Study of wide bandgap crystal LiCaAlF 6 by IR-reflection spectroscopy and ab initio calculations

Polarized IR-reflection spectra and results of ab initio calculations of vibrational and electronic properties of LiCaAlF6 single crystal are presented. It is shown that the crystal band gap is direct. Experimental and theoretical parameters are obtained for dipole-active and all phonons, respectively, including silent modes. Experimental IR-reflection and Raman spectra are well described in the frame of results obtained by ab initio calculations. The peculiarities are discussed concerning the structure of electronic bands, the interatomic interactions, the character of lattice vibrations, and the phonon dispersion. Isostructural crystals with general chemical formula LiMM′F6 (M = Ca,Ba,Sr; M′ = Al, Ga) possess a very wide band gap (Eg = 12.65 eV for LiCaAlF6 [1]). That’s why they are perspective for applications in UV spectroscopic region. The crystal LiCaAlF6, pure as well as doped by dand f-ions, is intensively studied. The list of its applications includes laser emission, scintillators, visualization of neutrons, UV optical windows, etc. The phonon spectrum of LiCaAlF6, necessary for the description of relaxation processes and vibronic spectra of f-d transitions [2], is not sufficiently studied. Raman phonons were studied in Ref. [3]. In this work, we present results of IR spectroscopic study as well as of ab initio calculations of LiCaAlF6. Crystal structure of LiCaAlF6 is trigonal [4], Li2ZrF6 type, space group is P31c (D3d, No. 163), z=2. Irreducable representations of phonon modes are classified as following Г=3A1g+5A2g+8Eg+4A1u+6A2u+10Eu [3], A1g and Eg being Raman active, A2u and Eu – IR-active, A2g and A1u – silent modes. Reflectivity spectra were registered using Brucker IFS66 fourier-spectrometer. Ab initio calculations were performed by two methods. Phonon properties in the center of Brillouin zone were calculated in the DFT-LCAO approximation with the use of CRYSTAL 14 code [5] and B3LYP hybrid pseudopotential [6]. Calculations of phonon and electronic dispersions across Brillouin zone were performed in the basis of plane waves using the code QUANTUM ESPRESSO [7]. Experimental reflectivity spectra of LiCaAlF6 are presented in Fig.1 by symbols. The experimental parameters of IR phonons were obtained by fitting the spectra in the code RefFIT [8] using Drude-Lorentz model for dielectric constant. Model spectra with the best fit are presented for comparison in Fig. 1 by grey curves. The following TO frequencies (cm–1) for IR active modes were obtained: 131, 284, 299, 400, 588 (A2u modes), 122, 190, * Corresponding author: mavrin@isan.troitsk.ru DOI: 10.1051/ , 0100 (2017) 713201007 EPJ Web of Conferences epjconf/201 SPECTROSCOPY.SU 2016 132 7 © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). 197, 269, 324, 360, 417, 451, 575 (Eu). We have to note that weak phonon modes (marked by arrows in Fig. 1) were reliably identified after ab initio calculation performed. Calculated by both ab initio methods structural parameters coincide with experimental ones [9] within 1 % accuracy. Partial densities of electronic states and dispersion of electronic states was obtained, according to which, LiCaAlF6 is a direct band-gap isolator. Theoretically obtained phonon frequencies are in good agreement with experimental ones, both IR and Raman active. The energies (in cm–1) of silent phonon modes were also obtained (LCAO gives 244, 275, 406, 579 (A1u modes), 31, 237, 305, 488, 726 (A2g)). Calculated IR reflectivity spectra (black curves at the bottom of Fig. 1) describe well experimental results. Analysis of TO-LO splittings was performed. One inverted phonon was found (Eu mode 451 cm–1). Due to the interaction of weak phonon, falling into the TOLO gap of the strong mode, with the last one, its TO-LO splitting inverts [10-13]. Fig. 1. Experimental (symbols), model (grey curves), and calculated ab initio (black curves at the bottom) reflectivity spectra of LiCaAlF6 for (а) A2u and (b) Eu vibrational modes. Arrows point to weak phonons The phonons with energies higher than 500 cm–1 are separated by a gap from lowfrequency ones. The analysis of atomic displacement in normal modes revealed predominantly stretching character of vibrations for high-frequency gapped modes. Phonon density of states calculated is also characterized by a gap in the 470-540 cm–1 spectral region. Phonon dispersion curves demonstrates strong softening (up to 3-5 cm–1), close to A-point of the Brillouin zone, for two phonon modes, namely, for the lowest-energy phonons with symmetry A2g (98 cm–1) and Eg (132 cm–1). This can be interpreted as quasidoubling of primitive cell along z axis. The authors are grateful to S.L. Korableva, who provided us with the LiCaAlF6 crystal.

Isostructural crystals with general chemical formula LiMM′F6 (M = Ca,Ba,Sr; M′ = Al, Ga) possess a very wide band gap (Eg = 12.65 eV for LiCaAlF6 [1]).That's why they are perspective for applications in UV spectroscopic region.The crystal LiCaAlF6, pure as well as doped by d-and f-ions, is intensively studied.The list of its applications includes laser emission, scintillators, visualization of neutrons, UV optical windows, etc.The phonon spectrum of LiCaAlF6, necessary for the description of relaxation processes and vibronic spectra of f-d transitions [2], is not sufficiently studied.Raman phonons were studied in Ref. [3].In this work, we present results of IR spectroscopic study as well as of ab initio calculations of LiCaAlF6.
Reflectivity spectra were registered using Brucker IFS66 fourier-spectrometer. Ab initio calculations were performed by two methods.Phonon properties in the center of Brillouin zone were calculated in the DFT-LCAO approximation with the use of CRYSTAL 14 code [5] and B3LYP hybrid pseudopotential [6].Calculations of phonon and electronic dispersions across Brillouin zone were performed in the basis of plane waves using the code QUANTUM ESPRESSO [7].
Experimental reflectivity spectra of LiCaAlF6 are presented in Fig. 1 by symbols.The experimental parameters of IR phonons were obtained by fitting the spectra in the code RefFIT [8] using Drude-Lorentz model for dielectric constant.Model spectra with the best fit are presented for comparison in Fig. 1 by grey curves.The following TO frequencies (cm -1 ) for IR active modes were obtained: 131, 284, 299, 400, 588 (A2u modes), 122,190,197,269,324,360,417,451,575 (Eu).We have to note that weak phonon modes (marked by arrows in Fig. 1) were reliably identified after ab initio calculation performed.
Calculated by both ab initio methods structural parameters coincide with experimental ones [9] within 1 % accuracy.Partial densities of electronic states and dispersion of electronic states was obtained, according to which, LiCaAlF6 is a direct band-gap isolator.Theoretically obtained phonon frequencies are in good agreement with experimental ones, both IR and Raman active.The energies (in cm -1 ) of silent phonon modes were also obtained (LCAO gives 244, 275, 406, 579 (A1u modes), 31, 237, 305, 488, 726 (A2g)).Calculated IR reflectivity spectra (black curves at the bottom of Fig. 1) describe well experimental results.Analysis of TO-LO splittings was performed.One inverted phonon was found (Eu mode 451 cm -1 ).Due to the interaction of weak phonon, falling into the TO-LO gap of the strong mode, with the last one, its TO-LO splitting inverts [10][11][12][13].The phonons with energies higher than 500 cm -1 are separated by a gap from lowfrequency ones.The analysis of atomic displacement in normal modes revealed predominantly stretching character of vibrations for high-frequency gapped modes.Phonon density of states calculated is also characterized by a gap in the 470-540 cm -1 spectral region.Phonon dispersion curves demonstrates strong softening (up to 3-5 cm -1 ), close to A-point of the Brillouin zone, for two phonon modes, namely, for the lowest-energy phonons with symmetry A2g (98 cm -1 ) and Eg (132 cm -1 ).This can be interpreted as quasidoubling of primitive cell along z axis.
The authors are grateful to S.L. Korableva, who provided us with the LiCaAlF6 crystal.