Piezoreflectance and - 新穎光電半導體材料、奈米微細結構及其元件構造之光學特性研究

photoreflectance study of annealing effects on GaAs

_0.916

Sb

_0.084

and GaAs

_0.906

Sb

_0.075

N

_0.019

films

on GaAs grown by gas-source molecular beam epitaxy

H. P. Hsu¹, Y. N. Huang¹, Y. S. Huang^{*, 1}, Y. T. Lin², T. C. Ma², H. H. Lin², K. K. Tiong³, P. Sitarek⁴, and J. Misiewicz⁴

1 Department of Electronic Engineering, National Taiwan University of Science and Technology, Taipei 106, Taiwan

2 Department of Electrical Engineering, National Taiwan University, Taipei 106, Taiwan

3 Department of Electrical Engineering, National Taiwan Ocean University, Keelung 202, Taiwan

4 Institute of Physics, Wrocław University of Technology, Wybrzeże Wyspiańskiego 27, Wrocław 50-370, Poland Received 3 July 2008, revised 19 November 2008, accepted 24 November 2008

Published online 25 March 2009 PACS 78.20.Ci, 78.66.Fd, 81.15.Hi

*Corresponding author: e-mail [email protected], Phone: +886 2 27376385, Fax: +886 2 27376424

1 Introduction Dilute nitride III–V semiconductors alloys such as GaNAs and GaInNAs have attracted much attention in recent past due to their unique physical proper-ties and potential applications for achieving the telecom-munication wavelengths of 1.3 – 1.5 μm with GaAs-based materials [1 – 3]. It is known that a dilute amount of nitro-gen offers a unique feature of reducing simultaneously the band gap and the lattice parameter of a given III – V alloy.

Nevertheless, N-containing materials are seriously limited by the degradation of their optical properties that is gener-ally observed as the N content is increased. Therefore, it remains a challenge to get a reasonable material quality for N concentrations higher than 1%. Consequently, the inter-est of N-containing materials for device applications has restricted to alloys of N composition less than 2 – 3%. In order to achieve the emission wavelength beyond 1.3 μm,

the amount of N incorporated into the Ga(In)As alloy must be higher and can result in severe degradation of crystal quality of such materials [4]. Recently, Ungaro et al. [5]

proposed a new alternative alloy, namely, GaAsSbN which offers capability to achieve longer wavelength. Like GaInAs, GaAsSb is a ternary alloy where the introduction of large size Sb substitutional atoms results in a band gap narrowing and increase of lattice parameter. When a small amount of N is incorporated into GaAsSb, as in other dilute III – V nitrides, a band gap reduction and a lattice contraction are observed. Their works demonstrate that GaAsSbN grown on GaAs substrate can be used to pre-pare optical devices emitting in 1.3 – 1.5 μm region at room temperature [6 – 8]. Subsequent studies related to GaAsSbN applications in long wavelength optical devices have also been reported [9 – 11]. In spite of the potential Thermal annealing effects of GaAs_0.916Sb_0.084 and

GaAs_0.906Sb_0.075N_0.019 films grown on GaAs substrates by gas-source molecular beam epitaxy have been characterized by piezoreflectance (PzR) and photoreflectance (PR). By a com-parison of relative intensity of PzR and PR spectra, the identi-fication of conduction to heavy-hole (HH) band and conduc-tion to light-hole (LH) band transiconduc-tions originated from the strained induced valence band splitting have been achieved. The near band edge transition energies are

blue-shifted, and the splitting of HH and LH bands is reduced after thermal annealing treatment. The annealing effects of GaAs_0.906Sb_0.075N_0.019 are found to be more pronounced than that of GaAs_0.916Sb_0.084. The temperature dependences of near band edge transition energies are analyzed using Varshni and Bose – Einstein expressions in the temperature range from 15 K to 300 K. The parameters that describe the temperature variations of the near band edge transition energies are evalu-ated and discussed.

Phys. Status Solidi A 206, No. 5 (2009) 831

applications, very little work has been done on the optical properties related to thermal annealing effects in Sb/In containing materials. For example, the optical transition features originated from the valence band splitting and their temperature dependent near band edge interband tran-sitions properties is only little known. Hence, further study on the thermal annealing effects of GaAsSb and GaAsSbN alloys is not only interesting but also necessary and impor-tant.

In this study, we present piezoreflectance (PzR) and photoreflectance (PR) study of thermal annealing effects on strained GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019 films grown on GaAs substrates by gas-source molecular beam epitaxial (MBE). By a comparison of the relative intensity of PzR and PR spectra, the identification of conduction to heavy-hole (HH) band and conduction to light-hole (LH) band transitions originated from the strained induced va-lence band splitting has been achieved. The annealing effects of the near band edge transitions are studied. The temperature dependent behaviors of these transition ener-gies in the range from 15 – 300 K are also observed. The parameters that describe the temperature variations of the near band edge transition energies are evaluated and dis-cussed.

2 Experimental detail The GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019 layers were grown on (100) semi-insulating GaAs substrate via a VG-V80 gas-source MBE system. An EPI Sb cracking cell was used to provide mixed dimmer and monomer Sb beam. As2 beam was from a gas cell with a cracking temperature of 1000 °C. The precursor was AsH3. Ga flux, calibrated using an ion gauge to keep the growth rate at 1 μm/h, was provide by an EPI uni-bulb RF plasma K-cell operating at a radio frequency of 13.56 MHz. A PBN shutter was placed in front of the K-cell to reduce the ionized species. The thickness of the samples in this study is 1 μm with growth temperature at 490 °C. The compositions of the GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019 were quantified by electron probe X-ray microanalyzer (EPMA) with GaAs, GaN, and GaSb as standards for ZAF (atomic number Z, absorption A, and fluorescence F) correction. The GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019 were ex situ annealed at temperature 800 °C for 300 s in N2 ambient during the thermal anneal-ing process.

For the PzR measurements, the thin samples were glued on a 0.15 cm thick lead zirconate titanate piezoelec-tric transducer driven by a 200Vrms sinusoidal source at 200 Hz. The alternating expansion and contraction of the transducer subjected the sample to an alternating strain with a typical rms Δl/l value of ~10^–5. PR measurements were achieved using an internally modulated 670 nm laser diode served as the pumping beam. The radiation from a 150 W tungsten-halogen lamp filtered by a 0.25 m mono-chromator provided the monochromatic light. The reflected light was detected by an InGaAs photodetector. The dc output of photodetector was maintained constant by a

servo mechanism of variable neutral density filter. A dual-phase lock-in amplifier was used to measure the detected signals. Multiple scans over a given photon energy range was programmed until a desired signal-to-noise level has been attained with computer controlled data acquisi- tion procedure. Detailed PzR and PR configurations have been described elsewhere [12, 13]. For temperature de-pendent measurement, a closed-cycle cryogenic refrigera-tor equipped with a digital thermometer controller was used for the low temperature measurements with a tem-perature stability of 0.5 K or better.

3 Results and discussion Figure 1 show PzR and PR spectra for the as-grown and annealed GaAs0.916Sb0.084

(Fig. 1(a)) and GaAs0.906Sb0.075N0.019 (Fig. 1(b)) samples at room temperature. The dotted lines are the experimental data and full curves are the least-squares fits to the deriva-tive Lorentzian line-shape function of the form [14, 15]

1 of the transitions, and the value of n depends on the origin of the transitions. For derivative functional form, n = 2 is appropriate for the bound states such as excitons.

As shown in Fig. 1, the lineshape fits for both PzR and PR spectra clearly show two structures (indicate with arrows) near the band edge of GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019. To identify the physical origin of the doublets, we make a spectral comparison of the PzR and PR measurements.

For the PzR measurement under [001]-symmetry co-planar stress, the intensity ratio between conduction to light-hole (LH) band transitions and the conduction to heavy-hole (HH) band has been shown to follow the rela-tion [16]

The variable S refers to the modulating stress applied to the sample. The parameters a and b represent the hydrostatic and the shear deformation potentials, respectively, and λ = –2S12/(S11 + S12), where Sij is the elastic compliance constant. We have used a number of relevant parameters from literature to deduce reasonable estimates for that of GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019. The deformation potentials of the ternary/quaternary materials were ob-tained by linear interpolation of values of the end-point semiconductors GaAs, GaSb and GaN: for GaAs, using a = –9.8 eV [17], b = –2.0 eV [17], S11 = 1.16 × 10^–6 bar^–1

832 H. P. Hsu et al.: Annealing effects on GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019 films on GaAs

physica

p s s

^status

Figure 1 PR and PzR spectra for the as-grown and annealed (a) GaAs0.916Sb0.084 and (b) GaAs0.906Sb0.075N0.019 samples at 300 K.

The obtained values of the transition energies are indicated by the arrows.

the numerical values of a = –9.53/–9.52 eV, b = –2.00/

–2.00 eV, S11 = 1.20/1.18 × 10^–6 bar^–1, and S12 = –0.38/

–0.37 × 10^–6 bar^–1, are deduced. The ratio of KPzR for GaAs0.916Sb0.084/GaAs0.906Sb0.075N0.019 is then evaluated to be about 2.22/2.24 from Eq. (2). This result indicates that the transition of light hole is more sensitive than that of heavy hole under the [001]-symmetry coplanar piezo-modulation. To further confirm this observation we have also studied the PR spectrum of the same sample. The rela-tive intensity of the LH and HH transitions for the PR measurements is insensitive to strain and has been used as a reference. The ratio (dELH/dS)/(dEHH/dS) of the ampli-tudes of the features is KPR in PR and KPzR in PzR; clearly, KPzR≠ KPR. In PR, which is a nonselective modulation technique, the modulation parameter is the same for all the transitions, so that the ratio K = KPzR/KPR directly gives the ratio of the piezomodulation parameters of the LH and HH transitions. From the PzR measurements for both of GaAsSb and GaAsSbN samples, the enlarged feature appeared at the higher energy side with respect to the HH feature indicates the presence of a compressive-type stress in the samples. The KPzR is determined to be 2.1(2.0)/2.2(2.3) for as-grown (annealed) GaAsSb/as-grown (annealed) GaAsSbN samples. The experimentally deduced values of KPzR agreed reasonably well with that of the theoretical calculation.

The identified transition features are denoted as HH and LH and indicated by vertical arrows in Fig. 1. As can be seen in Fig. 1, for the GaAs0.916Sb0.084 film the annealing effect is weak: the transition energy of HH remains un-changed and the separation between HH and LH transi-tions is slightly reduced (from 12 meV to 9 meV). For GaAs0.906Sb0.075N0.019, the annealing effect is more pro-nounced: a 13 meV blue shift of HH transition with the separation between HH and LH reduces from 34 meV to 21 meV, and the broadening parameter decreases. The re-sults show strain relaxation in GaAsSb and GaAsSbN lay-ers and improvement in compositional homogeneity after thermal annealing treatment.

The sensitivity of the GaAsSbN alloy to thermal an-nealing is remarkable in contrast with the behaviour of the GaAsSb. Similarly, in the case of GaInNAs material, strong blue shift was reported after annealing [22 – 24].

Since this phenomenon is essentially observed in the qua-ternary materials, short-distance interactions between the constituents of the quaternary alloys are likely to be in-volved. For GaInNAs, it was suggested that annealing favours the formation of N-In bonds, replacing as-grown N-Ga bonds [22, 25, 26]. This change in the distribution of N bonds was predicted to reduce the band gap bowing in GaInNAs, and can therefore explain a significant blue-shift induced by thermal treatments. Short range atomic rearrangements are indeed also expected in GaAsSbN to

Figure 2 Experimental PR spectra (dotted curves) of GaAs_0.916Sb_0.084 (a) as-grown and (b) annealed at 800 °C samples at several temperatures between 15 K and 300 K. The full lines are least-squares fits to Eq. (1). The obtained values of the transi-tion energies are indicated by the arrows.

Phys. Status Solidi A 206, No. 5 (2009) 833

Figure 3 Experimental PR spectra (dotted curves) of GaAs0.906Sb0.075N0.019 (a) as-grown and (b) annealed at 800 °C samples at several temperatures between 15 K and 300 K. The full lines are least-squares fits to Eq. (1). The obtained values of the transition energies are indicated by the arrows.

minimize the local strain around N or Sb sites. However, GaAsSbN has a notable difference with GaInNAs, since substitutional N atoms can only bond to Ga. Therefore, the degree of influence of As or Sb environment around N at-oms is certainly of second order as group V atat-oms are sec-ond nearest neighbours of N. As a consequence, it can be speculated that the electronic structure of GaAsSbN would be less sensitive to thermal treatment than GaInNAs. This speculation is in contradiction with the present experimen-tal results. A different origin such as a band gap shift due to the chemical change of the second-nearest-neighbouring atoms to the N atoms can be considered to explain the presently observed blue-shift.

Figures 2 and 3 show the temperature dependent PR spectra of GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019, respectively. Displayed by dotted curves in Figs. 2(a)/3(a) and 2(b)/3(b) are, respectively, the experi-mental PR spectra of as-grown and annealed GaAs0.916Sb0.084/GaAs0.906Sb0.075N0.019 samples at several temperatures between 15 K and 300 K. The solid lines are the fitted spectral data to Eq. (1) with n = 2, which yield transition energies indicated by arrows. As the general property of most semiconductors, when the measuring temperature is increased, the HH and LH transitions in the PR spectra exhibit an energy red-shift characteristic.

The temperature variations of the experimental PR val-ues for HH and LH transitions with representative error

Figure 4 Temperature variations of the experimental PR values for HH and LH transition with representative error bars for GaAs0.916Sb0.084 (a) as-grown and (b) annealed 800 °C samples as circles and open and diamonds, respectively. The full curves are least-squares fits to Eq. (3) and the dashed lines are least-squares fits to Eq. (4).

bars for as-grown and annealed GaAs0.916Sb0.084/ GaAs_0.906Sb_0.075N_0.019are depicted in Figs. 4(a) – (b)/5(a) – (b) as open-diamonds and circles. The full curves are the tem-perature dependence of the HH and LH near band edge transition energies fitted by Varshni semi-empirical rela-tionship [27]

( ) (0) ⁱ ,

i i

E T E T

T α

= -β

+ (3)

where Ei (0) are the conduction-to- heavy-hole and light-hole bands transition energies at 0 K. The constants αi is related to the electron (exciton)-average phonon interaction strength and βi is closely related to the Debye temperature.

The temperature dependence of near band edge transi-tion energies can also be described by a Bose – Einstein ex-pression (dashed lines) [28, 29]

B B

( ) (0) 2 ,

[exp ( / ) 1]

i i i

E T E a

= - Θ

- (4)

where Ei (0) are the transition energies for the conduction to heavy-hole and conduction to light-hole bands transition energies at 0 K, aiB represents the strength of the electron (exciton)-average phonon interaction, and ΘiB corresponds to the average phonon temperature. The values obtained for E0(0), αi, and βi from Varshni semi-empirical relation

834 H. P. Hsu et al.: Annealing effects on GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019 films on GaAs

physica

p s s

^status

Table 1 Values of the Varshni – and Bose – Einstein type fitting parameters, which describe the temperature dependence of near band edge transition energies for as-grown and annealed GaAs0.916Sb0.084/GaAs0.909Sb0.07N0.021 alloys. The parameters of GaAs0.90Sb0.08N0.02, GaAs0.9703N0.0297, GaAsSb and GaAs are also included for comparison.

samples feature Ei (0) αi βi a iB ΘiB –dEi/dT

(eV) (meV/K) (K) (meV) (K) –(meV/K)

GaAs_0.916Sb_0.084^a as-grown

a This study (photoreflectance), ^b Ref. [30] (photoluminescence), ^c Ref. [31] (absorption), ^d Ref. [32] (absorption),^e Ref.[33] (absorption).

ship, aiB and ΘiB from Bose−Einstein-type expression are listed in Table 1. For comparison, the parameters for band gap energies of GaAs0.90Sb0.08N0.02 [30], GaAs0.9703N0.0297

[31], GaAsSb [32], and GaAs [33] are also listed in Ta-ble 1.

Figure 5 Temperature variations of the experimental PR values for HH and LH transition with representative error bars for GaAs0.906Sb0.075N0.019 (a) as-grown and (b) annealed 800 °C sam-ples as circles and open diamonds, respectively. The full curves are squares fits to Eq. (3) and the dashed lines are least-squares fits to Eq. (4).

The parameter αi of Eq. (3) can be related to aiB and ΘiB in Eq. (4) by taking the high-temperature limit of both expressions. This yields αi = 2aiB/ΘiB. Comparison of the numbers presented in Table 1 show that this relation is fairly satisfied. From Eq. (4), it is straightforward to show the high temperature limit of the slope of Ei(T) vs. T curve approaches a value of –2aiB/ΘiB. The calculated value of –2aiB/ΘiB for conduction to heavy-hole/light-hole near band edge transition energies are –0.42/–0.40 and –0.42/

–0.40 (–0.33/–0.33 and –0.40/–0.39) meV/K for as-grown and annealed GaAs_0.916Sb_0.084 (GaAs_0.906Sb_0.075N_0.019) sam-ples, respectively, which agrees well with the values of [dEHH(LH)/dT] = –0.39/–0.37 and –0.38/–0.37 (–0.28/–0.28 and –0.41/–0.40) meV/K as obtained from the linear trapolation of the high temperature (150 – 300 K) PR ex-perimental data.

As shown in Table 1, the parameters that describe the temperature variations of near band edge transition ener-gies of GaAs0.916Sb0.084 and GaAs0.909Sb0.07N0.021 alloys are quite similar to that reported by Bian et al. [30]. The values of αi and dEi/dT for GaAsSbN are similar to that of GaAs0.9703N0.0297 [31] and GaAsSb [32] films and smaller than that of GaAs which was attributed to the temperature induced shift of the band edge transition energies and is substantially reduced by the presence of nitrogen/antimony.

4 Summary We have studied the thermal annealing effects of GaAs0.916Sb0.084 and GaAs0.906Sb0.075N0.019 films grown on GaAs substrate by PzR and PR techniques. A comparison of the PzR and PR spectra has led to the identi-fication of HH and LH transitions originated from the va-lence band splitting of the samples. Thermal annealing treatment results in a reduction of the strain induced sepa-ration of HH and LH valence splitting and an improvement in compositional homogeneity. The annealing effects of GaAs0.906Sb0.075N0.019 are found to be more pronounced than

Phys. Status Solidi A 206, No. 5 (2009) 835

that of GaAs0.916Sb0.084. The temperature dependences of these near band edge transition energies are analyzed using the Varshni expression and an expression containing the Bose – Einstein occupation factor for phonons. The pa-rameters that describe the temperature dependences of GaAsSbN alloys are similar to that of GaAsN and GaAsSb and smaller than that of GaAs. This has been attributed to the incorporation of nitrogen/antimony into the GaAsSbN alloys.

Acknowledgements The authors acknowledge the sup-ports of National Science Council of Taiwan under Project No. NSC 96-2221-E-011-030 and the project-based personnel exchange program between the NSC and PAS: Project No.

NSC96-2911-I-011-003. Electron. Lett. 35, 1246 (1999).

[6] J. C. Harmand, G. Ungaro, L. Largeau, and G. Le Roux, Appl. Phys. Lett. 77, 2482 (2000).

[7] J. C. Harmand, G. Ungaro, J. Ramos, E. V. K. Rao, G. Saint-Girons, R. Teissier, G. Le Roux, L. Largeau, and G. Patriarche, J. Cryst. Growth 227/228, 553 (2001).

[8] J. C. Harmand, A. Caliman, E. V. K. Rao, L. Largeau, J. Ramos, R. Teissier, L. Travers, G. Ungaro, B. Theys, and I. F. L. Dias, Semicond. Sci. Technol. 17, 778 (2002).

[9] K. H. Tan, S. F. Yoon, W. K. Loke, S. Wicaksono, K. L.

Lew, A. Stöhr, O. Ecin, A. Poloczek, A. Malcoci, and D. Jäger, Appl. Phys. Lett. 90, 183515 (2007).

[10] S. Wicaksono, S. F. Yoon, W. K. Loke, K. H. Tan, K. L.

Lew, M. Zegaoui, J. P. Vilcot, D. Decoster, and J. Chazelas, J. Appl. Phys. 102, 044505 (2007).

[15] F. H. Pollak, in: Handbook on Semiconductors, edited by M. Balkanski, Vol. 2 (North Holland, Amsterdam, 1994), p. 527.

[16] H. Mathieu, J. Allègre, and B. Gil, Phys. Rev. B 43, 2218 (1991).

[17] H. Qiang, F. H. Pollak, and G. Hickman, Solid State Com-mun. 76, 1087 (1990).

[21] S. P. Łepkowski and J. A. Majewski, Solid State Commun.

131, 763 (2004).

[22] P. J. Klar, H. Grüning, J. Koch, S. Schäfer, K. Volz, W. Stolz, W. Heimbrodt, A. M. Kamal Saadi, A. Lindsay, and E. P. O’Reilly, Phys. Rev. B 64, 121203 (2001).

[23] R. Kudrawiec, E.-M. Pavelescu, J. Wagner, G. Sęk, J. Mi-siewicz, M. Dumitrescu, J. Konttinen, A. Gheorghiu, and M.

Pessa, J. Appl. Phys. 96, 2576 (2004).

[24] R. Kudrawiec, E.-M. Pavelescu, J. Andrzejewski, J. Misie-wicz, A. Gheorghiu, T. Jouhti, and M. Pessa, J. Appl. Phys.

96, 2909 (2004).

[25] S. Kurtz, J. Webb, L. Gedvilas, D. Friedman, J. Geisz, J. Ol-son, R. King, D. Joslin, and N. Karam, Appl. Phys. Lett. 78, 748 (2001).

[26] E.-M. Pavelescu, J. Wagner, H. P. Komsa, T. T. Rantala, M. Dumitrescu, and M. Pessa, J. Appl. Phys. 98, 083524 (2005).

[27] Y. P. Varshni, Physica (Amsterdam) 34, 149 (1967).

[28] P. Lautenschläger, M. Garriga, S. Logothetidis, and M. Car-dona, Phys. Rev. B 35, 9174 (1987).

[29] P. Lautenschläger, M. Garriga, and M. Cardona, Phys. Rev.

B 36, 4813 (1987).

[30] L. F. Bian, D. S. Jiang, P. H. Tan, S. L. Lu, B. Q. Sun, L. H.

Li, and J. C. Harmand, Solid State Commun. 132, 707 (2004).

[31] K. Uesugi, I. Suemune, T. Hasegawa, T. Akutagawa, and T. Nakamura, Appl. Phys. Lett. 76, 1285 (2000).

在文檔中新穎光電半導體材料、奈米微細結構及其元件構造之光學特性研究 (頁 50-56)