Non-cubic Er centres in ZnSe studied by electron paramagnetic resonance and optical analysis

Abstract

EPR and optical ( 4 f 4 photoluminescence and photoluminescence excihdon) speara due to Er centres are studied on bulk &Se crystals, which were gmwn by the high-pressure Bridgman technique and doped with ErF? (and pMly. in addition. with Liz-) during crystal growth. Besides the well known. almost isotropic signal with 8 = 5.94, which has been assignedto isolated E*+ on Zn lattice sires. we observe three strongly anisotropic EPR s p c h a (A, B and B' ) which are due to haositions in non-cubic Er'+ (spectrum A) and @+ (specua B and B') centres. The symmetry axes of these centres have directions close to (111) (A) and parallel to (I 10) @ and B'). The angular dependences of the signals are influenced by twinning effects.In the erystals doped with ErF3 alone. only a single type of centre, which is obviously identical with the type A EPR cenho, manifests itself in photoluminescence. A crystal-field analysis of the companding BPR and optical spectra shows that this antre has a rs-type ground level and is chamtenzed by a clystal-field parameter ratio A 6 ( r 6 ) / A j ( / ) = -0.22 We thhk that this centre is a complex consisling of E?+ on a Zn site and F on a nearest-neighbour interstitial site.The g-factors found for the EPR signals B and B' can be explained on the basis of non-Kramers doublet ground levels of E?+ 4fL2 which result from the splitting of cubic rs triplets due to non-cubic crystal-field components. These two signals are ascribed (0 E?+ on the two zincblende-type interstitial sites, respectively, each forming a complex with some other kind of atom on its next-nearest-neighbour interstitial site.A discussion is given of the result that the EPR spectra (unlike the optical spectra) were detected only in lhe case of Li2C0, codoping.

Electron paramagnetic resonance and photoluminescence study of Er-impurity complexes in Si

“…Despite the absence of any hyperfine lines as a positive signature of Er, the high degree of anisotropy is indicative of a paramagnetic center where the orbital angular momentum is not appreciably quenched, such as a rare earth ion. Furthermore the similarity of the principal g values for centers OEr-1, OEr-1Ј, and OEr-3 to those reported [26][27][28] for other Er 3ϩ centers strongly suggest that these centers are Er-related centers. In addition the idea that the centers, and also centers OEr-2, OEr-2Ј, and OEr-4, involve Er 3ϩ , rather than the non-Kramers ion Er 2ϩ , is further confirmed by the fact that they were fitted by the spin Hamiltonian in Eq.…”

“…If the overall crystal field has less than cubic symmetry, the principal components of the g tensor can be related to the g value predicted for cubic symmetry g c by [26][27][28] …”

“…26,27 This analysis is also only valid where the g value in cubic symmetry is independent of the strength of the crystal field, which in this case is for the ⌫ 6 or ⌫ 7 states. This approach has been used successfully in the interpretation of rare-earth EPR spectra.…”

“…This approach has been used successfully in the interpretation of rare-earth EPR spectra. [26][27][28] Small differences ͑Ͻ0.04͒ between g c and g av are usually explained in terms of interactions with higher-lying energy levels or the effects of covalency. 28 If the difference between the value of g c and g av is large (Ͼ0.1), 28 the assumption that the low symmetry components of the crystal field are small compared to when the cubic crystal field breaks down.…”

Electron paramagnetic resonance of erbium doped silicon

“…͑1͒ would have to be used. Furthermore, the similarity of the principal g values to those reported [18][19][20][21][22] for some Er 3ϩ centers strongly suggest that the resonances are due to Er 3ϩ although the hyperfine lines associated with the 23% abundant Er 167 isotope with nuclear spin Iϭ7/2 could not be clearly resolved above the noise level. Since Er 3ϩ has eleven 4 f electrons, the ground state in a crystal field of any symmetry will be at least a doublet in accordance with Kramers' theorem.…”