First-Principles Calculation on the Emission Energy Level of Ruby Based on DV-Xα Molecular Orbital Method and Ligand Field Theory
Mega Novita,Akane Ito,Kazuyoshi Ogasawara
DOI: https://doi.org/10.1149/ma2016-02/42/3176
2016-09-01
ECS Meeting Abstracts
Abstract:The transition metal 3d 3 ions activated in various host crystals have been studied in a great number of experiments and calculations. Their energy levels in strong crystal field namely 4 A 2 , 2 E, 2 T 1 , 2 T 2 , 4 T 2 , and 4 T 1a play important roles in the luminescence process. Especially, the transition energy from 2 E state to 4 A 2 state or R-line has been utilized for the emission process. Since the development of ligand field theory (LFT), the first-principles calculations have been helpful to understand the behavior of R-line. In 1982, Ohnishi and Sugano [1] carried out the semi-empirical first-principles calculation based on the discrete variational Xα (DV-Xα) method [2] to estimate the emission energy of ruby (Al 2 O 3 doped with Cr 3+ ) at various bond lengths under O h symmetry. In their estimation, the barycenter of t 2 g 3 configuration was an important parameter to estimate the R-line energy. Although the results are reasonable, the barycenter was defined as the average of only 3 multiplet states i.e. 2 E, 2 T 1 , and 2 T 2 states. However, according to LFT, the t 2 g 3 configuration consists of 4 multiplet states not only 2 E, 2 T 1 , and 2 T 2 states but also the ground state, 4 A 2 . Therefore, here we proposed a new calculation formula to improve the accuracy of R-line energy by considering 4 multiplet states such as 4 A 2 , 2 E, 2 T 1 , and 2 T 2 states. Two types of model clusters, non-optimized and optimized model clusters consisting of 63 atoms were used for the calculations. The impurity Cr 3+ ion was located at the center of the model clusters. The non-optimized model cluster was constructed based on the experimental crystal structures of α-Al 2 O 3 under pressures with hexagonal A 2 X 3 structure. On the other hand, since the local structure in the actual materials is different from the original ones, the optimized model cluster was constructed to consider the effect of lattice relaxation. It was estimated by performing geometry optimization using the CAmbridge Serial Total Energy package (CASTEP) code [3]. Firstly, the structural optimizations of pure α-Al 2 O 3 crystal under various pressures were carried out and then followed by the geometry optimization of α-Al 2 O 3 crystal doped with Cr 3+ ion. The molecular orbitals energies were then estimated by performing one-electron calculations using DV-Xα method for both model clusters. During the calculations, the local C 3 symmetry was maintained to minimize the computational error. Next, the R-line energies were then estimated by two different approaches i.e., (1) based on the approach proposed by Ohnishi’s group and (2) based on LFT. Figure 1 shows the theoretical R-line energies of ruby under pressures calculated using the optimized model clusters. E R 1 and E R 2 indicate the estimated R-line energies based on the approach proposed by Ohnishi’s group and those based on LFT, respectively. For comparison, the R-line energies of ruby at various Cr-O bond length calculated under O h symmetry [1] and the experimental R-line energies of ruby under pressures observed within C 3 symmetry [4] are shown together in Fig. 1. Our results show that by both approaches, the R-line energies decrease as the applied pressure increases. Accordingly, simple estimation based on one-electron calculations has successfully reproduced the tendency of ruby’s emission energies. Compare to the experimental values, the R-line energies estimated based on Ohnishi’s approach were significantly underestimated. On the other hand, the calculated R-line energies based on LFT improved the accuracy towards the experiment excellently. References [1]. S.Ohnishi and S.Sugano. Jpn. J. Appl. Phys. 21 (1982) L309-L331. [2]. H. Adachi, M. Tsukada, and C. Satoko. J. Phys. Soc. Jpn. 45 (1978) 875-883. [3]. M. C. Payne, M. P. Teter, D. C. Allan, T. A. Arias, and J. D. Joannopoulos. Rev. Mod. Phys. 64 (1992) 1045-1097. [4]. J. H. Eggert, K. A. Goettel, and I. F. Silvera. Phys. Rev. B 40 (1989) 5724-5732. Figure 1