† Corresponding author. E-mail:
Project supported by the Foundation of Education Department of Shaanxi Province, China (Grant No. 16JK1461).
The EPR parameters of trivalent Er3+ ions doped in hexagonal GaN crystal have been studied by diagonalizing the 364×364 complete energy matrices. The results indicate that the resonance ground states may be derived from the Kramers doublet Γ6. The EPR g-factors may be ascribed to the stronger covalent bonding and nephelauxetic effects compared with other rare-earth doped complexes, as a result of the mismatch of ionic radii of the impurity Er3+ ion and the replaced Ga3+ ion apart from the intrinsic covalency of host GaN. Furthermore, the J–J mixing effects on the EPR parameters from the high-lying manifolds have been evaluated. It is found that the dominant J–J mixing contribution is from the manifold 2K15/2, which accounts for about 2.5%. The next important J–J contribution arises from the crystal–field mixture between the ground state 4I15/2 and the first excited state 4I13/2, and is usually less than 0.2%. The contributions from the rest states may be ignored.
In recent years much attention has been paid to the investigation of rare-earth (RE) doped GaN semiconductors, due to their potential in micro- and optoelectronics devices applications.[1–11] Of particular interest has been the electrical and optical properties of trivalent Er3+ in GaN host,[5–11] since the 1.54 μm luminescence center arising from the 4I13/2 → 4I15/2 transition is in the region of minimal loss in optical fibers.
It is well known that the open 4f shell of the trivalent 4f11 ion Er3+ is spatially protected from the 5s and 5p shells due to a greater radial extent. Accordingly, the central RE ions are less affected by the surrounding ligand ions when doped in a certain host, thus resulting in a larger spin–orbit interaction than the crystal field if compared with 3dN transition-metal complexes.[12–16] However, the case may be different for the RE doping from one host to another because of the distinct features between the central ion and the ligand ion, e.g., ionic radius, charge, mass, etc.
To gain a better insight into the local structures and optical properties of the RE materials, electronic paramagnetic resonance (EPR) is regarded as a suitable tool to investigate the interrelation between the electronic structure and molecular structure. For instance, Palczewska et al. has performed the EPR measurement on Er3+-doped single GaN crystal, and the corresponding EPR parameters g∥ = 2.861 ± 0.003, g⊥ = 7.645 ± 0.003, A∥ = (110 ± 5) × 10−4 cm−1, and A⊥ = (290 ± 5) × 10−4 cm−1 were obtained.[17] It should be noted that a rule of average g-factor gav = 1/3(g∥ + 2g⊥), obtained from experiment, may be applied to predict the assignment of the lowest Kramers doublet by comparing with the theoretical one.[18–20] Namely, if the average g-factor gav ≈ 6.0, where the contributions from all high excited states are ignored and only the pure 4I15/2 manifold is taken into account, the ground resonance states may be assigned to Kramers doublet Γ7 in Td symmetry, while the lowest states may be attributed to Γ6 if gav ≈ 6.8 for the same site symmetry.[21] The rule is appropriate for an axial C3v or C4v symmetry.[22,23] Consequently, Palczewska et al. assigned the ground states to Γ7 for the tetrahedral Er3+ center in GaN epilayers based on the fact that the experimental average g-factor (gav ≈ 6.05) is approximately equal to the expected gav of the doublet Γ7.[17] In order to account for these EPR parameters, Maâlej et al. recently carried out a theoretical study on the EPR parameters of the Er3+ ions in GaN crystal by means of the first-order perturbation method,[24] where only the ground manifold 4I15/2 is included. They also claimed that the lowest Kramers doublet belongs to the Γ7 based on the calculated average g-factor gav ≈ 6.14. However, to the best of our knowledge, the experimental gav of the most Er3+-doped complexes is generally less than the expected value 6.0 or 6.8 for the ground Kramers doublet Γ7 or Γ6, respectively.[19] Hence, the average g-factor gav for the hexagonal Er3+ center in GaN seems to be somewhat abnormal if the ground doublet is Γ7. For that matter, it is necessary to perform a detailed theoretical analysis on the EPR parameters of GaN:Er3+ system to examine the actual assignment associated with the ground Kramers doublet.
In this work, we will perform a quantitative theoretical study on the J–J mixing effects, arising from the high-lying excited states (such as the terms 4I13/2, 2K15/2, 2L15/2, etc.), on the EPR parameters based on a complete energy matrix method, where all excited states are included. Furthermore, the relation between the EPR g-factors and the orbit reduction factor k reflecting the covalent bonding and nephelauxetic effects is studied systematically. The plausible abnormal g-factors particularly for the average g-tensor gav of the Er3+ ion in GaN are reasonably expounded.
In general, the perturbation Hamiltonian for a 4f11 configuration ion Er3+ in an axial crystal field can be expressed as[25–29]
To study the EPR parameters without spoiling the overall coupling of the various interactions, the actual Zeeman operator has been included in Eq. (
Likewise, the magnetic hyperfine interaction ĤHF, which arises from the interaction between the unfilled electrons and nucleus, may be written as
In this work, a refined least-squares fitting method, rather than the regular fitting routine, has been developed by simulating the EPR parameters and optical spectra simultaneously, which may reflect the macro- and microscopic optical properties of the GaN:Er3+ system, respectively. In order to reduce the number of adjustable parameters in our calculations, all free-ion parameters except the spin–orbit coupling parameter ζ are taken as the same values as the free-ion parameters of Er3+ obtained by Carnall et al,[34] in view of the fact that the small change of free-ion parameters only affects the displacement of the centre of gravity of various manifolds, and hardly affects the Stark splitting and the EPR parameters. The obtained crystal–field parameters together with the free-ion parameters are tabulated in Table
With the aid of the operator-equivalent technique developed by Stevens et al.,[37] the J–J mixing effects on the EPR parameters from the manifolds 4I15/2, 4I13/2, 2K15/2, and 2L15/2 have been evaluated and collected in Table
Early studies have indicated that trivalent erbium ion Er3+ will occupy the position of Ga3+ site with local C3v symmetry when implanted in GaN host.[5] However, there is considerable difference of the ionic radii for the impurity ion Er3+ and the replaced cation Ga3+ of host (rEr3+ ≈ 0.89 Å, rGa3+ ≈ 0.62 Å). More overlap of electronic cloud between the central Er3+ ion and the ligand ions N3 − in the GaN:Er3+ system can be expected as a result of the larger ionic radius of impurity Er3+, and thus produce a stronger covalent bonding effect to some extent apart from the dominant covalency of host GaN. In the paper, in order to evaluate the covalent and nephelauxetic effects characterized by an average orbital reduction factor k as a whole,[38,39] the calculated results of the EPR g-factors (g∥, g⊥, gav) as a function of the orbital reduction factor k are obtained and collected in Table
A complete theoretical method has been established by diagonalizing the 364×364 energy matrix to study the EPR parameters of the trivalent Er3+ ions doped in hexagonal GaN crystal. The results show that the ground states may be derived from the Kramers doublet Γ6 for the GaN: Er3+ system, and the strong covalent bonding and nephelauxetic effects, as a result of the mismatch of ionic radii of the impurity Er3+ and replaced Ga3+ ion, may be mainly responsible for the larger deviation of the average EPR g-tensor gav from the theoretical value (gav ≈ 6.8) as expected. Moreover, the J–J mixing contribution to the EPR parameters from the high-lying excited states has been evaluated. Differing from the previous studies, it is found that the predominant J–J mixing contribution is the state 2K15/2, which accounts for about 2.5%, and the second major J–J contribution arises from the crystal–field mixture between the ground state 4I15/2 and the first excited state 4I13/2, and is usually less than 0.2%. The next J–J mixing contribution is from the state 2L15/2 which merely accounts for about 0.023% for g∥ and 0.022% for g⊥, respectively. The contributions from the rest states such as the 4I13/2, 4I11/2, 4I9/2, 4F9/2, and 2K13/2 may be ignored. It follows that the J–J mixing effects from the high manifolds could not really add more contribution to the final EPR parameters, which may be due to the larger spin–orbit coupling interaction than the crystal–field interaction for the RE-doped complexes.
1 | |
2 | |
3 | |
4 | |
5 | |
6 | |
7 | |
8 | |
9 | |
10 | |
11 | |
12 | |
13 | |
14 | |
15 | |
16 | |
17 | |
18 | |
19 | |
20 | |
21 | |
22 | |
23 | |
24 | |
25 | |
26 | |
27 | |
28 | |
29 | |
30 | |
31 | |
32 | |
33 | |
34 | |
35 | |
36 | |
37 | |
38 | |
39 |