Percolation picture for long wave phonons in zinc-blende mixed crystals: from (Zn, Be) chalcogenides to (Ga, In)As

Abstract

We propose a simplified version of the one-bond two-mode percolation model originally developed for the long wave phonons related to the stiff Be–VI bond in (Zn, Be) chalcogenides, which open the class of random mixed crystals with contrast in the bond stiffness. This is deduced from the comparison between the Raman responses from the stiff Be–VI bond and the soft Zn–VI one in these systems. The simplified version is tested on (Ga, In)As, made of soft-like bonds only and taken here as a representative challenging system. This results in a successful reinterpretation of the puzzling multi-phonon behaviour in the Raman/infrared spectra of this alloy, that has been a subject of debate. The discussion is supported by contour modelling of the TO and LO Raman lineshapes by applying the Hon and Faust treatment to a version of the modified-random-element-isodisplacement model generalized to multi-oscillators. Also, the assignment of the long wave phonons in (Ga, In)As is supported by atomistic calculations of the bond length distributions of the minority bond species in large (Ga, In)As supercells corresponding to alloy compositions close to the In–As (In ∼0.19) and Ga–As (In ∼0.81) bond percolation thresholds. The configurations are analysed to distinguish between isolated and connected bonds, not in the usual terms of next nearest neighbours.

Combined Raman study of InGaAsN from the N-impurity and InGaAs-matrix sides

“…8 lacking so far, probably due to the fact that the traditional 1-bond→ 1-mode description of the phonon mode behavior in random InGaAs, which makes a natural reference here, has been itself a much debated subject. 14 In this work, we report a combined Raman study of In 0.08 Ga 0.92 As 0.98 N 0.02 layers within both N-and InGaAsrelated spectral ranges. The N-related phonons are assigned from ab initio calculations of the zone-center phonon density of states ͑ZC-DOS͒ of GaAs-based supercells with ͑In, N͒ motifs among the most likely to occur in real crystals.…”

“…An unexplored issue of whether the N-free In atoms distribute randomly or not in GaAs is addressed by comparing the transverse and longitudinal optical ͑TO,LO͒ Raman signals to reference line shapes calculated via our phenomenological 1-bond→ 2-phonon "percolation" model recently proposed for random InGaAs. 14 Raman measurements are performed on two ͑001͒ InGaAsN / GaAs pseudomorphic layers grown by MBE, 15 of similar composition, i.e., ͑ϳ2% N, ϳ8% In͒, as determined by x-ray diffraction. For an insight into the N-related phonon modes we use a thin ͑100 nm͒ perfectly strained layer.…”

“…We stress that the input parameters used to calculate the whole set of LO lines come from the original curve fitting in the TO symmetry, which, to our view, provides the only safe basis for reliable analysis of multiphonon systems. 14 The corresponding curves in the random case are added as dotted lines, for reference purpose. These were calculated directly from Fig.…”

Lattice vibrations of Zn1 − x Cd x Se semiconductor alloy in the two-mode and percolation models of rearrangement of the vibrational spectrum with the composition

“…For alloy crystals with the modified two mode rearrange ment of the vibrational spectrum, the MREI model is invalid. To analyze the rearrangement of the vibra tional spectrum of the alloy with the above indicated composition, Pagés et al [7] proposed a fundamen tally new concept based on the percolation site theory [8]. In the percolation (1 bond 2 mode) scheme proposed for the vibrational spectrum, the A 1 -x B x C alloy crystal is considered as a composite medium formed of two coexisting AC and BC like regions that appear as a result of natural local fluctuations in the alloy composition.…”

Description of vibrational properties of random alloy ZnTe1–xSex within the percolation model

“…In a recent work, we applied this approach successfully to the In 1-x Ga x As [6] and In 1-x Ga x P [7] random alloys. The purpose of this paper is to investigate whether the percolation model is also able to account for the TO behavior of the ZnTeSe alloy -on a quantitative basis, which the traditional MREI approach fails to do (refer above).…”