Advertisement

Temperature-Insensitive Band-Gap III-V Semiconductors: Tl-III-V and III-V-Bi

  • Hajime Asahi
Part of the Springer Handbooks book series (SPRINGERHAND)

Abstract

Thallium-containing III-V (Tl-III-V) and bismuth-containing III-V (III-V-Bi) alloy semiconductors were first proposed as novel functional semiconductors. They are alloys consisting of semiconductors (III-V) and semimetals (Tl-V, III-Bi), and are important materials for the fabrication of temperature-insensitive lasing wavelength laser diodes as well as long wavelength infrared (LWIR) optical devices. However, the growth conditions for alloys containing Tl and Bi are very strict and the growth windows are narrow. In this chapter, the expected properties of these semiconductors and the experimental results for the growth, characterization, and device applications are described.

Wavelength division multiplexing (WDM ) technology is very important for an optical fiber communication system for increasing transport capacity and obtaining flexible network management. However, one of the problems encountered when using InGaAsP/InP laser diodes (LD s) in the WDM system is that the lasing wavelength fluctuates with ambient temperature variation mainly due to the temperature dependence of the band-gap energy. Therefore, LDs in WDM systems must be equipped with Peltier elements to stabilize the LD temperature.

To solve this problem, Oe et al. proposed the use of temperature-insensitive band-gap semiconductors as an active layer of LDs [23.1]. Such a temperature-insensitive band-gap property was observed in the II-VI semiconductor, HgCdTe , as shown in Fig. 23.1 [23.2]. HgCdTe is usually used for the long wavelength infrared (LWIR ) detectors. CdTe is a semiconductor and HgTe is a semimetal. With the increase of HgTe composition, their band-gap energy decreases. At the CdTe-rich side of HgCdTe, the band-gap energy decreases with the increase of temperature, while at the HgTe-rich side of HgCdTe, the band-gap energy increases with the increase of temperature. Therefore, at a certain HgTe composition (0.48) region, the temperature-insensitive behavior is expected [23.3]. Similar behavior was also observed for HgCdSe, where the very small temperature dependence of refractive index was also confirmed [23.4].
Fig. 23.1

Band-gap energy versus alloy composition for HgCdTe as a function of temperature. (After [23.2])

The origin of this temperature-insensitive band-gap energy is considered as follows. Figure 23.2 shows the alloy composition dependence of the band structure for HgCdTe alloys [23.5]. At the CdTe side of HgCdTe, the Γ6 band behaves as a conduction band and the Γ8 band behaves as a valence band. Also in this composition region, the band-gap energy decreases with increasing temperature similar to usual semiconductors. With increasing HgTe composition, the Γ6 band goes down and the band-gap energy decreases. By further increasing HgTe composition, the Γ6 and Γ8 bands cross and the band inversion occurs. After this band inversion, the Γ6 band becomes a valence band and the Γ8 light hole band forms a conduction band. The Γ8 heavy hole band still behaves as a valence band and HgCdTe becomes a semimetal. Therefore, the energy difference between the two bands increases with increasing temperature so that the band-gap energy for HgCdTe semiconductors at the HgTe side composition increases with increasing temperature. At some intermediate alloy composition, we can observe the temperature-insensitive band-gap energy as shown in Fig. 23.1 [23.2].
Fig. 23.2

Alloy composition dependence of the band structure for HgCdTe alloys. (After [23.5])

From the above-mentioned discussion, the alloy semiconductor consisting of a semiconductor and a semimetal is expected to show the temperature-insensitive band-gap behavior. Oe et al. suggested GaAsBi and InGaAsBi alloys as new semiconductors showing temperature-insensitive band-gap energy for possible application to the temperature-stable lasing wavelength LDs [23.1]. Asahi et al. also proposed TlInGaAs and TlInGaP as new semiconductors showing temperature-insensitive band-gap energy [23.6, 23.7]. Similar to HgCdTe, it is also expected that these alloy semiconductors can be used for the LWIR detectors.

This chapter describes the expected properties of the Tl-III-V and III-V-Bi alloy semiconductors and the present status of their growth, characterization, and device applications.

23.1 Tl-III-V Alloy Semiconductors

Alloy semiconductor HgCdTe is widely used for LWIR focal-plane arrays (FPA ). However, HgCdTe is a weakly bonded II-VI compound with material growth and processing problems dominated by native defects that limit FRA performance and yield [23.8]. Van Schilfgaarde et al. proposed thallium (TI)-based III-V ternary alloy semiconductors, TlInSb, TlInAs, and TlInP as the possible candidates for future LWIR detector materials [23.10, 23.11, 23.8, 23.9]. Their theoretical calculations of TlSb, TlAs, and TlP showed that they have negative band gaps similar to HgTe (semimetal). Therefore, it is expected that TlInSb, TlInAs, and TlInP should exhibit LWIR band gaps. Furthermore, they are III-V materials and are compatible with the conventional III-V device fabrication technologies. Therefore, it was expected that these Tl-based alloys could overcome the difficulties encountered in the widely used LWIR semiconductor HgCdTe.

Interesting extensions of TlInSb, TlInAs, and TlInP alloys are quaternary alloys TlInGaP and TlInGaAs [23.12, 23.6, 23.7]. These quaternary alloys can be lattice matched to GaAs, InP, and InAs substrates. Lattice matching is very important to obtain good quality crystals. Furthermore, they are expected to have temperature-insensitive band-gap energies at some alloy compositions and they can be applied to the fabrication of temperature-insensitive lasing wavelength LDs [23.6]. Further extension is a quinary alloy semiconductor TlInGaAsN , which consists of a semiconductor GaInNAs and a semimetal TlAs and is also expected to be applicable to fabricate the temperature-stable lasing wavelength and temperature-stable threshold LDs [23.13] as discussed below.

23.1.1 Expected Properties of Tl-Based III-V Alloys

Figure 23.3 shows the expected relationships between the band-gap energy (Eg) and the lattice constant for Tl-based III-V alloys [23.12, 23.13, 23.6] based on the theoretical results for TlInSb, TlInAs, and TlInP by van Schilfgaarde et al. [23.11, 23.8]. The lines for TlGaAs and TlGaP were predicted by Asahi et al. [23.6]. The data for other III-V compound materials are also plotted in the figure for reference.
Fig. 23.3

Relationship for the band-gap energy versus alloy composition for Tl-III-V alloys as well as usual III-V semiconductors. (After [23.12, 23.13, 23.6])

Van Schilfgaarde et al. [23.10, 23.11, 23.8, 23.9] have theoretically calculated the lattice constant, cohesive energy, elastic constant, band-gap energy, effective mass, and electron mobility by the local density approximation first principle theory. They obtained the following results: (1) TlAs and TlP are formed in the zincblende structure, while TlSb is formed in the CsCl structure; (2) the lattice constants of TlSb (0.659 nm), TlAs (0.618 nm), and TlP (0.596 nm) are only 1.5–2% larger than those of InSb (0.648 nm), InAs (0.606 nm), and InP (0.587 nm), respectively; (3) the cohesive energies are about twice that of HgTe; therefore, bond strengths are higher; (4) TlSb, TlAs, and TlP are semimetals with negative band gaps of −1.60, −1.34, and −0.27 eV, respectively, similar to HgTe (−0.3 eV); (5) TlInSb, TlInAs, and TlInP exhibit a direct band gap of 0.1 eV (12.4 μm) at the Tl compositions of 0.09, 0.15, and 0.67, respectively. These properties are promising for application to LWIR detectors.

Thermodynamical properties of TlInP and TlInAs were theoretically estimated [23.14]. It was predicted that only low concentrations of Tl can be achieved in TlInP (< 5% at 350C) using gas-source molecular beam epitaxy (GSMBE ). Although much less than 1% Tl was expected to be soluble in TlInAs, the error estimates in calculation indicated that obtaining 15% Tl is possible.

Quaternary alloys of TlInGaP and TlInGaAs proposed by Asahi et al. [23.6, 23.7] can be easily lattice matched to InP, GaAs, and InAs substrates as shown in Fig. 23.3. TlInGaP lattice matched to InP can cover the band-gap energy from 1.24 eV (0.92 μm at 0 K, InP) to less than 0.1 eV (12.4 μm, TlGaP), and over the whole lattice-matched composition the direct band-gap is expected to form. The addition of Al and As can further extend the energy (wavelength) range. TlInGaAs lattice matched to InP can cover from 0.75 eV (1.65 μm, InGaAs) to about 0 eV (TlGaAs). These quaternary alloys can be applied to LWIR detectors and emitters. They are expected to exhibit the temperature-insensitive band-gap energy behavior at some alloy compositions similar to HgCdTe [23.6]. This is because these semiconductors are alloys consisting of semiconductors (InGaP, InGaAs) and semimetals (TlP, TlAs). These temperature-insensitive band-gap energy characteristics are very promising for the fabrication of LDs whose wavelength does not change with or is insensitive to ambient temperature variation. The band lineup of lattice-matched TlInGaP/InP heterostructures was suggested to be type-I and the conduction band discontinuities are larger than the valence band offsets [23.6], also proved by the theoretical calculation [23.15].

Quinary alloy semiconductor TlInGaAsN was proposed by Asahi et al. to fabricate the temperature-stable lasing wavelength and temperature-stable threshold current TlInGaAsN/AlGaAs LDs [23.13]. TlInGaAsN is an alloy of a semiconductor InGaNAs and a semimetal TlAs as shown in Fig. 23.3. InGaNAs/AlGaAs heterostructures are widely studied to fabricate temperature-stable threshold current LDs in the wavelength range of 1 μm because of their large conduction band discontinuity [23.16]. However, there exists a long wavelength limit of 1.3 μm in this material system. This is because of the degradation of the crystal quality as well as the reduction of valence band offset with increasing N composition [23.17]. To overcome this problem, new alloy semiconductor TlInGaAsN/AlGaAs heterostructures were proposed [23.13].

As can be clearly seen in Fig. 23.3, much lower band-gap energy, therefore, longer wavelength, can be easily obtained for this new semiconductor compared with GaInNAs. TlInGaAsN/AlGaAs heterostructures also have a large conduction band discontinuity similar to GaInNAs/AlGaAs ones. Furthermore, TlInGaAsN is expected to have temperature-insensitive band-gap energy because it is an alloy of a semiconductor and a semimetal, similar to TlInGaAs. Therefore, the fabrication of temperature-stable lasing wavelength and temperature-stable threshold current LDs is expected by using TlInGaAsN/AlGaAs heterostructures.

23.1.2 Growth of Tl-III-V Alloys

The growth of Tl-based III-V alloys has been attempted by metal-organic vapor phase epitaxy (MOVPE ) [23.18, 23.19, 23.20], gas source MBE (GSMBE) [23.21, 23.22, 23.23, 23.24, 23.25, 23.26, 23.6], and solid source MBE (SSMBE ) [23.27, 23.28].

In the MOVPE growth of TlInSb and TlInP, cyclopentadiethylthallium (CPTl) and trimethylindium (TMIn) were used as Tl and In precursors, and trimethylantimony (TMSb) or tertiarybutyldemethylantimony (TBDMSb) and phosphine (PH3) as Sb and P precursors, respectively [23.18, 23.19, 23.20, 23.29]. The vapor pressure of CPTl is as low as about 3 Torr even at 120C. Therefore, high bubbler temperature and heated valves, and stainless delivery lines are needed.

In GSMBE or SSMBE growth, elemental Tl, In, and Ga were used as group III sources [23.21, 23.22, 23.23, 23.24, 23.25, 23.26, 23.27, 23.28, 23.6]. Thermally cracked arsine (AsH3), phosphine (PH3), or elemental As and Sb were used as group V sources. Tl vaporizes from the melt state, which means that the controllability of the Tl flux is much better than that of Sb. The problem in using Tl metal is its ease of oxidation. The toxicity of Tl is known to be higher than that of As, so extreme care must be taken during treatment [23.6]. Tl-condensed surfaces should not be touched with bare hands and Tl flakes or vapors should not be inhaled: legal intake is documented to be at 600 mg. From this toxicity point of view, growth by MBE is much safer than that by MOPVE. In MBE, the solid Tl is used and is stored in the growth chamber, whereas in MOVPE, the Tl is supplied and vented as a form of metal organic (MO ) gas. It is considered that there is no essential difference between GSMBE and SSMBE both in safety and growth mechanism, though growth conditions may be slightly different.

Growth of Tl-III-Sb Ternary Alloys

The growth of TlInSb was first conducted by Choi et al. [23.18]. They grew TlInSb layers with good surface morphology on GaAs and InSb substrates at 455C by low-pressure MOVPE. X-ray diffraction measurements showed resolved peaks of InSb and TlInSb. However, in contrast to the theoretical prediction, an increase in Tl incorporation led to a decrease in the lattice constant (higher diffraction angle). This was also observed by Karam et al. [23.20]. The latter obtained TlInSb with Tl composition of 0.1 (10%). Incorporation of Tl was also confirmed with Auger electron spectroscopy and Rutherford backscattering (RBS ) measurements. Infrared transmission measurements showed a shift of the absorption edge toward longer wavelengths with increasing Tl composition for TlInSb [23.18, 23.20, 23.29].

Infrared photoconductors were fabricated from TlInSb/InSb grown on (100) semi-insulating GaAs substrates [23.20, 23.30, 23.31]. A clear shift of response to longer wavelength was observed. The cutoff wavelength extended to 11 μm at 300 K for the TlInSb (Tl composition of 6%) photoconductor [23.30]. The maximum detectivity was \({\mathrm{7.6\times 10^{8}}}\,{\mathrm{cm{\,}Hz^{1/2}{\,}W^{-1}}}\) at 77 K.

Growth of Tl-III-As Ternary Alloys

The growth experiments of TlInAs were conducted by GSMBE [23.25, 23.26, 23.32] and SSMBE [23.27, 23.28]. Takenaka et al. [23.25] confirmed the successful growth of TlInAs on InAs substrate at 430C by double crystal x-ray diffraction measurement (40 arcsec shift) as well as by reflection high-energy electron diffraction (RHEED ) intensity oscillation measurement, where the oscillation period was reduced by the addition of Tl flux during the growth of InAs (1.3% reduction). Lubyshev et al. [23.28] also reported the growth of TlInAs with Tl composition of 0.6%. However, Antonell et al. [23.26, 23.32] and Lange et al. [23.27] reported that the incorporation of Tl into InAs was difficult. The growth of high Tl composition TlInAs has not been accomplished. The window for the appropriate growth conditions, especially substrate temperature, was narrow.

The growth of TlGaAs was conducted on GaAs substrates by SSMBE by Lubyshev et al. [23.28]. They reported the successful growth of TlGaAs with a Tl composition of 0.05 (diffraction peak separation of 1260 arcsec between TlGaAs and GaAs in the (400) double crystal x-ray diffraction (DCXRD )) and a second metal phase (Tl+Ga+As) on the surface. They observed that the growth of TlGaAs is easier than that of TlInAs.

Growth of Tl-III-P Ternary Alloys

TlInP growth was conducted on InP substrates by both MOVPE and GSMBE. Razeghi et al. [23.29] reported the successful growth of ( TlP3) x  ( InP)1−x layers at 520C by MOVPE and observed clear photoresponse with onset wavelengths of 5.5 and 8.0 μm in infrared photoconductivity measurements for two samples (estimated compositions: x = 0.6 and 0.64), respectively. Yamamoto et al. confirmed the GSMBE growth of TlInP by the DCXRD measurement [23.21] and by the RHEED intensity oscillation measurement [23.22]. The growth temperature was about 450C and (2 × 4) RHEED patterns and specular surface morphology with SEM were observed [23.21]. With increasing Tl∕(Tl+In) flux ratio, the XRD peak was shifted to lower diffraction angles, indicating the increase in Tl incorporation. On the other hand, Antonell et al. [23.26] reported the difficulty of the growth of TlInP similar to TlInAs. Narrow growth window for the growth of these materials is probably the reason for this discrepancy.

TlGaP layers closely lattice matched to GaAs were successfully grown on GaAs substrates at about 450C by GSMBE [23.23]. However, TlGaP on GaAs showed the phase separation into three stable compositions: TlGaP nearly lattice matched to GaAs, GaP-like TlGaP and TlP-like TlGaP. However, by adjusting Tl and Ga fluxes, TlGaP layers without phase separation and nearly lattice matched to GaAs were grown. Photoconductance measurement on TlGaP showed the absorption in the 1.3 μm wavelength region, agreeing with the expected band-gap energies for the TlGaP lattice matched to GaAs. Furthermore, a small temperature variation of band-gap energy was observed as expected.

Growth of Tl-III-V Quaternary Alloys

Quaternary TlInGaP layers were successfully grown on InP substrates by GSMBE [23.22, 23.33, 23.34]. During growth, the RHEED pattern showed (2 × 4) reconstructions [23.22]. The DCXRD measurement showed a lower angle shift for the TlInGaP peak with increasing Tl flux [23.33, 23.34], agreeing with the increase of Tl composition. No phase separation was observed. The Tl composition obtained was less than 0.1. The increase of photoconductivity (PC) was observed in the longer wavelength region compared with the PC of InP [23.24], clearly indicating the formation of narrow band-gap semiconductors.

Quaternary TlInGaA layers were also successfully grown on InP substrates by GSMBE [23.25, 23.33, 23.34]. The growth temperature was 450C. The growth condition for alloys containing Tl was very strict. For example, the growth below 400C resulted in segregation of Tl, and the growth above 460C failed to incorporate Tl due to desorption of Tl atoms from the growing surface. During the growth at 450C, the RHEED pattern showed (2 × 2) reconstructions. The grown samples showed mirror surface and incorporation of Tl. With increasing Tl flux, the x-ray diffraction peak shifted toward lower angles and the Tl composition was increased (Fig. 23.4) [23.33]. It is considered that during the growth of TlInGaAs, the large part of incident Tl atoms is re-evaporated from the surface and only a small part of them is incorporated into films, very similar to the case of MBE growth of GaN under usual growth conditions [23.35].
Fig. 23.4

Dependence of Tl composition in TlInGaAs on Tl flux during MBE growth. (After [23.33])

Growth of TlInGaAsN Quinary Alloys

The growth of TlInGaAsN was conducted by GSMBE with thermally cracked AsH3 and plasma-enhanced N2 as As and N sources [23.36]. TlInGaAsN/GaAs quantum well (QW ) structures were grown on GaAs (100) substrates. High Tl flux was needed for the incorporation of Tl into the films.

The TlInGaAsN/TlGaAsN QW samples having higher N concentration in the QWs showed higher Tl incorporation [23.37]. The addition of N species during growth induced the increase of Tl incorporation into TlInGaAs. This is considered due to the strong bonds of Tl–N.

23.1.3 Temperature Dependenceof Physical Properties

TlInGaAs: Photoluminescence

Photoluminescence (PL ) emission was observed from the grown TlInGaAs/InP samples in the temperature range of 10–300 K. With increasing Tl composition, the PL peak energy was shifted toward lower energy [23.38]. As the Tl composition increased, the temperature variation of the PL peak energy decreased and was much smaller than that of InAs, as shown in Fig. 23.5. It is noteworthy that the band-gap energy (≈ 0.8 eV) of TlInGaAs is larger than that (0.356 eV) of InAs. In general, the wider the band gap a material has, the larger is the variation of band-gap energy with temperature.
Fig. 23.5

Temperature variation of the PL peak energy for the TlInGaAs/InP DH samples as a function of Tl composition. (After [23.38])

The temperature variation of as small as \(-{\mathrm{0.03}}\,{\mathrm{meV/K}}\) was observed in Fig. 23.5, which corresponds to the wavelength variation of 0.04 nm ∕ K. This value is much smaller than those of lasing wavelengths for the InGaAsP/InP Fabry–Pérot (FP) LDs (0.4 nm ∕ K) as well as the InGaAsP/InP-distributed feedback (DFB ) LDs (0.1 nm ∕ K). The InGaAsP/InP DFB LDs are presently used in the optical fiber communication systems. With further increase of Tl composition, real temperature-insensitive band-gap energy temperature dependence is expected.

TlInGaAs: Refractive Indexand Band-Gap Energy

A spectroscopic ellipsometer (SE ) was used to obtain the refractive indices of TlInGaAs in the photon energy range of 1.2–2.0 eV (above band-gap energy region) [23.39]. Optical reflectance and absorption measurements were carried out with a double-beam spectrometer in the wavelength range of 1700–2500 nm (below band-gap energy region) to determine the refractive indices n and the band-gap energies E0 (Eg) of TlInGaAs, respectively [23.40].

At the energies above band gap, the refractive index n increases with photon energy [23.39]. However, the temperature variation of n for TlInGaAs (Tl = 0.039) greatly decreases as compared to InGaAs, as can be seen in Fig. 23.6 at the wavelength of 1.0 μm.
Fig. 23.6

Temperature variation of refractive indices for TlInGaAs (with different Tl composition x) and InGaAs at a wavelength of 1.0 μm. (After [23.39])

The photon energy dependencies of n at the energies below band gap for TlInGaAs (Tl = 0.043) and InGaAs at 300 and 340 K together with the calculated results by the first-order Sellmeier equation are shown in Fig. 23.7 [23.40]. The value of n of TlInGaAs is greater than that of InGaAs. The n value increases with increasing temperature. However, the temperature variation of n of TlInGaAs is remarkably smaller than that of InGaAs. Therefore, the temperature variation of n decreases with the increase of Tl composition at the photon energy both above and below Eg.
Fig. 23.7

Refractive index dispersions at T = 300 and 340 K. Gray-colored circles and triangles represent experimental refractive indices at 300 K and brown-colored circles and triangles represent those at 340 K for TlInGaAs (Tl = 0.043) and InGaAs, respectively. Solid and dashed lines represent the calculated results with the first-order Sellmeier equation. (After [23.40])

Table 23.1 [23.40] summarizes the temperature coefficients of the E0 edge, dE0 ∕ dT, and those of the refractive index, \(n^{-1}(\mathrm{d}n/\mathrm{d}T)\), for TlInGaAs (Tl = 0 and 0.043), together with those from [23.41]. Table 23.1 shows that the TlInGaAs alloy, although the Tl composition is limited to small value, exhibits temperature stability not only in the E0 edge but also in n.
Table 23.1

Temperature coefficients of the direct-band edge E0 and the refractive index [n ( λ → ∞ ) ] of InGaAs and TlInGaAs. (After [23.40])

Material

dE0 ∕ dT (\({\mathrm{10^{-4}}}\,{\mathrm{eV/K}}\))

\(n^{-1}(\mathrm{d}n/\mathrm{d}T\)) (\({\mathrm{10^{-5}}}\,{\mathrm{K^{-1}}}\))

Ref.

InGaAs

−3.48

 

[23.41]

−3.5

9.8

[23.40]

TlInGaAs

−2.1

5.0

[23.40]

TlInGaAsN: Photoluminescenceand Electroluminescence

TlInGaAsN/GaAs and InGaAsN/GaAs double quantum well (DQW ) LEDs with GaAs layers as barrier layers were fabricated [23.36]. Reduction in the temperature variation of electroluminescence (EL ) peak energy was observed in the temperature range of 288–328 K by the addition of Tl into InGaAsN/GaAs DQW layers; \(-{\mathrm{0.62}}\,{\mathrm{meV/K}}\) for the InGaAsN/GaAs DQW and \(-{\mathrm{0.53}}\,{\mathrm{meV/K}}\) for the TlInGaAsN/GaAs DQW. By replacing GaAs barrier layers with TlGaAs barrier layers, further reduction was obtained; \(-{\mathrm{0.35}}\,{\mathrm{meV/K}}\) for TlInGaAsN/TlGaAs DQW LEDs [23.36]. SIMS measurements indicated that this improvement is caused by the increased incorporation of Tl into the DQW layers by using TlGaAs barrier layers.

TlInGaAsN/TlGaAsN triple QWs (TQWs) with TlGaAsN layers as barrier layers were investigated [23.37]. The TQW samples having higher N concentration in the TQWs have the highest Tl incorporation without deterioration of the crystalline quality. The temperature dependence of the PL peak energy was found to be the least for the highest Tl containing TQW sample.

23.1.4 Tl-III-V LD Application

TlInGaAs/InP double heterostructure (DH ) LD structures were grown by GSMBE on (100) InP substrates [23.42]. Temperature variation of the EL peak wavelength as small as 0.06 nm ∕ K was confirmed. This is much smaller than the 0.4 nm ∕ K observed for InGaAsP/InP FP LDs as well as the 0.1 nm ∕ K observed for InGaAsP/InP DFB LDs.

Laser operation was achieved under the pulsed condition from 77 to 310 K [23.42]. The threshold current density at room temperature was approximately 7 kA ∕ cm2, and the lasing wavelength was about 1660 nm. The characteristic temperature T0 value for the temperature dependence of the threshold current in the temperature range of 77–300 K was approximately 94 K [23.42], which is similar to that of InGaAsP/InP LDs.

Several longitudinal-mode peaks were observed in the lasing spectra. Each longitudinal mode peak wavelength is determined by the effective refractive index of the TlInGaAs/InP waveguide and the cavity length of the LD. The temperature variation of each longitudinal-mode peak wavelength was as small as 0.06 nm ∕ K [23.42]. However, in the wide temperature range, the main peak of the lasing spectrum moved to another peak of longer wavelength with increasing temperature. The average temperature variation of the main peak wavelength was 0.3 nm ∕ K [23.42]. The temperature variation of the main peak for the FP LDs is mainly determined by the temperature variation of the gain peak. The observed average temperature variation of 0.3 nm ∕ K is due to the temperature variation of the gain peak.

Laser light propagates along both TlInGaAs active layer and InP cladding layer; hence the temperature variation of lasing wavelength is influenced by their temperature variations. The addition of Tl into InP cladding layer results in the smaller temperature-dependent refractive index of cladding layer. The refractive index for TlInP was measured with the SE and it was confirmed that the refractive index of TlInP is larger than that of InP [23.43].

The InP/TlInP/TlInGaAs/TlInP/InP separate confinement heterostructure (SCH ) LDs were fabricated and the improved temperature dependence of EL peak wavelength was obtained as shown in Fig. 23.8 [23.43]. Temperature dependence of the EL peak wavelength was reduced to 0.01 nm ∕ K. The insertion of TlInP layer also increased the incorporated Tl composition in the TlInGaAs active layer.
Fig. 23.8

Temperature dependence of EL peak wavelength for the TlInGaAs/TlInP/InP SCH LED. (After [23.43])

The lasing spectra as a function of temperature and the temperature dependence of main peak wavelength for the SCH LDs under pulsed operation are shown in Fig. 23.9 [23.43]. The temperature variation of main peak wavelength is as small as 0.06 nm ∕ K within the experimental temperature range (297–302 K). This value is much smaller than those of InGaAsP/InP FP LDs (0.4 nm ∕ K) and InGaAsP/InP DFB LDs (0.1 nm ∕ K). This reduced temperature variation is considered to be due to the reduced temperature dependence of refractive indices for TlInGaAs and TlInP compared with those for InGaAs and InP [23.43].
Fig. 23.9

(a) Lasing spectra as a function of temperature and (b) temperature dependence of lasing peak wavelength for the TlInGaAs/ TlInP/InP SCH LD under pulsed operation. (After [23.43])

The observed temperature variation of the main peak wavelength for the TlInGaAs/TlInP/InP SCH LD is larger than that of the EL peak wavelength, which corresponds to that of gain peak or that of band-gap energy [23.43]. The main peak in the FP LD is one of the longitudinal-mode peaks, which are determined by the effective waveguide refractive index, that is, refractive indices of TlInGaAs active layers, TlInP cladding layers, and InP cladding layers, as well as the cavity (waveguide) length. Therefore, the temperature variation of the main peak cannot be smaller than that of the longitudinal-mode peak [23.43].

23.2 III-V-Bi Alloy Semiconductors

Bismuth-containing III-V (III-V-Bi) semiconductors were first studied for the applications to mid-infrared and far-infrared optical devices [23.44]. InSbBi [23.45, 23.46, 23.47, 23.48], InAsBi [23.49, 23.50], and InAsSbBi [23.49] were grown by various growth methods, such as the Czochralski bulk growth method, RF sputtering, MBE, and MOVPE [23.44]. The solid solubility of Bi in InSb was reported to be as low as 2.6% in the bulk growth [23.45]. Low temperature MBE growth increased the Bi incorporation up to 5% [23.47].

Oe et al. [23.1] suggested the unique possible properties (temperature-independent band-gap energy characteristics) of III-V-Bi semiconductors and proposed the application to the temperature-insensitive wavelength LDs.

23.2.1 Expected Propertiesof Bi-Based III-V Alloys

Figure 23.10 shows the expected relationships between the band-gap energy and the lattice constant for the III-V-Bi alloy semiconductors [23.44]. As already described, the alloy semiconductors consisting of a semiconductor and a semimetal are expected to exhibit the temperature-insensitive band-gap energy similar to HgCdTe. The temperature-insensitive band gap is caused by the reverse temperature dependence of band gap (overlap) energy between the semiconductor and semimetal [23.1]. InGaAsBi was first suggested as a candidate for the temperature-insensitive band-gap energy semiconductor. GaAs and InAs are semiconductors and GaBi and InBi are semimetals.
Fig. 23.10

Expected relationships between the band-gap energy and lattice constant for the III-V-Bi alloy semiconductors. (After [23.44])

To adjust the lasing wavelength of GaAsBi and related alloys to the waveband of optical fiber communication, a quaternary alloy of GaNAsBi was proposed which can be lattice matched to GaAs [23.44]. On the other hand, InGaAsBi is lattice matched to InP, but an additional element such as Al has to be added to InGaAsBi to adjust its band gap to the waveband of optical fiber communication [23.44]. GaAsBi and GaNAsBi have a characteristic of large spin–orbit interaction, which helps in reducing the Auger recombination in LDs [23.51].

23.2.2 Growth of III-V-Bi Alloys

Growth of III-V-Bi Ternary Alloys

In MBE growth of InSbBi, it is imperative to control the V∕III ratio (Sb∕In flux ratio) in the range of near unity to achieve Bi incorporation into InSb [23.44, 23.47]. Bi incorporation was suppressed at Sb∕In ratio greater than unity despite the existence of Bi atoms on the growing surface. The strength of the In–Sb bond is considered to be stronger than that of In–Bi. Therefore, the Sb atoms are preferentially incorporated and the Bi atoms are difficult to be incorporated when the Sb∕In flux ratio is greater than unity. Furthermore, the In–Bi bond is easily replaced by the In–Sb bond. Another important condition is that the growth temperature has to be set lower than 400C to avoid re-evaporation of Bi atoms from the growing surface [23.44, 23.47]. The InBi molar fraction increases by decreasing the substrate temperature (Tsub) [23.47].

After the proposal of GaAsBi alloys as a temperature-insensitive band-gap semiconductor [23.1], GaAsBi was grown on GaAs substrates by MOVPE [23.53]. To obtain the GaAsBi alloys, growth at relatively low temperatures was needed because Bi has high vapor pressure. However, in the MOVPE growth, the insufficient decomposition of MOs at low temperatures induced the incorporation of carbon (C) impurity species into GaAsBi films. On the other hand, in the MBE growth, elemental Ga, As, and Bi are used as sources [23.52, 23.54], and the incorporation of C impurity species is expected to be avoided.

GaAsBi was grown on GaAs (001) substrates at Tsub less than 400C by MBE [23.52, 23.54]. During the growth of GaAsBi, the RHEED pattern revealed the surface reconstruction of (2 × 1), suggesting the existence of Bi–Bi or Bi–As dimers [23.44, 23.55]. The unusual metallic nature of Bi-induced (2 × 1) was proposed as the origin of the surfactant nature of Bi [23.44, 23.56]. RHEED intensity oscillations observed during growth suggest a layer-by-layer growth [23.57].

For the growth of GaAsBi, the GaBi molar fraction decreased with increasing Tsub under constant Ga and As fluxes as shown in Fig. 23.11a [23.52]. At Tsub higher than 400C, Bi atoms are not incorporated into GaAs. Therefore, the GaAsBi alloys have to be grown at low temperature. However, the GaBi molar fraction increases with Bi flux, followed by the saturation for the constant Ga and As fluxes at the constant Tsub as shown in Fig. 23.11b. The unincorporated Bi atoms are desorbed from the surface without the formation of Bi droplets below critical Bi supply, although excess Bi supply forms Bi-containing droplets on the growth surface [23.52]. The droplet formation causes severe surface roughness and inhomogeneity.
Fig. 23.11

(a) GaBi molar fraction in GaAsBi against substrate temperature for the constant Bi flux. (b) GaBi molar fraction in GaAsBi against Bi flux for the same substrate temperature. (After [23.52])

The range of the As flux is also very narrow to obtain GaAsBi films [23.52]. For the growth at \(T_{\text{sub}}={\mathrm{380}}\,{\mathrm{{}^{\circ}\mathrm{C}}}\) with \(\text{Ga flux}={\mathrm{3\times 10^{-7}}}\,{\mathrm{Torr}}\) and \(\text{Bi flux}={\mathrm{2\times 10^{-8}}}\,{\mathrm{Torr}}\), only GaAs without Bi incorporation was grown above As flux of 10−5 Torr, while below 10−6 Torr the surface of the film becomes rough and only heavily degraded film was grown. Therefore, to grow the device quality GaAsBi, careful adjustment of the source supply and Tsub is needed.

Growth of III-V-Bi Quaternary Alloys

GaNAsBi quaternary alloys can be lattice matched to GaAs substrate with the band gap of optical fiber communication wavelengths. They were grown on GaAs (100) substrate by MBE using solid sources of Ga, As, and Bi and RF-plasma-enhanced N2 gas at Tsub of 350–400C [23.58]. Precise control of As flux within a limited range and low temperature growth was crucial as in the case of GaAsBi. During growth, the (2 × 1) RHEED pattern was observed also similar to GaAsBi, implying Bi-stabilized reconstruction. The GaBi molar fraction decreased with the increase of Tsub for the constant Ga, As, and Bi fluxes, which is similar to the case of GaAsBi growth. Lattice matching between GaNAsBi film and GaAs substrate was achieved by adjusting GaN molar fraction [23.59].

The InGaAsBi quaternary alloy is another extension of GaAsBi and can be lattice matched to InP [23.60]. Considering the band gaps of InAs (0.36 eV) and GaAs (1.42 eV), the band gap of InGaAsBi is expected to cover the whole wavelength range used in the optical fiber communication systems, and InGaAsBi is a promising semiconductor to fabricate the temperature-insensitive wavelength LDs [23.60].

InGaAsBi was grown on InP (100) substrate by MBE using solid sources of In, Ga, As, and Bi [23.60]. Tsub was ranging from 270 to 350C, which is lower by 50C than that for GaAsBi growth, corresponding to the lower desorption temperature of Bi atoms from the InGaAs surface than that from the GaAs surface. During the growth of InGaAsBi, the (1 × 3) streaky RHEED pattern was observed. The XRD pattern showed a well-defined diffraction peak with distinct Pendellösung fringes, indicating the homogeneous alloy composition and high epitaxial quality with smooth surface. RBS channeling angular scans showed that Bi atoms are located exactly on the substitutional sites in InGaAs.

23.2.3 Optical Properties of III-V-Bi Alloys

Although the GaAsBi films were grown at low temperatures below 400C by both MOVPE [23.53] and MBE [23.58, 23.61, 23.62], intense PL emission was observed at room temperature. This is considered to be due to the surfactant effect of Bi atoms during growth. Figure 23.12 shows the PL spectra as a function of temperature and the temperature variation of the PL peak energy for the MOVPE-grown GaAsBi with GaBi molar fraction of 0.024 [23.53]. The temperature variation of the PL peak energy is as small as 0.1 meV ∕ K, which is much smaller than that of GaAs as can be seen in Fig. 23.12b. The results suggest that the band gap of GaAsBi is less sensitive to temperature than the usual III-V semiconductors [23.53]. The PL peak energy decreased with the increase of GaBi molar fraction [23.58]. Reduction in the temperature variation of PL peak energy was also observed for the MBE-grown GaAsBi samples [23.61, 23.62]. The temperature coefficient of PL peak energy was 0.23 meV ∕ K [23.61] and 0.15 meV ∕ K [23.62] for the GaBi molar fraction of 0.013 and 0.025, respectively. The temperature coefficients of GaAsBi band gaps obtained by photoreflectance spectroscopy were 0.24, 0.23, and 0.15 meV ∕ K between 150 and 300 K for samples with GaBi = 0.005, 0.013, and 0.026, respectively [23.63]; that is, with the increase of GaBi molar fraction the temperature coefficient of band gap decreased.
Fig. 23.12

(a) PL spectra as a function of temperature and (b) temperature variation of PL peak energy for the GaAsBi (GaBi = 0.024). For comparison, the band gap of GaAs is also shown. (After [23.53])

To obtain intense PL emission for the grown GaNAsBi, post-growth thermal annealing was needed, similar to GaInNAs. PL intensity of the GaNAsBi (GaBi = 0.02) film was increased by approximately fivefold after thermal annealing as shown in Fig. 23.13 [23.58]. PL peak energy of GaNAsBi showed no notable variation even after thermal annealing. The PL peak energy was decreased with the increase of GaBi molar fraction [23.58]. The temperature coefficient of the PL peak energy for GaNAsBi was very close to that for GaAsBi with the same GaBi molar fraction. Figure 23.14 shows the temperature variation of PL peak energy for three samples [23.58]. The temperature variation of PL peak energy for GaNAsBi is smaller than that of InGaAsP with similar band-gap energy (near 1300 nm). It can be said that the temperature coefficients of GaAsBi and GaNAsBi are governed by GaBi molar fraction and not by GaN molar fraction [23.58].
Fig. 23.13

PL spectra of as-grown and annealed GaNAsBi (GaBi = 0.02). (After [23.58])

Fig. 23.14

Temperature variation of PL peak energy for GaNAsBi (GaBi = 0.02, 0.037, 0.047) and InGaAsP. (After [23.58])

23.2.4 III-V-Bi LD Application

Photo-pumped lasing oscillation at 983 nm was achieved in the MBE-grown GaAs/GaAsBi/GaAs (GaBi = 0.025) DH structure sample [23.62]. The temperature coefficient of the lasing wavelength decreased to 40% of that of the band gap of GaAs in the temperature range of 150–240 K. By using the GaAs/GaAsBi/AlGaAs DH structure sample to improve the carrier confinement, the photo-pumped lasing oscillation at the longer wavelength of up to 1200 nm was achieved [23.65]. The temperature coefficient of the lasing wavelength was about 40% to that of the 1300 nm InGaAsP LDs in the temperature range of 20–80C.

Electrically pumped lasing oscillation at 947 nm was demonstrated in the GaAsBi (GaBi = 0.022) FP-LD grown by MOVPE [23.66], though the temperature coefficient of the lasing wavelength is unclear. GaAsBi (GaBi = 0.03 or 0.04)/AlGaAs SCH FP-LD was grown by MBE [23.64]. Lasing oscillation was observed in the temperature range of 15–40C. The characteristic temperature T0 for the GaAsBi (GaBi = 0.03)/AlGaAs SCH FP-LD was calculated to be 125 K, which is higher than that of 1300 nm InGaAsP FP-LD (T0 = 60 K) and is similar to that for the 980 nm InGaAs/GaAs LDs. High T0 is because of the stronger carrier confinement of the GaAsBi/AlGaAs system than that of the InGaAsP/InP system [23.64]. The temperature coefficient of lasing wavelength was reduced to 0.17 nm ∕ K, which is 45% of that (0.37 nm ∕ K) for the GaAs FP-LD as shown in Fig. 23.15 [23.64].
Fig. 23.15

Lasing wavelength shift for the GaAsBi and GaAs FP-LDs from the lasing wavelength at 15C in the temperature range of 15–25C. (After [23.64])

23.3 Summary

Tl-III-V and III-V-Bi alloy semiconductors are important materials for the fabrication of temperature-insensitive lasing wavelength LDs as well as LWIR optical devices. This chapter described the expected interesting properties of these semiconductors and the present status of their growth, characterization, and device applications. It was shown that these alloy semiconductors have a potential to have temperature-insensitive band gap and to fabricate temperature-insensitive lasing wavelength LDs.

References

  1. 23.1
    K. Oe, H. Asai: IEICE Trans. Electron. E79-C, 1751 (1996)Google Scholar
  2. 23.2
    D. Long: Energy Band in Semiconductors (Wiley, New York 1968)Google Scholar
  3. 23.3
    W.S. Pelouch, L.A. Schline: Appl. Phys. Lett. 68, 1389 (1996)CrossRefGoogle Scholar
  4. 23.4
    C.J. Summers, J.G. Broerman: Phys. Rev. B21, 559 (1980)CrossRefGoogle Scholar
  5. 23.5
    R. Dornhaus, G. Nimtz: Springer Tracts in Modern Physics, Vol. 78 (Spring, Berlin 1976)Google Scholar
  6. 23.6
    H. Asahi, K. Yamamoto, K. Iwata, S. Gonda, K. Oe: Jpn. J. Appl. Phys. 35, L876 (1996)CrossRefGoogle Scholar
  7. 23.7
    H. Asahi: Compound Semicond. 2, 34 (1996)Google Scholar
  8. 23.8
    M. van Schilfgaarde, A. Sher, A.B. Chen: Appl. Phys. Lett. 62, 1857 (1994)CrossRefGoogle Scholar
  9. 23.9
    A. Sher, M. van Schilfgaarde, S. Krishnamurthy, M.A. Berding, A.B. Chen: J. Electron. Matter. 24, 1119 (1995)CrossRefGoogle Scholar
  10. 23.10
    A.B. Chen, M. van Schilfgaarde, A. Sher: J. Electron. Matter. 22, 843 (1993)CrossRefGoogle Scholar
  11. 23.11
    M. van Schilfgaarde, A.B. Chen, S. Krishnamurthy, A. Sher: Appl. Phys. Lett. 65, 2714 (1994)CrossRefGoogle Scholar
  12. 23.12
    H. Asahi, K. Yamamoto, M. Fushida, H. Koh, K. Asami, S. Gonda, K. Oe: Proc. 9th Int. Conf. InP Relat. Mater. (1997) pp. 448–451Google Scholar
  13. 23.13
    H.J. Lee, A. Mizobata, O. Maeda, K. Konishi, K. Asami, H. Asahi: Jpn. J. Appl. Phys. 41, 1016 (2002)CrossRefGoogle Scholar
  14. 23.14
    M.A. Berding, M. van Schilfgaarde, A. Sher, M.J. Antonell, C.R. Abernathy: J. Electron. Matter. 26, 683 (1997)CrossRefGoogle Scholar
  15. 23.15
    W. Mönch: Appl. Phys. Lett. 71, 1231 (1997)CrossRefGoogle Scholar
  16. 23.16
    M. Kondow, M.C. Larson, T. Kitatani, K. Nakahara, K. Uomi: Proc. Conf. Indium Phosphide Relat. Mater. (1998) p. 233Google Scholar
  17. 23.17
    C.K. Kim, T. Miyamoto, Y.H. Lee: Jpn. J. Appl. Phys. 37, 5994 (1998)CrossRefGoogle Scholar
  18. 23.18
    Y.H. Choi, C. Besikci, R. Sudharsanan, M. Razeghi: Appl. Phys. Lett. 63, 361 (1993)CrossRefGoogle Scholar
  19. 23.19
    K.T. Huang, R.M. Cohen, G.B. Stringfellow: J. Cryst. Growth 156, 320 (1995)CrossRefGoogle Scholar
  20. 23.20
    N.H. Karam, R. Sudharsanan, T. Parodos, M.A. Dodd: J. Electron. Matter. 25, 1209 (1996)CrossRefGoogle Scholar
  21. 23.21
    K. Yamamoto, H. Asahi, M. Fushida, K. Iwata, S. Gonda: J. Appl. Phys. 81, 1704 (1997)CrossRefGoogle Scholar
  22. 23.22
    H. Asahi, M. Fushida, K. Yamamoto, K. Iwata, H. Koh, K. Asami, S. Gonda, K. Oe: J. Cryst. Growth 175/176, 1195 (1997)CrossRefGoogle Scholar
  23. 23.23
    M. Fushida, H. Asahi, K. Yamamoto, H. Koh, K. Asami, S. Gonda, K. Oe: Jpn. J. Appl. Phys. 36, L665 (1997)CrossRefGoogle Scholar
  24. 23.24
    H. Koh, H. Asahi, M. Fushida, K. Yamamoto, K. Asami, S. Gonda, K. Oe: J. Cryst. Growth 188, 107 (1998)CrossRefGoogle Scholar
  25. 23.25
    K. Takenaka, H. Asahi, H. Koh, K. Asami, S. Gonda, K. Oe: Jpn. J. Appl. Phys. 38, 1078 (1999)CrossRefGoogle Scholar
  26. 23.26
    M.J. Antonell, C.R. Abernathy, A. Sher, M. Berding, M. van Schilfgaarde, A. Sanjuro, K. Wong: J. Cryst. Growth 188, 13 (1998)CrossRefGoogle Scholar
  27. 23.27
    M.D. Lange, D.F. Storm, T. Cole: J. Electron. Matter. 27, 36 (1998)CrossRefGoogle Scholar
  28. 23.28
    D.L. Lubyshev, W.Z. Cai, G.L. Catchen, T.S. Mayer, D.L. Miller: Proc. IEEE 24th Intern. Symp. Compound Semicond. (1997) p. 125Google Scholar
  29. 23.29
    M. Razeghi, J.D. Kim, S.J. Park, Y.H. Choi, D. Wu, E. Michel, J. Xu, E. Bigan: Inst. Phys. Conf. Ser. 145, 1085 (1996)Google Scholar
  30. 23.30
    J.D. Kim, E. Michel, S. Park, J. Xu, S. Javadpour, M. Razeghi: Appl. Phys. Lett. 69, 343 (1996)CrossRefGoogle Scholar
  31. 23.31
    P.T. Staveteig, Y.H. Choi, G. Labeyrie, E. Bigan, M. Razeghi: Appl. Phys. Lett. 64, 60 (1994)CrossRefGoogle Scholar
  32. 23.32
    M.J. Antonell, C.R. Abernathy, A. Sher, M. Berding, M. van Schilfgaarde: Proc. of 9th Intern. Conf. InP Relat. Mater. (1997) pp. 444–447Google Scholar
  33. 23.33
    H. Asahi, H. Koh, K. Takenaka, K. Asami, K. Oe, S. Gonda: J. Cryst. Growth 201/202, 1069 (1999)CrossRefGoogle Scholar
  34. 23.34
    H. Asahi, K. Takenaka, H. Koh, K. Oe, K. Asami, S. Gonda: Inst. Phys. Conf. Ser. 162, 541 (1999)Google Scholar
  35. 23.35
    H. Asahi, S. Hasegawa: IPAP Book 2 (The Institute of Pure and Applied Physics, Tokyo 2005) p. 91Google Scholar
  36. 23.36
    T. Matsumoto, D. Krishnamurthy, A. Fujiwara, S. Hasegawa, H. Asahi: J. Cryst. Growth 295, 133 (2006)CrossRefGoogle Scholar
  37. 23.37
    K.M. Kim, S. Emura, D. Krishnamurthy, S. Hasegawa, H. Asahi: J. Appl. Phys. 108, 053501 (2010)CrossRefGoogle Scholar
  38. 23.38
    A. Ayabe, H. Asahi, H.J. Lee, O. Maeda, K. Konishi, K. Asami, S. Gonda: Appl. Phys. Lett. 77, 2148 (2000)CrossRefGoogle Scholar
  39. 23.39
    A. Imada, H.-J. Lee, A. Fujiwara, S. Emura, S. Hasegawa, H. Asahi: J. Appl. Phys. 94, 6976 (2003)CrossRefGoogle Scholar
  40. 23.40
    A. Imada, H.J. Lee, A. Fujiwara, T. Mukai, H. Asahi: Appl. Phys. Lett. 84, 4212 (2004)CrossRefGoogle Scholar
  41. 23.41
    E. Zielinski, H. Schweizer, K. Streubel, H. Eisele, G. Weimann: J. Appl. Phys. 59, 2196 (1986)CrossRefGoogle Scholar
  42. 23.42
    A. Fujiwara, H.J. Lee, A. Imada, S. Hasegawa, H. Asahi: Jpn. J. Appl. Phys. 42, L1359 (2003)CrossRefGoogle Scholar
  43. 23.43
    A. Fujiwara, D. Krishnamurthy, T. Matsumoto, S. Hasegawa, H. Asahi: J. Cryst. Growth 301, 109 (2007)CrossRefGoogle Scholar
  44. 23.44
    M. Yoshimoto, K. Oe: Molecular beam epitaxy of GaAsBi and related alloys. In: Molecular Beam Epitaxy, ed. by M. Henini (Elsevier, Oxford 2013) p. 159CrossRefGoogle Scholar
  45. 23.45
    B. Joukoff, A.M. Jean-Louis: J. Cryst. Growth 12, 169 (1972)CrossRefGoogle Scholar
  46. 23.46
    J.L. Zilko, J.E. Greene: J. Appl. Phys. 51, 1549 (1980)CrossRefGoogle Scholar
  47. 23.47
    K. Oe, S. Ando, K. Sugiyama: Jpn. J. Appl. Phys. 20, L303 (1981)CrossRefGoogle Scholar
  48. 23.48
    A.J. Noreika, W.J. Takei, M.H. Francombe: J. Appl. Phys. 53, 4932 (1982)CrossRefGoogle Scholar
  49. 23.49
    Y.K. Ma, Z.M. Fang, D.H. Jaw, R.M. Cohen, G.B. Stringfellow: Appl. Phys Lett. 55, 2420 (1989)CrossRefGoogle Scholar
  50. 23.50
    H. Ofuchi, T. Kubo, M. Tabuchi, Y. Takeda, H. Okamoto, K. Oe: Jpn. J. Appl. Phys. 38, 545 (1999)CrossRefGoogle Scholar
  51. 23.51
    M. Yoshimoto, T. Fuyuki: Localized states in GaAsBi and GaAs/GaAsB heterostructures. In: Bismuth-Containing Compounds, ed. by H. Li, Z.M. Wang (Springer, New York 2013) pp. 201–224CrossRefGoogle Scholar
  52. 23.52
    M. Yoshimoto, S. Murata, A. Chayahara, Y. Horino, J. Saraie, K. Oe: Jpn. J. Appl. Phys. 42, L1235 (2003)CrossRefGoogle Scholar
  53. 23.53
    K. Oe, H. Okamoto: Jpn. J. Appl. Phys. 37, L1283 (1998)CrossRefGoogle Scholar
  54. 23.54
    S. Tixier, M. Adamcyk, T. Tiedje, S. Francoeur, A. Mascarenhas, P. Wei, F. Schiettekatte: Appl. Phys. Lett. 82, 2245 (2003)CrossRefGoogle Scholar
  55. 23.55
    P. Laukkannen, M.P.J. Punkkinen, H.-P. Komsa, M. Ahola-Tuomi, K. Kokko, M. Kuzumin, J. Adell, J. Sadowski, R.E. Peraelae, M. Ropo, T.T. Rantala, I.J. Vaeyrynen, M. Pessa, L. Vitos, J. Kollar, S. Mirbt, B. Johansson: Phys. Rev. Lett. 100, 086101 (2008)CrossRefGoogle Scholar
  56. 23.56
    M.P.J. Punkkinen, P. Laukkannen, H.-P. Komsa, M. Ahola-Tuomi, M. Raesaenen, K. Kokko, M. Kuzmin, J. Adell, J. Sadowski, R.E. Peraelae, M. Ropo, T.T. Rantala, I.J. Vaeyrynen, M. Pessa, L. Vitos, J. Kollar, S. Mirbt, B. Johansson: Phys. Rev. B 78, 195304 (2008)CrossRefGoogle Scholar
  57. 23.57
    Y. Tominaga, Y. Kinoshita, K. Oe, M. Yosimoto: Appl. Phys. Lett. 93, 131915 (2008)CrossRefGoogle Scholar
  58. 23.58
    W. Huang, K. Oe, G. Feng, M. Yoshimoto: J. Appl. Phys. 98, 053505 (2005)CrossRefGoogle Scholar
  59. 23.59
    W. Huang, M. Yoshimoto, Y. Takehara, J. Saraie, K. Oe: Jpn. J. Appl. Phys. 43, L1350 (2004)CrossRefGoogle Scholar
  60. 23.60
    G. Feng, M. Yoshimoto, K. Oe, A. Chayahara, Y. Horino: Jpn. J. Appl. Phys. 44, L1161 (2005)CrossRefGoogle Scholar
  61. 23.61
    G. Pettinari, A. Polimeni, M. Capizzi, J.H. Blokland, P.C.M. Christianen, J.C. Maan, E.C. Young, T. Tiedje: Appl. Phys. Lett. 92, 262105 (2008)CrossRefGoogle Scholar
  62. 23.62
    Y. Tominaga, K. Oe, M. Yosimoto: Appl. Phys. Express 3, 062201 (2010)CrossRefGoogle Scholar
  63. 23.63
    J. Yoshida, T. Kita, O. Wada, K. Oe: Jpn. J. Appl. Phys. 42, 371 (2003)CrossRefGoogle Scholar
  64. 23.64
    T. Fuyuki, K. Yoshida, R. Yoshida, M. Yoshimoto: Appl. Phys. Express 7, 082101 (2014)CrossRefGoogle Scholar
  65. 23.65
    T. Fuyuki, R. Yoshida, K. Yoshida, M. Yoshimoto: Appl. Phys. Lett. 103, 202105 (2013)CrossRefGoogle Scholar
  66. 23.66
    P. Ludewig, N. Knaub, N. Hossain, S. Reinhard, L. Nattermann, I.P. Marko, S.R. Jin, H. Hild, S. Chatterjee, W. Stols, S.J. Sweeney, K. Volt: Appl. Phys. Lett. 102, 242115 (2013)CrossRefGoogle Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  1. 1.Inst. of Scientific and Industrial ResearchOsaka UniversityOsakaJapan

Personalised recommendations