Long-wave phonons inZnSe−BeSemixed crystals: Raman scattering and percolation model

Abstract

We extend to longitudinal-optical (LO) phonons the percolation model set for the basic understanding of the atypical transverse-optical (TO) one-bond→ two-mode behavior observed by Raman scattering in the Be-Se spectral range of the random Zn 1−x Be x Se alloy ͑0 ഛ x ഛ 1͒, which opens the class of mixed crystals with contrast in the bond stiffness. The study is supported by contour modeling of the TO and LO Raman line shapes. This is achieved via application of the Hon and Faust treatment to a version of the modified-randomelement-isodisplacement model generalized to multioscillators. While the TO signal clearly discriminates between Be-Se vibrations within the hard Be-rich region and the soft Zn-rich one, complexity arises in the LO symmetry due to vibration coupling via the 1ong-range longitudinal polarization field. In particular this generates a massive transfer of oscillator strength from the low-frequency ͑LO − ͒ (hard, soft)-mixed mode to the high-frequency ͑LO + ͒ one, which results in an apparent LO + single-mode behavior. Moreover the contrasts between the Zn-Se and Be-Se bond lengths and bond stiffness are proposed to force a Verleur and Barker-like (VB) discrete multimode Raman response from each region. Accordingly LO − and LO + intramode transfers of oscillator strength superimpose to the LO − → LO + intermode one. This accounts for the spectacular distortions of the LO + line shape. On the whole, the puzzling LO behavior can be regarded as the result of a cooperative phenomenon between two discrete assemblies of polar LO phonons, driven by the long-range longitudinal polarization field. Also, the Verleur and Barker description accounts for subtle unexplained behaviors in the TO symmetry. More generally it appears to provide a much attractive area for the discussion of the asymmetries of the TO and LO Raman line shapes in random alloys, as a possible alternative to the much debated spatial correlation model or to internal/external strain effects.

Percolation model for long wave phonons in random zincblende alloys: from II‐VI's to III‐V's

“…Also, the red-asymmetry of the overall Zn-Te signal, due to the extra mode, increases with x, as expected if the strengths of the extra and dominant Zn-Te TO modes scale as the corresponding Be-Te counterparts, i.e. as x and 1-x, respectively [1]. In fact the Zn-Te TO lineshapes calculated via Eq.…”

“…In particular the dominant Zn-Te peak continuously blue-shifts when x increases, while the extra mode has a fixed frequency at x ≤ 0.19 ('fractal-like' regime, [1]) and blue-shifts for higher x ('normal' regime, [1]), as shown in Fig. 1.…”

“…In the past few years we have developed a percolation approach for the basic understanding of the atypical 1-bond→2-mode behavior that exhibits the stiff Be-VI bond in the Raman spectra of the (Zn,Be)-chalcogenides, that open the class of zincblende mixed crystals with contrast in the bond stiffness [1]. Basically the low-and high-frequency Be-VI Raman modes were assigned as Be-VI vibrations within the randomly-formed hard Be-rich and soft Zn-rich host regions, respectively.…”

An EXAFS study of the structure of the Zn1-xBexSe alloy system

“…The value of S 1 /S 1 v (S 1 v is the slope of the first coordination bond length described with Vegard's law under VCA) was 18% in GaAs x P 1-x [7]. In Zn 1-x Raman measurements and theoretical calculations indicate that, the stiffer Be-Se bond is expected to show a two mode behavior corresponding to its location in the hard and soft regions of the mixed crystal [3] and are ~ 2.24 Å and 2.28 Å for "x" close to the lower threshold of percolation. The fact that good fits to the χ(q) data are obtained by using a single Be-Se bond, indicates that in the present experiment, EXAFs measurements donot explicitly distinguish the presence of the two different bond lengths by the EXAFS data analysis.…”

“…Besides its technological importance, the Raman data of this material system shows a novel and atypical three mode behavior, ie, a single mode for Zn-Se bond and a two-mode behavior for the Be-Se bond [3] in complete contradiction to the Modified Random Element Isodisplacement (MREI) model [3]. This observation has been explained by invoking the fact that the Be atoms, in the mixed crystal, can be in a Be rich domain (Be atoms connected to each other forming a chain and referred to as hard like (HL) region) and a Zn rich domain (isolated Be atom or small clusters surrounded by Zn atoms referred to as soft like (SL) region) [3]. The HL and SL regions in the system result in two different force constants for the Be-Se bonds, which show up as two different modes in the Raman data.…”