 8 . The mechanical response of monolayers in the elastic regime was determined by calculating the in-plane stiffness, 2 = 11 2 − 12 2 11 2 , where the cij's are the elastic constants including hydrostatic and shear terms 9 . The Young modulus of bulk crystals were calculated by using opensource code 10 based on Voigt-Reuss-Hill averaging scheme 11 . A denser k-point mesh was taken for calculations on mechanical properties.

CVD Growth of NbS2 and Wet Transfer Method
NbS2 flakes were grown using ambient pressure chemical vapor deposition method on a c-cut sapphire substrate. Before the growth, the sapphire substrates are cleaned using acetone, isopropanol, water and dried by blowing N2 gas. Niobium pentoxide (Nb2O5) powder, crushed salt (NaCl) and Sulfur are used as the growth precursors. Nb2O5 and NaCl are milled together using mortar and pestle. The mixture is transferred in an alumina boat and the sapphire substrate is placed over the boat facing down at a few millimeters to the precursor powder. Sulphur precursor is placed in a different alumina boat. Both precursors are positioned in a quartz tube and the tube is slid into the tube furnace. Figure S1 shows the schematic of the growth setup.

Figure S1
Schematic of CVD growth chamber for 3R-NbS2 crystals showing different temperature zones and configuration of sample and precursors.
Once the chamber is sealed, we purged the chamber using Ar gas flow rate 499 sccm for 6 minutes to get rid of the air inside the chamber. After 6 minutes, we set Ar at 20 sccm flow rate and started the CVD chamber heater then waited for it to reach at 785 o C. After 785 o C is achieved, Ar and H2 flow 100 sccm and 12 sccm respectively. After 10-15 minutes, we stopped the H2 flow and set Ar again at 20 sccm during cool down. We got large area thin NbS2 flakes which are exceedingly difficult to obtain via mechanical exfoliation. Figure S2 shows some examples of the 3R-NbS2 crystals on sapphire substrate. Figure S2 a and b show the optical microscope micrograph of 3R-NbS2 thin crystals grown on ccut sapphire via CVD. Scale bar is 10 µm.
XPS and EDX maps taken from the NbS2 flakes show X-Ray Na intercalation within the crystals. Figure S3 shows the XPS surveys for Nb, S and, Na after various etching cycles. We hypothesize that there might be Na intercalation between the layers as there are reports in the literature showing Na-NbS2 crystals.

Wet Transfer of NbS2 Crystals
Sapphire substrate with CVD grown NbS2 is spin coated with Poly(methyl methacrylate) PMMA 495 (A4) at 1300rpm. The substrate is heated at 180 o C for 6 minutes. Then, the substrate is dipped into buffered oxide etch (BOE) for 20 minutes to release the PMMA film along with the CVD grown crystals from the substrate. After rinsing, the film is released onto water surface by wedge transfer method and subsequently picked up by PDMS stamp. The PMMA film is dissolved in acetone for 20 minutes to transfer NbS2 crystals successfully. Finally, the desired crystal was transferred onto the holes using deterministic dry transfer technique as mentioned in main text. We performed Raman spectroscopy to confirm the phase of the crystals studied in this work. Figure S5 shows the exemplary Raman spectra we obtained for each crystal. Dashed lines indicated on the spectra shows the matching lines with the literature 12,13 . Although the Raman spectrum of NbS2 matches with the spectrum reported for 3R-NbS2 in the literature, as we discuss in the following sections, X-ray photoelectron spectroscopy (XPS) and energy dispersive spectroscopy (XPS) shows that the NbS2 crystals we have might have Na atoms interstitially between the layers. Figure S6 Optical microscope micrograph of a. 3R-NbS2 b. 2H-TaS2 c. 1T-NbTe2 and d. 1T-TaTe2. Scale bar is 10 µm.

Calibration of AFM Probes using GetReal Calibration Method
We used built-in GetReal method for automated calibration of AFM probes that uses Sader's method 14 and thermal noise method for the calibration of cantilever's spring constant and sensitivity (InvOLS) respectively.

Force-Indentation Loading Curve
A full cycle of the − loading curve is given in Figure S7. Important points in the curve are marked on the graph. First the tip starts above the membrane. When the tip is sufficiently close, the tip snaps onto the membrane and indentation continues till the mechanical failure of the film. Then, the tip rapidly descends to the bottom of the hole and starts applying force there again.

SEM Micrographs of AFM tips After a Single Measurement
We characterized the changes on a single AFM tip using SEM imaging. We first acquired SEM image of a pristine tip. Then, scanned the sample and took another SEM image. We installed the same tip to the AFM, this time we scanned the sample and performed an indentation and took an SEM image. Finally, we used the same tip for a scan, indentation and breaking and acquired an SEM image. Figure S10 shows the series of SEM images after each step and gradual degradation of the AFM tip.
Figure S10 a. SEM micrograph of a pristine tip. The tip radius is measured as 12 nm. b. SEM micrograph of the same tip in a after a single scan typically performed to map the crystal. Tip radius is 23 nm. c. Same tip in b after a second scan followed by an indentation. Tip radius is 28 nm. d. Same tip in c after a third scan, indentation and a breaking. Tip is now blunt.
Before the measurements we picked up 10 random tips from the box and measured the diameter of the AFM tips. We found out that the tip radii are within the specifications of the manufacturer (10.0 ± 1.0 nm). We used each tip for a single indentation and breaking measurement and after every other measurement we measured the tip radius using SEM imaging ( Figure S11). For the measurements we have the measured tip value, we used the measured value, otherwise we used the averaged tip radius 22.0 ± 2.0 nm.

XPS Survey on bulk metallic TMDCs
We performed XPS surveys on exfoliated TDMCs right after exfoliation and after keeping the samples under ambient for an hour. Except TaS2 other metallic TMDCs showed significant signs of oxidation. Figure S13 shows XPS surveys of the chalcogen and metal constituents of the metallic TMDCs.