Figure 3.1 presents the first-order Raman spectra of Zn1–xCdxSe (x = 0, 0.06, 0.11, 0.18, 0.25, and 0.32) epilayers. The spectra were all obtained at room temperature and ambient pressure under (z x y x y z+ , + ) backscattering geometry. The Raman selection rule for this geometry forbids the TO and allows the emission of LO only. However, a weak TO feature which appears at 205.4 , 203.3, and 201.6 cm-1 for x = 0, 0.06, and 0.11 samples, respectively, can be attributed to a slight deviation from perfect backscattering geometry. Table 3.1 presents the LO and TO phonon frequencies and the linewidth of the LO phonons of all Zn1–xCdxSe samples. As the Cd content increases, the LO and TO phonon frequency decreases.
Asymmetric broadening of the peaks of the LO phonon at high Cd content (x ≧ 0.18) is observed, attributed primarily to the disorder of the alloy [16], and becomes more evident under pressure. Figure 3.2 shows the dependence of LO phonon frequencies and the full width at half maximum (FWHM) on the Cd concentration (x), respectively. The LO phonon frequency falls as x increases, and is accompanied by an increase in FWHM.
The zone-center optical phonon mode of ZB Zn1–xCdxSe was controversial in previous studies. Alonso et al. [17], Avendaño-López et al. [5], and Camacho et al. [18] found that in addition to the LO and TO phonon modes, an impurity mode I was present, and could be attributed to the impurity modes of Zn in CdSe for x close to 1 and to that of Cd in ZnSe for x close to 0. Therefore, they concluded that ZnxCd1–xSe exhibited mixed-mode behavior.
However, Meredith et al. [16]and Li et al. [19] indicated that ZnxCd1–xSe exhibited single mode behavior because they observed no impurity mode and their experimental results were consistent with the mass criterion of one mode behavior [16].Although at room temperature, only LO and TO phonons of Zn1–xCdxSe (x = 0 to 0.32) are observed in the Raman spectra, as shown in Fig. 3.1. However, the low-temperature Raman scattering measurements exhibit the impurity mode I (Fig. 3.3). In particular, at x ≧ 0.25, under the RRS condition, the impurity mode I is clearly present. Moreover, the frequency of the I phonon falls as the Cd content increases, supporting the results given in Ref. 17. Accordingly, the vibration mode of Zn1–xCdxSe is an intermediate.
Figure 3.4 presents the up-stroke pressure-dependent Raman spectra of Zn1–xCdxSe epilayer (x = 0.06) at room temperature. A strong LO phonon and a weak TO feature are observed at ambient pressure, while the TO phonon becomes more intense as the pressure increases. The x = 0 and 0.11 samples exhibit the same behavior because the sample chips deviate from the perfect backscattering geometry when the samples are pressured in the diamond cell. The applied pressure reduces the lattice constant and the crystal volume.
Therefore, the frequencies of the LO and TO phonons shift to higher frequencies, accompanied with decreasing in intensity. The LO phonon disappears and the sample becomes opaque at approximately 13.0 GPa. These facts are evidences of the phase transition from the semiconductor to the metal phase. However, a TO phonon is still observed in the Raman spectra when the semiconductor becomes metallic. It can be ascribed to the fact that
transverse surface lattice vibrations are allowed in both semiconductor and metal, even if the skin depth (or penetration depth) of the metal into which the laser penetrates is merely several tens of angstroms [10,11]. Similar experimental results were obtained from our samples with low Cd content (x ≦ 0.18) and in earlier investigations of nonmagnetic and magnetic II-VI ternary compounds, ZnFeSe [10], ZnSeTe [11], and ZnMnSe [12] crystals.
In our previous investigations, the TO phonon splitting occurred in the up-stroke pressure process and was attributed to the pressure-induced formation of an additional phase.
Despite numerous careful search; however, no TO splitting observed in any of the studied Zn1–xCdxSe samples during the up-stroke process, which were examined several times by reloading them and changing the pressure medium to de-ionized (DI) water. Nevertheless, in the down-stroke process of Zn0.94Cd0.06Se, shown in Fig. 3.5, a split TO phonon mode begins to develop as the pressure is reduced to around 11.5 GPa. It becomes more pronounced at 8.3 GPa. The split TO phonon mode is slightly blueshifted and its intensity falls as the pressure is released. The splitting of the TO mode in the down-stroke process also implies the formation of an additional high-pressure phase. Recent EDXD experiments by Pellicer-Porres et al.
[20,21] demonstrated that ZnSe and ZnSexTe1–x exhibited a cinnabar structure between the ZB and RS structures, becoming apparent only in the down-stroke process. Côté et al. [22] also theoretically calculated the existence of a fourfold-coordinated cinnabar phase between ZB and RS phases, which could be found in the down-stroke process. Accordingly, this work presents the observation of a cinnabar phase in Zn1–xCdxSe by Raman scattering in the down-stroke process.
Table 3.2 lists the pressure-dependent LO phonon frequencies obtained by fitting a quadratic polynomial equation to our measurements,
ωLO = ω0 + ap + bp2, (1) where ω0 is the LO phonon frequency at ambient pressure, p is the pressure in gigapascal, and
a and b are the vibrational pressure coefficients for this mode. The pressure dependence of a
mode frequency ωLO can be defined in terms of the dimensionless Grüneisen parameter (γLO), which is given by [23]
where K0 is the bulk modulus, defined as the inverse of the isothermal volume compressibility (β), and V is the molar volume in cm3/mol. Since the bulk modulus (K0) of Zn1–xCdxSe is unknown, K0 (ZnSe) = 62.4 GPa is used [24]. Table 3.2 presents the γLO values of all samples at ambient pressure. dω/dp = 3.31 cm-1/GPa, and γLO = 0.82 and dω/dp = 4.53 cm-1/GPa, and γLO = 1.18 are obtained for ZnSe and Zn1–xCdxSe (x = 0.32), respectively.
Camacho et al. [18]noted that no pressure-dependent Raman results for bulk CdSe or epilayer had been published. However, Alivisatos et al. [25] found dω/dp = 4.30 cm-1/GPa and γLO = 1.1 for ZB-phase CdSe nanocrystals. The ionization of the lattice can be deduced from γLO
[26], and the ionicity (fi)of CdSe (0.70) is larger than that of ZnSe (0.63) [Ref. 27]. By contrast, higher Cd concentration samples exhibit larger lattice ionization and higher pressure variation on the phonon frequency. In Table 3.2, we can see that the applied pressure tends to reduce the variation of phonon frequency and the lattice ionization due to the second negative term of dω/dp.
Figure 3.6 shows the pressure-dependent Raman spectra of Zn1–xCdxSe epilayer (x = 0.25) at room temperature. The TO phonon is not found and the intensity of LO phonons does not monotonously decline as the pressure increases. Samples with a high Cd content (x = 0.18, and 0.32) exhibit a similar trend, because of the pressure-driven RRS effect, which occurs when the incident laser energy is sufficiently close to the energy of the electronic excitations, described as
laser LO( , ) exc( , )
hν −mhν x p ≈E x p , (3)
where hνlaser is the photon energy of the incident laser (2.41 eV). Eexc( , )x p and
LO( , )
hν x p are the electronic transition energy and the LO phonon energy, respectively, as functions of Cd content (x) and applied pressure (p). The m denotes the overtone order of LO phonons. The exciton energy of Zn1–xCdxSe was controlled by tuning the Cd concentration during the growth, or by manipulating the pressure and temperature of the sample chamber. In this investigation, hνlaser was fixed at 2.41 eV and the exciton energy increased with pressure.
As the external pressure was gradually tuned toward 6.6 GPa, the exciton energy approached the laser energy and an increase in intensity of LO phonon was observed. By further increasing the pressure, the exciton energy exceeds the incident laser energy (at around 7.5 GPa in Fig. 3.6) and begins to move away from the RRS condition, and the intensity of LO phonon decreases monotonously as the pressure increases. At around 11.0 GPa, the LO phonon disappears, revealing the phase transformation from the semiconductor to the metal phase.
Figure 3.7 displays the pressure-dependent PL spectra of the Zn1–xCdxSe epilayer (x = 0.25) at 300 K. At 0.2 GPa, the broad peak at 2.08 eV is attributed to the near band edge (NBE) emission. As the pressure is increased to 2.2 GPa, the NBE emission shifts to higher energy. Further increasing the pressure causes the NBE emission to overlap the LO phonon peaks. The intensity of the LO phonon peaks of the Stokes side increases and the RRS effect occurs. The inset in Fig. 3.7 plots the peak energies of the room-temperature PL spectra before the phase transition as a function of pressure. The PL spectra were excited using an Ar+ 488 nm laser line (2.54 eV) to observe the NBE emission above the 2.41 eV energy of the Ar+ 514.5 nm laser line, which was used to excite phonon Raman scattering. As the pressure further increases to above 6.6 GPa, far from the RRS condition, the intensity of the 1LO phonon decreases abruptly and the PL spectra disappears because the NBE energy exceeds the excitation laser energy. When the pressure reaches the value for the phase transition (11.0
GPa), the LO phonons at both sides of Raman spectra disappears. This result is a strong evidence of the semiconductor-to-metal phase transition. The phase transition pressure of this sample, at which the LO phonon disappeared, was also verified using a 632.8 nm ruby laser as an excitation source, this wavelength is far from the RRS effect.
The pressures at the onset of semiconductor-to-metal phase transition of all Zn1-xCdxSe epilayers discussed herein are listed in Table 3.2. As the Cd concentration increases to 0.32, the phase transition pressure falls from 13.6 to 9.4 GPa. Figure 3.8 plots the decline in the phase transition pressure as a function of x. The solid line represents a quadratic fit, given by Pt (GPa) = 13.6 – 6.8x – 20.3x2. The fall of Pt with x suggests a reduction in structural stability. The structure becomes destabilized as the substituted element content increases; in addition, this relationship has been observed in other ZnSe-based ternary semiconductors. For instance, Yang et al. found that for ZnSe1–xTex epilayers with x up to 0.54, the pressure at the onset of semiconductor-to-metal phase transition decreased from 13.7 to 7.3 GPa. For Zn1-xMnxSe crystals, the structural transition from zincblende (B3) to rocksalt (B1) occurred at about 11.8 and 9.9 GPa for x = 0.07 and 0.24, respectively [11,12].
IV. Conclusions
The pressure-dependent Raman and PL spectra of Zn1–xCdxSe epilayers (0 ≦ x ≦ 0.32) were investigated. Low-temperature experiments reveal that the optical phonons of Zn1–xCdxSe exhibit intrinsic intermediate-mode behavior. The splitting of the TO mode in the down-stroke pressurized process may have been caused by the formation of an additional high-pressure phase. The pressure-dependent LO phonon frequencies are fitted with a quadratic polynomial equation and the Grüneisen parameters are obtained. The pressure-driven RRS effect was observed in samples with a high Cd concentration (x ≧ 0.18).
Both the Stokes and anti-Stokes sides of the LO phonons disappear as the crystal phase
changes from semiconductor-to-metal phase. The pressure at the onset of semiconductor-to-metal phase transition is found to decrease as the Cd content increases.
References
[1] M. A. Haase, J. Qiu, J. M. Depuydt, and H. Cheng, Appl. Phys. Lett. 59, 1272 (1991).
[2] J. Ding, H. Jeon, T. Ishihara, M. Hagerott, A. V. Nurmikko, H. Luo, N. Samarth, and J. K.
Furdyna, Phys. Rev. Lett. 69, 1707 (1992).
[3] W. Faschinger and J. Nürnberger, Appl. Phys. Lett. 77, 187 (2000).
[4] S. -B. Che, I. Nomura, A. Kikuchi, and K. Kishino, Appl. Phys. Lett. 81, 972 (2002).
[5] J. Avendaño-López, F. L. Castillo-Alvarado, A. Escamilla-Esquivel, G. Contreras-Puente, J. Ortiz-López, and O. Zelaya-Angel, Solid State Commun. 100, 33 (1996).
[6] U. Lunz, J. Kuhn, F. Goschenhofer, U. Schüssler, S. Einfeldt, C. R. Becker, and G.
Landwehr, J. Appl. Phys. 80, 6861 (1996).
[7] N. Samarth, H. Luo, J. K. Furdyna, R. G. Alonso, Y. R. Lee, A. K. Ramdas, S. B. Qadri, and N. Otsuka, Appl. Phys. Lett. 56, 1163 (1990).
[8] S. Fujita, Y. Wu, Y. Kawakami, and S. Fujita, J. Appl. Phys. 72, 5233 (1992).
[16] W. Meredith, G. Horsburgh, G. D. Brownlie, K. A. Prior, B. C. Cavenett, W. Rothwell, and A. J. Dann, J. Cryst. Growth 159, 103 (1996).
[17] R. G. Alonso, E. -K. Suh, A. K. Ramdas, N. Samarth, H. Luo, and J. K. Furdyna, Phys.
Rev. B 40, 3720 (1989).
[18] J. Camacho, I. Loa, A. Cantarero, K. Syassen, I. Hernández-Calderón, and L. González, Phys. Status Solidi B 235, 432 (2003).
[19] W. S. Li, Z. X. Shen, D. Z. Shen, and X. W. Fan, J. Appl. Phys. 84, 5198 (1998).
[20] J. Pellicer-Porres, A. Segura, V. Muñoz, J. Zúñiga, J. P. Itié, A. Polian, and P. Munsch, Phys. Rev. B 65, 012109 (2001).
[21] J. Pellicer-Porres, D. Martínez-García, Ch. Ferrer-Roca, A. Segura, V. Muñoz-Sanjosé, J.
P. Itié, A. Polian, and P. Munsch, Phys. Rev. B 71, 035210 (2005).
[22] M. Côté, O. Zakharov, A. Rubio, and M. L. Cohen, Phys. Rev. B 55, 13025 (1997).
[23] M. Blackman and W. B. Daniels, Light Scattering in Solids IV, edited by M. Cardona and G. Güntherodt (Springer, Berlin, 1984), Chap. 8.
[24] S. Ves, K. Strössner, N. E. Christensen, C. K. Kim, and M. Cardona, Solid State Commun.
56, 479 (1985).
[25] A. P. Alivisatos, T. D. Harris, L. E. Brus, and A. Jayaraman, J. Chem. Phys. 89, 5979 (1988).
[26] R. W. Meulenberg and G. F. Strouse, Phys. Rev. B 66, 035317 (2002).
[27] J. C. Phillips, Phys. Rev. Lett. 27, 1197 (1971).
Table 3.1 LO and TO phonon frequencies and FWHM of LO phonon for Zn1–xCdxSe epilayers.
Cd content (x)
LO phonon frequency
(cm-1)
TO phonon frequency
(cm-1)
FWHM of LO phonon (cm-1)
0 252.2 205.4 9.3
0.06 248.4 203.3 10.1 0.11 247.9 201.6 10.9 0.18 244.6 - 14.3 0.25 242.1 - 16.4 0.32 239.1 - 17.7
Table 3.2 Pressure-dependent LO phonon frequencies (ωLO), dωLO/dp, calculated mode Grüneisen parameters (γLO) and phase transition pressures for Zn1-xCdxSe epilayers.
Cd content
FIG. 3.1 Raman spectra of Zn1–xCdxSe epilayers (0 ≦ x ≦ 0.32) at 300 K and ambient pressure.
FIG. 3.2 Dependence of LO phonon frequencies (open circle) and FWHM (solid triangle) on Cd concentration (x).
FIG. 3.3 Raman spectra of Zn1–xCdxSe epilayers (0 ≦ x ≦ 0.32) at 100 K and ambient pressure. The impurity (I) modes, indicated by black arrows, appear at low temperature. The LO and TO phonons of the GaAs substrate are also labeled.
FIG. 3.4 Up-stroke pressure-dependent Raman spectra of Zn1–xCdxSe (x = 0.06) at room temperature. The LO phonon disappears at around 13.0 GPa, revealing a phase change from semiconductor to metal.
FIG. 3.5 Down-stroke pressure-dependent Raman spectra of Zn1–xCdxSe (x = 0.06) at room temperature. The splitting of the TO phonon, labeled by black arrows, was observed clearly as the pressure was released.
FIG. 3.6 Up-stroke pressure-dependent Raman spectra of Zn1–xCdxSe (x = 0.25) at room temperature. The pressure-driven resonant Raman scattering effect occurred as the pressure was increased. The LO phonon was found to disappear at about 11.0 GPa.
FIG. 3.7 Pressure-dependent photoluminescence spectra of Zn1–xCdxSe (x = 0.25) at 300 K.
The 514.5 nm (2.41 eV) Ar+ laser was fixed as an excitation source. The black dashed arrow at 2.41 eV labels the energy of the excitation laser, whereas the Stokes and anti-Stokes Raman spectra occur at the lower and higher energy sides of the laser, respectively. The inset plots the pressure dependence of PL energies, and the dashed line indicates the energy of the excitation laser.
FIG. 3.8 Cd concentration (x)-dependent phase transition (semiconductor-to-metal) pressure of Zn1–xCdxSe epilayers. The solid curve represents a quadratic polynomial fit.