In spite of the attractive interest of α-MoS3 based nanomaterials for numerous applications in catalysis and energy, the amorphous nature of the MoS3 phase makes it challenging to control and understand its chemical reactivity[1]. In particular, the type of the structural building unit such as Mo3 triangular[2,3] vs. Mo chain[4,5] is still debated[6], while the ambivalent interpretation of the nature of sulfur species (S2, S22) and MoMo bonds leads to ambiguous interpretations of spectroscopic data and reactivity[2]. By density functional theory (DFT), we simulate the energetic, structural, and spectroscopic features of relevant 0D-, 1D- and 2D-MoS3 triangular, chain-like polymorphs, including unprecedented ones (ring, wave and helix) and revisit the interpretation of EXAFS, IR-RAMAN, and XPS experimental data.
To optimize the geometries, we use DFT as implemented in VASP relying on the Perdew-Burke-Ernzerhof (PBE)[7] functional within the framework of generalized gradient approximation (GGA). The long-range interactions were included through a density-dependent dispersion correction (dDsC)[8]. The projector augmented-wave (PAW)[9] method was chosen to describe the electron-ion interaction. Spin-polarized calculations were performed to obtain the electronic ground state of the clusters.
The exploration of some complex structures has been refined by ab initio molecular dynamics (AIMD) with the scaled velocity Verlet algorithm to solve Newton's equations of motions.
For frequency calculations, the structures have been optimized with tighter convergence criteria. To calculate the frequencies and corresponding infra-red intensities, the linear response method with density functional perturbation theory (DFPT) has been used as implemented in VASP.
The thermodynamic stability (including zero-point energy and entropy corrections), of the various oligomers have been calculated w.r.t. Mo3S9 unit of triangular oligomer as reference.
We analyze how MokS3k clusters of a few k atoms may grow up to infinite MoS3 polymorphs (figure 1(g)). The evolution of the growth energy and the computed IR spectra level suggest the coexistence of various polymorphs in the MoS3 phase as a function of sizes (figure 1(f)). Molecular dynamics simulations reveal how the small triangular MokS3k oligomers may transform into a more condensed MoS3 patches resembling embryos of the 2D 1T’-MoS2 phase (figure 1(c)). Finally, we discuss some plausible transformation pathways from one polymorph to another.[10]
As a perspective of this works, we are currently investigating the effect of an oxide support (alumina, used in many catalysts, on the transformation of MoO3 precursor into MoS3.

References
1. C. G. Morales-Guio, X. Hu, Acc. Chem. Res. 2014, 47, 8, 2671-2681.
2. Th. Weber et al., J. Phys. Chem. 1995, 99, 22, 9194-9200.
3. H. Jiao and Y-W. Li, et al., J. Am. Chem. Soc. 2001, 123, 7334-7339.
4. K. S. Liang, S. P. Cramer et al., J. Non-Cryst. Solids, 1980, 42, 345-356.
5. S. J. Hibble, G. B. Wood, J. Am. Chem. Soc. 2004, 126, 3, 959-965.
6. P. D. Tran et al. Nature Materials, 2016, 15, 640-647.
7. J. P. Pperdew et al. Phys. Rev. Lett. 1996, 77, 3865-3868.
8. S. N. Steinmann, C. Corminboeuf, J. Chem. Theory Comput. 2010, 7, 1990-2001.
9. G. Kresse et al. Phys. Rev. B 1999, 59 , 1758-1775.
10. A. Sahu, S. Steinmann, P. Raybaud. Cryst. Growth Des. (in press)