X-ray absorption spectroscopy for lattice-matched and strain-relaxed GaAsPBi epi-layers on GaAs (001) substrates

Saharat Chomdech ^a, Chalermchai Himwas *^a, Wenich Pumee ^a, Suphakan Kijamnajsuk ^b, Waraporn Tanthanuch ^c and Songphol Kanjanachuchai *^a
aSemiconductor Device Research Laboratory, Department of Electrical Engineering, Faculty of Engineering, Chulalongkorn University, 254 Phayathai Road, Bangkok 10330, Thailand. E-mail: Chalermchai.H@chula.ac.th
^bNational Metal and Materials Technology Center, Thailand Science Park, 114 Paholyothin Rd., Klong 1, Klong Luang, Pathumthani 12120, Thailand
^cSynchrotron Light Research Institute (Public Organization), Muang, Nakhon Ratchasima 30000, Thailand

First published on 13th August 2024

---

---

Introduction

Bi-containing materials are a relatively new system developed recently due to the difficulties in incorporating large Bi atoms into semiconducting compound hosts. Significant effects caused by alloying Bi in semiconducting hosts are band gap reduction, valence-band modification, and enhanced optical stability. Adding Bi in a GaAs matrix can reduce the band gap for 80–90 meV per %Bi.^1,2 The electron mobility of GaAsBi is high and close to that of GaAs host³ because the addition mainly perturbs the valence band.⁴ The alloy is promising in optoelectronics because it possesses luminescent peak stability against temperature modification.⁵ However, homogeneously incorporating Bi atoms in III–V compounds is complicated. Atomic ordering^6–8 and surface droplets^9,10 can occur during the epitaxial growths of GaAsBi films, especially for high Bi concentrations, owing to the large Bi atoms and the low electronegativity. Scientists attempt to improve the Bi incorporation to make the alloy suitable for various applications, but there is a tradeoff between the Bi concentration and the crystal quality. The growth parameters must be elaborately adjusted.^11–15

One way to increase the Bi concentration without sacrificing the crystal quality is to synthesize the GaAsPBi epitaxial layer instead of the GaAsBi. However, the synthesis of GaAsPBi is not straightforward. There is a catalytic effect of the Bi atoms towards the existing P atoms, making it hard to control the alloy composition. Furthermore, orthogonal dislocations and running droplets readily occur if grown without using the reported strategic procedure.¹⁶ Another challenge when incorporating more elemental atoms in an epitaxial layer is the uncertainty in the compositional estimation. Nattermann et al. estimated the GaAsPBi alloy composition by comparing the optical results against the structural model¹⁷ when our earlier works relied on synchrotron-based X-ray photoelectron spectroscopy (XPS) for the quaternary alloy compositional estimation under a cation–anion stoichiometric assumption.^16,18 XPS is a surface-sensitive technique that measures all ejected electrons from the surface, hence unsuitable for probing inhomogeneous epitaxial layers. Besides, Bi-containing materials grown at low temperatures usually host more defects/dislocations if the growth parameters are not elaborately adjusted. These defects/dislocations invalidate the carriers participating in the radiative luminescent mechanism. Thermal annealing in hydrogen-containing gas aims to permeate hydrogen atoms into the epilayer to neutralize these defects/dislocations and free up carriers, thus improving luminescent intensity.^19,20 Since high-temperature device fabrication processes can easily disrupt the epitaxial layer and degrade the device's quality, another benefit of the annealing observations is to set the maximum temperature that the epitaxial can tolerate. Here, we propose a means to estimate the compositions of the GaAsPBi epitaxial layers and study the ex situ annealing of the lattice-matched GaAsPBi epitaxial layer. This work reports the study of X-ray absorption near-edge structures (XANES) and compositional estimations using extended X-ray absorption fine structures (EXAFS) for defective and lattice-matched GaAsPBi epitaxial layers by implying information obtained from surface and structural investigations. This innovative technique considers only bound atoms in the local structure, making it suitable for estimating the composition of both defective and homogeneous compounds. Subsequently, we report the effect of ex situ annealing on the structural properties, optical properties, XANES spectra, and EXAFS spectra of the lattice-matched GaAsPBi on GaAs (001).

Experimental details

Epi-ready GaAs (001) substrates were outgassed in an ultra-high vacuum at 450 °C for 1 hour. The residual surface oxides were subsequently desorbed in the growth chamber (MBE, Riber 32P) at the substrate temperature (T_s) of 600 °C for 30 min under the As₄ beam equivalent pressure (BEP_As4) of 8.0 × 10⁻⁶ Torr. The two-step thermal desorption resulted in oxide-free corrugated GaAs surfaces. 300 nm-thick GaAs buffer layers were deposited to obtain atomically flat surfaces before synthesizing 200 nm-thick GaAsPBi films. The buffer layers and the quaternary alloy films were synthesized in the growth chamber with solid sources to vaporize Ga atoms, Bi atoms, As₄ molecules, and P₂ molecules. The GaAs buffer layers were grown at T_s = 580 °C, with BEP_Ga = 1.1 × 10⁻⁷ Torr and BEP_As4 = 8.0 × 10⁻⁶ Torr. Two quaternary alloy samples (A and B) were grown at the identical T_s = 320 °C, BEP_Ga = 1.1 × 10⁻⁷ Torr, and BEP_Bi = 5.0 × 10⁻⁷ Torr but different BEP_As4 and BEP_P2. For sample A (sample B), the supplied BEP_As4 was 3.3 × 10⁻⁶ Torr (4.0 × 10⁻⁶ Torr), and BEP_P2 was 5.8 × 10⁻⁷ Torr (1.5 × 10⁻⁷ Torr). The reported T_s values were derived from the non-contact thermocouple temperatures calibrated against the GaAs surface reconstruction and oxide desorption temperatures. We real-time assessed the surface morphologies during growth by high-energy electron diffraction (RHEED) patterns. From earlier studies, the growth conditions for sample A triggered dislocations and inhomogeneity owing to the lattice-mismatch¹⁶ while those for sample B allowed to have the quaternary alloy with a lattice constant close to that of the substrate (lattice-matching condition).¹⁸

The surfaces of samples A and B were measured using an atomic force microscope (AFM, SPA-400, Seiko Instruments) and a differential interference contrast (DIC) optical microscope (OM). The structural properties of both samples were probed by high-resolution X-ray diffraction (HR-XRD, Rigaku). The chemical formations were characterized by advanced X-ray absorption spectroscopy (XAS). XAS spectra were acquired at the SUT-NANOTEC-SLRI XAS beamline (BL5.2) of the SLRI, Thailand. The X-ray light has an energy ranging from 1240–12100 eV and a photon flux ranging from 10⁸–10¹⁰ photons per s/100 mA. An InSb (111) double-crystal monochromator was used for energy scanning. The Bi–M5 edge XANES calibration was performed using a Mo foil in transmission mode. Spectral acquisition was conducted from −30 eV to 100 eV relative to E₀ at 2580 eV, with an energy step of 0.2. For P-K edge XAS measurement, the elemental phosphorus standard was used for calibration. P-K edge XANE acquisition covered a range from −20 eV to 90 eV of E₀ at 2146 eV with an energy step of 0.2. In P-K edge EXAFS spectral acquisitions, the energy scan was set up in a range of −100, −20, 30, and 10.6 k of E₀ at 2146 eV with an energy step of 5, 0.3, 0.05 k, with a corresponding time step of 1, 1, and 3 s. Data preprocessing and theoretical fitting were performed consecutively using Athena-IFEFFIT and Artemis-IFEFFIT packages.²¹

For the ex situ annealing study of lattice-matched GaAsPBi, sample B is cut into four pieces, then annealed for ∼3 hours under the flow of hydrogen-containing gas at different annealing temperatures (T_A): 450 °C (sample C), 550 °C (sample D), and 650 °C (sample E). The structural properties of samples B, C, D, and E were measured using AFM, OM, and high-resolution XRD. The optical quality of the samples was assessed by photoluminescence at 30 K. The samples were mounted on the cold finger of a close cycle He cryostat and excited by Ventus solid-state laser (λ = 532 nm) at an average power density of 50 W cm⁻². Focused PL emitted from the samples was dispersed in a monochromator (Horiba iHR320) and collected by a cooled (77 K) InGaAs detector using a standard lock-in detection technique. PL intensity acquired for each sample was normalized against a standard sample to allow meaningful comparisons. Finally, we report the effect of ex situ annealing on the XAS spectra.

Results and discussion

The AFM micrograph of sample A (root-mean-square (RMS) roughness = 11.2 nm) shows stripes along the [110] direction (see Fig. 1(a)). The lower magnification OM image (Fig. 1(b)) reveals that stripes also occur along the [10] direction. Thus, the surface exhibits orthogonal dislocations called cross-hatched patterns (CHPs).²² The CHPs emerged from the strain relaxation, possessing a strain variation around the dislocation lines. This strain variation affects the adatom incorporations, resulting in in-plane compositional inhomogeneity. On the contrary, the AFM micrograph of sample B shows a flat surface (see Fig. 1(c)) with the RMS roughness of 0.74 nm, agreeing with the lower magnification OM image (Fig. 1(d)). Fig. 1(e) depicts the HR-XRD scans around the symmetric (004) reflections of GaAsPBi epitaxial layers (samples A and B). Each diffractogram shows two peaks associated with the GaAs buffer and GaAsPBi epitaxial layers. Theoretically, the 2θ diffraction peak for GaAs (GaP) binary compound is 66.05° (68.85°), and that for GaBi is 58.28° which corresponds to the lattice constant of 6.33 Å. Both diffractograms show identical 2θ diffraction peaks for GaAs at 66.05°. The 2θ diffraction peak of GaAsPBi for sample A is 68.06°, approaching that of GaP which suggests that the epitaxial layer contains a high P concentration. The highly lattice-mismatched epitaxial layer results in a relaxed quaternary alloy in the form of orthogonal dislocations (see Fig. 1(b)). The in-plane inhomogeneity triggers the broad-linewidth diffractogram. The 2θ diffraction peak of GaAsPBi for sample B is 66.33°. The epitaxial layer is almost lattice-matched to the substrate (GaAs). The narrow linewidth diffractogram suggests that the layer is homogeneous.

We use X-ray absorption spectroscopy (XAS) to characterize the structures of the quaternary alloy. Depending on the energy range extended from the absorption edge of an atom, XAS is categorized into XANES and EXAFS. The sample's thickness exceeding 300 μm, restricts the measurement of XAS spectra to the fluorescent mode. We skip characterizing the absorption edges for Ga and As atoms since the measuring signal potentially comes from the substrate, disturbing the fitting and interpretation. Furthermore, the beamline 5.2 configuration gives the highest photon flux within a limited energy range, making it suitable for study only at Bi M5- and P K-edges for this experiment.

The oxidation states of the synthesized quaternary alloy are measured using XANES at the Bi M5-edge and P K-edge and compared against their known standards. Fig. 2(a) shows Bi M5-edges for sample A (E₀ = 2599 eV) and for sample B (E₀ = 2601 eV). Additionally, Bi M5-edge for Bi₂O₃ (E₀ = 2598 eV), which theoretically corresponds to Bi³⁺, is added in the figure for comparison. Since the absorption edges of both samples are close to that of Bi₂O₃, the oxidation state of Bi in samples A and B is likely 3+. Fig. 2(b) shows P K-edges for samples A and B (E₀ = 2146 eV). Additionally, the P K-edges for Ca₃(PO₄)₂ (E₀ = 2154 eV), KH₂PO₄ (E₀ = 2154 eV), and phosphorus powder (P₄, E₀ = 2146 eV) are included in the figure for comparison. Theoretical assignments suggest that these correspond to P5+, P5+, and 0 oxidation states, respectively. It indicates that the synthesized quaternary alloy possesses a phosphorus oxidation state of 0, as the absorption energy aligns with that of P₄.

For the EXAFS, the measuring spectrum in terms of energy is converted to k-space, then to R-space by applying a Fourier transform. The R-space represents the radial distribution of neighboring atoms around the absorbing atom.

The k-space and R-space are fitted with a known model that includes parameters such as bond lengths, coordination numbers, and the Debye–Waller factor (σ²). The fitting results consider these parameters. The shifted energy (ΔE₀) should be within ±10 eV for well-fitted EXAFS. The coordination number represents the number of neighboring atoms around the absorber. Bond length (R) is the distance between the absorbing atom and its neighboring atoms. σ² value represents the Debye–Waller factor or mean square relative displacement (MSRD), which accounts for the thermal and static disorder in the distances between the absorbing atom and its neighboring atoms. Typical values of σ² for well-ordered structured structures are on the order of 0.001–0.01, while higher values indicate more disorder or higher temperature.^21,23 The amplitude reduction factor (S₀²) adjusts the overall amplitude of the EXAFS oscillations, which typically ranges between 0.7 and 1.0.^21,24R-Factor quantifies the agreement between the fitted model and the experimental data, aiming for a low value to indicate close agreement between the experimental and theoretical EXAFS spectra.²⁵

We characterize the local structures of samples A and B by using synchrotron-based EXAFS around the P K-edge. EXAFS around the Bi M5-edge is neglected because its energy is too close to the Bi M4-edge (<250 eV difference). Fig. 3(a) and (b) consecutively depict Fourier transformed of P K-edge EXAFS spectra (black spectra) for samples A and B. Fitting the spectra in R space allows revealing the local structures around the P atoms. For the fitting, we initially generate the theoretical EXAFS signal by the FEFF code performed on the cubic GaP crystal (prevalently experimentally observed structure). We execute the fitting using the theoretical signal in conjunction with the lattice constant value resulting from the X-ray diffractogram shown in Fig. 1(e).

For sample A, highly mismatched probed by XRD, cracks, and orthogonal dislocations evidenced in AFM attest to perform the diffractogram/lattice constant conversion under a fully relaxed assumption. The theoretical EXAFS signal assumes that the quaternary alloy is stoichiometric, i.e., P, As, and Bi atoms (anions) are interchangeable and are balanced by Ga atoms (cations). The Ga atoms in the first shell are intact, and the P atoms in the second shell are interchangeable with As and Bi atoms until the experimental and fitting spectra converge (red scatters in Fig. 3(a)). For sample A, the best fit is obtained when the second shell comprises 2.83 As atoms, 8.18 P atoms, and 0.99 Bi atoms (equivalent to GaAs_0.24P_0.68Bi_0.08). The resulting high-P-content quaternary alloy agrees with the XRD diffractogram. This composition, which measures the local structure of the four atomic types binding together, represents an average composition of the quaternary alloy. The discrepancy of the composition acquired using EXAFS and XPS is associated with the nature of the acquired signal. The composition estimated using XPS spectroscopy relies on all the atoms that appeared on the surface. Hence, the method is suitable only for homogeneous alloys.

For sample B, the adjacent diffraction peaks of the substrate and epitaxial layer (Fig. 1(e)) and the smooth AFM surface (Fig. 1(c)) suggest that the grown layer is pseudo-morphically synthesized on the substrate. We obtain the best fit (scatters in Fig. 3(b)) by applying the identical fitting technique reported earlier when the second shell comprises 7.95 As atoms, 3.00 P atoms, and 1.05 Bi atom (equivalent to GaAs_0.66P_0.25Bi_0.09).

To study the ex situ annealing effect of the lattice-matched GaAsPBi on GaAs (001), we cut sample B into four pieces, kept one piece as a controlled sample, and annealed the other three samples at 450 °C (sample C), 550 °C (sample D), and 650 °C (sample E). To anneal each sample, we placed it in the middle of a quartz tube, increased the T_A to the set point for one hour, dwelled at the set point for one hour, and decreased the T_A to room temperature for another hour. Thus, the complete annealing profile lasts three hours under hydrogen-containing gas (95% N₂, 5% H₂). The dissociated hydrogen atoms can neutralize dislocations and dangling bonds and free the carriers, thus improving the optical emission. Indeed, we annealed the samples using the identical chamber and annealing profiles as in the case of reported dislocations embedded GaAsPBi epilayer,²⁶ allowing us to compare the annealing effects from both studies.

Fig. 4(a)–(f) show the AFM micrographs (OM images) of latticed-matched films annealed at 450, 550, and 650 °C, respectively. The AFM images confirm that the surfaces of all the annealed samples remain unchanged from the as-grown surface (Fig. 1(c)). Their RMS roughness is in the range of 0.36–0.62 nm. For a lower magnification field of view, the OM images of annealed samples at T_A = 450 °C and 550 °C are similar to that of the as-grown sample, in agreement with the AFM results. Blistering defects emerge on the surface of sample E (T_A = 650 °C) (see Fig. 4(f)). The low density of the defects hinders us from observing them by AFM. The average diameter and density of the blisters are 41.65 μm and 1.31 × 10⁻⁵ μm⁻², respectively. The defects distributed randomly for the lattice-matched GaAsPBi/GaAs, differently from those that occurred on the surface with orthogonal dislocations reported elsewhere.²⁶

A 90°-tilted SEM micrograph of sample D (T_A = 550 °C, Fig. 4(g)) illustrates a smooth surface of the 280 nm-thick GaAsPBi epitaxial layer. The measured thickness is thicker than the 200 nm-thick prior determined by RHEED oscillations, owing to the different incorporating mechanisms during the growth and calibration explained earlier.¹⁸ 90°-tilted SEM micrographs of samples B (as-grown) and C (T_A = 450 °C), having the same structure as sample D, are not shown here. Fig. 4(h) shows a 90°-tilted SEM micrograph of sample E (T_A = 450 °C) with a blistering defect on the surface. This blistering defect has a depth of 3 μm, indicating that it destroys both the epitaxial layer and GaAs substrate.

HR-XRD scans around the symmetric (004) reflections of the annealed GaAsPBi epitaxial layers are depicted in Fig. 5. For samples C (T_A = 450 °C) and D (T_A = 550 °C), the GaAs and the quaternary alloy diffraction peaks are identical to those of the as-grown (sample B), confirming that the structures after annealing at T_A = 450 °C and 550 °C are intact. The structural stabilities agree with the surface results shown earlier. For sample E (T_A = 650 °C), the 2θ diffraction peak of GaAs is stable at 66.05° but that of GaAsPBi shifts to 66.36°. The XRD peak shift confirms that the structural degradation occurred at T_A = 650 °C, agreeing with the reported surface results. The XRD peak position suggests that the degraded film has comparatively lower As or Bi concentrations than other samples in the annealing series.

Fig. 6(a) shows normalized PL spectra of samples B (as-grown), C (T_A = 450 °C), D (T_A = 550 °C), and E (T_A = 650 °C). The PL spectra of all the samples show two luminescent peaks with energies centered around 1.32 eV and 1.47 eV, emitting from the epitaxial layer and GaAs substrate, respectively. Single Gaussian luminescence from the epitaxial layer suggests that the epitaxial layer is homogeneous, unlike sample A, which possesses multiple peak luminescence originating from the in-plane inhomogeneity.²⁶ The optical improvement of the annealing series is analyzed by comparing their normalized integrated photoluminescence (IPL) with the IPL of sample B (as-grown) (see Fig. 6(b), black bar graph). For comparison, we also summarize the optical improvement in the case of the annealed series of the inhomogeneous GaAsPBi epitaxial layers from ref. 26 (see Fig. 6(b), gray bar graph). For lattice-matched GaAsPBi/GaAs, annealing with hydrogen-containing gas can only slightly enhance the luminescence. The most luminescent improvement occurs at T_A = 550 °C, where the IPL is improved 1.3 times compared to the as-grown sample. The optimal annealing temperature (T_A = 550 °C) is identical to the case of the inhomogeneous quaternary alloy in that the IPL is improved 34 times from the as-grown condition that occurred along with the luminescent spectral feature modification.²⁶

Since the improvement comes mainly from the carriers released from the neutralized defects, the current annealing recipe can significantly improve the optical quality for the sample inherited with high defect density, for example, high orthogonal dislocation density as sample A. When the annealing temperature increases to T_A = 650 °C, the IPL decreases to 0.07 times compared to the as-grown sample. The reduction of the luminescence agrees with the degraded structural quality. After the annealing process, the XANES signal for all the annealed samples is similar to that for sample B (not shown). The oxidation state of Bi (P) for all the annealed samples C, D, and E is 3+ (0) since the Bi M5-edges (P K-edges) for the samples are at E₀ = 2601 eV (2146 eV).

Fig. 7(a)–(c) consecutively depict Fourier transformed of P K-edge EXAFS spectra (black spectra) for the annealed GaAsPBi samples C, D, and E. We execute the fitting using the theoretical signal with the lattice constant value resulting from the X-ray diffractogram shown in Fig. 5 – the identical fitting technique used earlier. The best fits (scatters in Fig. 7(a)–(c)) are obtained when the compositions of samples C, D, and E are equivalent to GaAs_0.66P_0.25Bi_0.09. The EXAFS fitting parameters for samples A, B, C, D, and E are listed in Table 1.

---

---

The room temperature band gap of GaAs_0.66P_0.25Bi_0.09 calculated by equations proposed by Luo et al.²⁷ is 1.17 eV. The GaAs_0.66P_0.25Bi_0.09 band gap at 25 K is estimated at 1.27 eV (an increase of ∼0.1 eV as the temperature decreases from 300 K to 25 K, the same as GaAs). The marked band gap position in Fig. 5(a) is within the luminescent curve but still lower than the experimental luminescent peak of 50 meV. The discrepancy is attributed either to inappropriate parameters in the fitting packages or to the equations proposed by Luo et al.,²⁷ resulting in a slight overestimation of the Bi content (or underestimation of the P content). The local structure of the lattice-matched GaAsPBi alloy does not change even at the highest annealing temperature (T_A = 650 °C), where we observe blisters on the surface and degraded luminescent quality, implying that the degradation occurred macroscopically while the local structure was intact. Here, we do not estimate the band gap of GaAs_0.24P_0.68Bi_0.08 for two reasons. First, the P concentration (0.68) is out of the application range for the equations proposed by Luo et al. (0.40).²⁷ Second, the sample is inhomogeneous, which makes the comparison between the optical and structural asserted band gaps invalid.

Conclusion

We report the synchrotron-based X-ray absorption near-edge structure of lattice-matched GaAsPBi and GaAsPBi with orthogonal dislocations on GaAs (001) acquired around the P K-edge and Bi M5-edge. Both quaternary compounds show similar oxidation states of P (0) and Bi (3+). The oxidation states of P and Bi are not affected by hydrogen-containing gas annealing even at T_A = 650 °C, where the surface is destroyed. The synchrotron-based extended X-ray absorption fine structure (EXAFS) spectra of lattice-matched GaAsPBi and GaAsPBi with orthogonal dislocations on GaAs (001) are acquired around the P K-edge. The simulated EXAFS spectra for both samples in R space are constructed based on the FEFF code performed on the cubic GaP crystal in conjunction with the experimental lattice constants obtained from the HR-XRD diffractograms. Fitting the spectra in R space with the simulated signal, granting the estimated compositions of GaAs_0.66P_0.25Bi_0.09 and GaAs_0.24P_0.68Bi_0.08 for lattice-matched GaAsPBi and GaAsPBi with orthogonal dislocations on GaAs (001), respectively. This innovative technique for finding alloy compositions applies to other complex alloy systems where a standard XRD technique cannot resolve the material composition. The ex situ annealing of lattice-matched GaAsPBi on GaAs (001) can slightly improve the luminescence (up to 1.3 times as compared with the as-grown condition at T_A = 550 °C). The luminescent efficacy drops to 0.07 times as compared with the as-grown condition when T_A = 650 °C, in agreement with the structural degradation. The finding indicates that the fabrication process for III–V-Bi layers must not exceed 600 °C to preserve the material quality. Contrarily, the annealing at high temperatures does not disturb the local structures of lattice-matched GaAsPBi on GaAs (001). The identical fitting technique performed on the EXAFS spectra shows that the estimated compositions for annealing samples are GaAs_0.66P_0.25Bi_0.09 (T_A = 450 °C), GaAs_0.66P_0.25Bi_0.09 (T_A = 550 °C), and GaAs_0.66P_0.25Bi_0.09 (T_A = 650 °C).

Author contributions

Saharat Chomdech: investigation, writing – original draft, writing – review & editing. Chalermchai Himwas: conceptualization, methodology, investigation, writing – original draft, writing – review & editing, resources. Wenich Pumee: investigation. Suphakan Kijamnajsuk: investigation, Waraporn Tanthanuch: investigation, writing – review & editing. Songphol Kanjanachuchai: resources.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was partially supported by Grants for Development of New Faculty Staff, Ratchadaphiseksomphot Endowment Fund, Chulalongkorn University (Chalermchai HIMWAS), the National Research Council of Thailand (N42A650261), and CUEE High-Performance Scholarship (Saharat Chomdech). The authors thank V. Yordsri and C. Thanachayanont for their technical support. Synchrotron beamtime was granted by Synchrotron Light Research Institute (Public organization), Thailand.

References

1. S. Tixier, et al., Molecular beam epitaxy growth of GaAs1−xBix, Appl. Phys. Lett., 2003, 82, 2245–2247 CrossRef CAS .
2. S. Francoeur, et al., Band gap of GaAs1−xBix, 0<x<3.6%, Appl. Phys. Lett., 2003, 82, 3874–3876 CrossRef CAS .
3. D. G. Cooke, F. A. Hegmann, E. C. Young and T. Tiedje, Electron mobility in dilute GaAs bismide and nitride alloys measured by time-resolved terahertz spectroscopy, Appl. Phys. Lett., 2006, 89, 122103 CrossRef .
4. Y. Zhang, A. Mascarenhas and L.-W. Wang, Similar and dissimilar aspects of III−V semiconductors containing Bi versus N, Phys. Rev. B: Condens. Matter Mater. Phys., 2005, 71, 155201 CrossRef .
5. K. Oe and H. Okamoto, New Semiconductor Alloy GaAs1-xBix Grown by Metal Organic Vapor Phase Epitaxy, Jpn. J. Appl. Phys., 1998, 37, L1283–L1285 CrossRef .
6. A. G. Norman, R. France and A. J. Ptak, Atomic ordering and phase separation in MBE GaAs1−xBix, J. Vac. Sci. Technol., B: Nanotechnol. Microelectron.: Mater., Process., Meas., Phenom., 2011, 29, 03C121 Search PubMed .
7. D. F. Reyes, et al., Bismuth incorporation and the role of ordering in GaAsBi/GaAs structures, Nanoscale Res. Lett., 2014, 9, 23 CrossRef .
8. A. Beyer, et al., Local Bi ordering in MOVPE grown Ga(As,Bi) investigated by high resolution scanning transmission electron microscopy, Appl. Mater. Today, 2017, 6, 22–28 CrossRef .
9. E. Sterzer, et al., Investigation of the microstructure of metallic droplets on Ga(AsBi)/GaAs, J. Cryst. Growth, 2014, 408, 71–77 CrossRef CAS .
10. C. R. Tait, L. Yan and J. M. Millunchick, Droplet induced compositional inhomogeneities in GaAsBi, Appl. Phys. Lett., 2017, 111, 042105 CrossRef .
11. X. Lu, D. A. Beaton, R. B. Lewis, T. Tiedje and M. B. Whitwick, Effect of molecular beam epitaxy growth conditions on the Bi content of GaAs1−xBix, Appl. Phys. Lett., 2008, 92, 192110 CrossRef .
12. W. Pan, L. Wang, Y. Zhang, W. Lei and S. Wang, MBE growth strategy and optimization of GaAsBi quantum well light emitting structure beyond 1.2 μ m, Appl. Phys. Lett., 2019, 114, 152102 CrossRef .
13. R. B. Lewis, M. Masnadi-Shirazi and T. Tiedje, Growth of high Bi concentration GaAs_1−xBi_x by molecular beam epitaxy, Appl. Phys. Lett., 2012, 101, 082112 CrossRef .
14. I. Moussa, H. Fitouri, A. Rebey and B. El Jani, Atmospheric-pressure metalorganic vapour phase epitaxy optimization of GaAsBi alloy, Thin Solid Films, 2008, 516, 8372–8376 CrossRef CAS .
15. H. Fitouri, I. Moussa, A. Rebey and B. El Jani, Study of GaAsBi MOVPE growth on (100) GaAs substrate under high Bi flow rate by high resolution X-ray diffraction, Microelectron. Eng., 2011, 88, 476–479 CrossRef CAS .
16. C. Himwas, et al., GaAsPBi epitaxial layer grown by molecular beam epitaxy, Semicond. Sci. Technol., 2020, 35, 095009 CrossRef CAS .
17. L. Nattermann, et al., MOVPE growth and characterization of quaternary Ga(PAsBi)/GaAs alloys for optoelectronic applications, Appl. Mater. Today, 2016, 5, 209–214 CrossRef .
18. C. Himwas, et al., Optical properties of lattice-matched GaAsPBi multiple quantum wells grown on GaAs (001), Semicond. Sci. Technol., 2021, 36, 045014 CrossRef CAS .
19. C. Himwas, S. Panyakeow and S. Kanjanachuchai, Optical properties of as-grown and annealed InAs quantum dots on InGaAs cross-hatch patterns, Nanoscale Res. Lett., 2011, 6, 496 CrossRef .
20. S.-W. Ryu, I. Kim, B.-D. Choe and W. G. Jeong, The effect of strain on the interdiffusion in InGaAs/GaAs quantum wells, Appl. Phys. Lett., 1995, 67, 1417–1419 CrossRef CAS .
21. B. Ravel and M. Newville, ATHENA, ARTEMIS, HEPHAESTUS : data analysis for X-ray absorption spectroscopy using IFEFFIT, J. Synchrotron Radiat., 2005, 12, 537–541 CAS .
22. S. Kanjanachuchai, M. Maitreeboriraks, C. C. Thet, T. Limwongse and S. Panyakeow, Self-assembled InAs quantum dots on cross-hatch InGaAs templates: Excess growth, growth rate, capping and preferential alignment, Microelectron. Eng., 2009, 86, 844–849 CrossRef CAS .
23. Y. Jiang, et al., Order and disorder in the local and long-range structure of the spin-glass pyrochlore, Tb₂Mo₂O₇, J. Phys.: Condens. Matter, 2011, 23, 164214 CrossRef .
24. J. Goulon, C. Goulon-Ginet, R. Cortes and J. M. Dubois, On experimental attenuation factors of the amplitude of the EXAFS oscillations in absorption, reflectivity and luminescence measurements, J. Phys., 1982, 43, 539–548 CrossRef CAS .
25. I. A. Sluchinskaya and A. I. Lebedev, An experimental and theoretical study of Ni impurity centers in Ba0.8Sr0.2TiO3, Phys. Solid State, 2017, 59, 1512–1519 CrossRef CAS .
26. C. Himwas, et al., Degradation mechanisms of annealed GaAsPBi films grown by molecular beam epitaxy, Surf. Interfaces, 2023, 40, 103031 CrossRef CAS .
27. G. Luo, K. Forghani, T. F. Kuech and D. Morgan, First-principles predictions of electronic properties of GaAs_1-x-yP_yBi_x and GaAs_1-x-yP_yBi_x-based heterojunctions, Appl. Phys. Lett., 2016, 109, 112104 CrossRef .

This journal is © The Royal Society of Chemistry 2024