Can calculated harmonic vibrational spectra rationalize the structure of TiC-based nanoparticles?

Nanoscale titanium carbide (TiC) is widely used in composites and energy applications. In order to design and optimize these systems and to gain a fundamental understanding of these nanomaterials, it is important to understand the atomistic structure of nano-TiC. Cluster beam experiments have provided detailed infrared vibrational spectra of numerous Ti x C y nanoparticles with well defined masses. However, these spectra have yet to be convincingly assigned to Ti x C y nanoparticle structures. Herein, using accurate density functional theory based calculations, we perform a systematic survey of likely candidate nanoparticle structures with masses corresponding to those in experiment. We calculate harmonic infrared vibrational spectra for a range of nanoparticles up to 100 atoms in size, with a focus on systems based on removing either four carbon atoms or a single titanium atom from rocksalt-structured stoichiometric TiC nanoparticles. Our calculations clearly show that Ti-deficient nanoparticles are unlikely candidates to explain the experimental spectra as such structures are highly susceptible to C-C bonding, whose characteristic frequencies are not observed in experiment. However, our calculated infrared spectra for C-deficient nanoparticles have some matching features with the experimental spectra but tend to have more complex infrared spectra with more peaks than those obtained from experiment. We suggest that the discrepancy between experiment and theory may be largely due to thermally induced anharmonicities and broadening in the latter nanoparticles, which are not be accounted for in harmonic vibrational calculations.