Table of contents

In this topical issue of Semiconductor Science and Technology, 18 invited papers have been devoted to reviewing the current status of research and development in III–N–V semiconductor alloys. Leading researchers in this field discuss the growth and processing, experimental characterization, theoretical understanding, and device design and fabrication of this recently developed class of semiconductor alloys.

When using molecular beam epitaxy to grow GaNAs and GaInNAs with excellent crystallinity, nitrogen radicals are used as the nitrogen precursor. To suppress the generation of non-radiative centres in as-grown crystals and to eliminate them, the conditions of both growth and post-annealing have been optimized. It has been found that a relatively low growth temperature, a high growth rate and a high V/III ratio are effective for improving the crystallinity of GaInNAs. This result is possibly due to the elimination of the phase separation, because the excess atomic migration on the growing surface is suppressed under these growth conditions. It has also been found that additional rapid thermal annealing at a relatively high temperature around 750 °C greatly improves the crystallinity of GaInNAs, i.e. up to the same level as conventional III–V GaInAs.

The growth of GaNAs and related alloys has been studied using metal–organic molecular beam epitaxy. The lattice mismatch of GaNAs layers grown on GaAs induces lattice relaxation, the details of which are investigated with a mapping measurement of x-ray diffraction. The experimental results are compared to the calculated value of the critical thickness for the generation of misfit dislocations. It is shown that the surface morphology changes with the nitrogen (N) incorporation and, especially, wire-like surfaces associated with lateral compositional modulation are observed in a certain range of N compositions and with a thickness of more than 1 μm. The effects of indium and selenium doping on the N incorporation in GaNAs alloys are also discussed.

The key issues for growing III–V compound layers, free of structural defects, on Si substrates are clarified. The technologies for overcoming the fundamental problems have been developed. As a result, it has been clarified that dislocation-free III–V–N alloys can be grown on Si substrates whose lattice constants are matched to those of Si. Device structures of the GaAsPN/GaPN quantum well structure and the Si/GaPN/Si structure have been successfully grown on a Si (100) substrate covered with a thin GaP initial layer. The grown layers and hetero-interfaces contained no threading dislocations and no misfit dislocations, respectively. Neither stacking faults nor anti-phase domains were observed. A key issue for application to novel devices is the increase in nitrogen composition without degrading optical and electrical properties.

III–N–V semiconductors are promising materials for use in next-generation multijunction solar cells because these materials can be lattice matched to substrates such as GaAs, Ge and Si, with a range of bandgaps that are complementary to those of other III–V semiconductors. Several potentially high-efficiency multijunction photovoltaic device designs using III–N–V materials are discussed. The main roadblock to the development of these solar cell devices is poor minority-carrier transport in the III–N–V materials. The present understanding of the material properties of GaInNAs lattice matched to GaAs and GaNPAs lattice matched to Si is reviewed.

Growth and properties of GaNAsSb alloys are investigated and compared with those of other dilute III–N–V alloys. Similar properties are observed including very high bandgap bowing, carrier localization at low temperature, sensitivity to thermal annealing and passivation of N-related electronic states by hydrogen. On the other hand, we point out some features of this alloy system and evaluate its potential for device applications. Probably, GaNAsSb can achieve emission at longer wavelengths than GaInNAs alloys grown to date. Its conduction- and valence-band offsets can be independently tuned by adjusting the N and Sb composition, respectively. Since this compound has a single group III element, its electronic structure should be less dependent on alloy configuration than GaInNAs.

Investigations on the synthesis of group III–N_x–V_1−x alloy including GaN_xAs_1−x, InN_xP_1−x and Al_yGa_1−yN_xAs_1−x using N ion implantation followed by rapid thermal annealing are reviewed. The fundamental band-gap energy for the ion beam synthesized III–N_x–V_1−x alloys is found to decrease with increasing N-implantation dose and can be quantitatively described by the anticrossing interaction between the localized N-states and the extended states of the semiconductor matrix. N activation efficiencies in these N ion synthesized alloy films are found to be low, ∼10% for GaN_xAs_1−x and ∼20% for InN_xP_1−x. A preliminary study showed that using pulsed laser melting followed by rapid thermal annealing greatly enhanced the N activation efficiency (∼50%) in N-implanted GaAs. The N-induced conduction band modification also results in an enhancement of the maximum free electron concentration in GaN_xAs_1−x. A maximum free electron concentration as high as 7 × 10¹⁹ cm⁻³ was observed in heavily Se-doped Ga_1−3xIn_3xN_xAs_1−x (x = 0.033) films, more than 20 times larger than that observed in GaAs films grown under similar conditions. A similar increase in free electron concentration was also achieved in a S-implanted GaN_xAs_1−x thin film. Combining the ion synthesis of diluted nitrides and S implantation doping techniques, we realized a large increase in the electrical activation of S co-implanted with N in GaAs within a thin near-surface region (∼500 Å), indicating the formation of a heavily doped thin diluted GaN_xAs_1−x alloy layer with x ∼ 0.3%. This result has important practical implications on the fabrication of low-resistance, non-alloyed ohmic contacts to n-type GaAs.

The role of hydrogen in altering the electronic properties of the In_xGa_1−xAs_1−yN_y/GaAs system has been investigated by photoluminescence (PL) spectroscopy. Several heterostructures whose nitrogen concentration, y, spans from the dilute (0.0001 ≤ y ≤ 0.001) to the alloy (0.01 ≤ y ≤ 0.052) limit have been studied. The most remarkable effects observed are a quenching of the PL lines related to exciton recombination in N complexes in the dilute limit, and a bandgap blueshift of the N-containing material towards that of the N-free reference samples in the alloy limit. Differences and similarities found between In-free and x = 0.25–0.41 samples are highlighted. In all cases, the system fully recovers by thermal annealing the optical properties it had before hydrogenation. These behaviours can be accounted for by the formation of N–H bonds, which leads to an effective electronic passivation of the N atoms in the lattice. An analysis of the annealing experiments provides some clues on the geometry of the N complexes in the dilute limit as well as an estimate of the N–H bond strength in both dilute and alloy limits. All these results show that the charge distribution around the N atoms maintains in the alloy limit the strongly localized character it has in the impurity limit.

In this paper, we carry out a comprehensive review of the nitrogen-induced modifications of the electronic structure of Ga_1−yIn_yN_xAs_1−x alloys. We study in detail the behaviour of the conduction-band effective mass as a function of Fermi energy, nitrogen content and pressure. From measurements of the plasma frequency for samples with different electron concentrations we have determined the dispersion relation for the lowest conduction band. We have also studied composition, temperature and pressure dependent optical absorption spectra on free-standing layers of Ga_1−yIn_yN_xAs_1−x (0 ≤ x ≤ 0.025 and 0 ≤ y ≤ 0.09) lattice-matched to GaAs. Spectroscopic ellipsometry measurements performed in a wide photon energy range from 1.5 to 5.5 eV have been used to determine the energy dependence of the dielectric function as well as the energies of E₁, E₀' and E₂ critical point transitions. Experiments have shown that nitrogen has a large effect on the dispersion relations and on the optical spectra for the conduction-band states close to the Γ point. A much smaller effect has been observed for X and L minima as well as for the valence-band states. We have compared our results with other available experimental data. The results are analysed in terms of the analytical band anti-crossing model as well as the local density approximation calculations and empirical pseudopotential models.

A brief review on our present knowledge of optical and magneto-optical properties of III–V–N alloys, in particular, Ga(In)NAs alloys with low nitrogen compositions is given. The main attention is focused on fundamental electronic parameters of the Ga(In)NAs alloys as well as key material-related issues which are relevant to device applications, such as identification of dominant recombination processes in the alloys, compositional dependence of electron effective mass and band alignment in Ga(In)NAs-based heterostructures.

Recent developments in Raman and resonant Raman studies of GaAs_1−xN_x are discussed with respect to the nitrogen localized vibrational mode and the abnormal behaviour of this material. The normalized Raman intensity of nitrogen LVM with respect to that of GaAs–LO phonon as well as ω_LVM exhibits a linear dependence on the nitrogen concentration for x ≤ 0.03, providing a useful calibration to determine the nitrogen composition in the ternary alloy GaAs_1−xN_x. The Raman intensity and linewidth resonances of the LO(Γ) phonon provide strong evidence for the formation of nitrogen-induced resonance states above the conduction-band minimum. Various phonons at the L and X zone boundaries not only emerge as distinct and sharp Raman features for excitations near the E₊ transition but also exhibit the same intensity resonance enhancement as observed for the zone-centre phonons, LO(Γ) and TO(Γ), which provides strong evidence of significant L and X components in the wave function of the nitrogen-induced E₊ state in GaAs_1−xN_x.

The unusual N-induced band formation and band structure of Ga(N, As) and (Ga, In)(N, As) alloys are also reflected in the electronic structure of quantum wells (QWS) and device structures containing these non-amalgamation-type alloys. This review is divided into three parts. The first part deals with band structure aspects of bulk Ga(N, As) and motivates the possibility of a k · p-like parameterization of the band structure in terms of the level repulsion model between the conduction band edge of the host and a localized N-level. The second part presents experimental studies of interband transitions in Ga(N, As)/GaAs and (Ga, In)(N, As)/GaAs QW structures addressing band offsets, electron effective mass changes and an intrinsic mechanism contributing to the blueshift of the (Ga, In)(N, As) band gap on annealing. The observed interband transitions can be well described using a ten-band k · p model based on the level repulsion scheme. The third part deals with (Ga, In)(N, As)-based laser devices. The electronic structure of the active region of vertical-cavity surface-emitting laser and edge-emitter laser structures is studied by modulation spectroscopy. The gain of such structures is measured by optical methods and analysed in terms of a model combining the ten-band k · p description of the band structure and generalized Bloch equations.

We present a unified picture of our studies of the nitrogen-specific properties of InGaAsN. Nitrogen-induced bandgap reduction and symmetry-breaking are examined with band-structure calculations and pressure-dependent optical measurements for an idealized, 'ordered' InGaAsN alloy. Including disorder and large conduction band random alloy fluctuations, we describe photoluminescence linewidths and electron mobilities and propose a N-induced mobility-edge. In transport and device experiments, extrinsic material properties (inhomogeneities, defects, etc) mask any intrinsic electron localization.

We review the empirical pseudopotential method and its recent applications to the III–V nitride alloys GaAsN, GaPN, GaInAsN and GaAsPN. We discuss how studies using this method have provided an explanation for many experimentally observed anomalous nitride phenomena, including sharp photoluminescence lines in dilute alloys, high effective masses, Stoke's shift between emission and absorption in higher concentration alloys for GaAsN and GaPN ternaries. We also discuss predictions of unusual effects that remain to be experimentally discovered in GaInAsN quaternaries and complex GaAsPN solid solutions.

In this paper we review the basic theoretical aspects as well as some important experimental results of the band anticrossing effects in highly electronegativity-mismatched semiconductor alloys, such as GaAs_1−xN_x and In_yGa_1−yAs_1−xN_x. The many-impurity Anderson model treated in the coherent potential approximation is applied to these semiconductor alloys, in which metallic anion atoms are partially substituted by a highly electronegative element at low concentrations. Analytical solutions of the Green's function provide dispersion relations and state broadenings for the restructured conduction bands. The solutions also lead to the physically intuitive and widely used two-level band anticrossing model. Significant experimental observations, including large bandgap reduction, great electron effective mass enhancement and unusual pressure behaviour of the bandgap, are compared with the predictions of the band anticrossing model. The band anticrossing model is extended over the entire Brillouin zone to explain the pressure behaviour of the lowest conduction band minimum in GaP_1−xN_x. Finally, we show that the band anticrossing can also account for the large bandgap bowing parameters observed in GaAs_xSb_1−x, InAs_ySb_1−y and GaP_xSb_1−x alloys.

We review how the tight-binding method provides a particularly useful approach to understand the electronic structure of GaInNAs alloys, and use it to derive a modified k·p model for the electronic structure of GaInNAs heterostructures. Using the tight-binding model, we first confirm that N forms a resonant defect level above the conduction band edge in Ga(In)As. We show that the interaction of the resonant N level with the conduction band edge accounts for the strong bandgap bowing observed in GaInN_xAs_1−x, in agreement with experimental analysis but contrary to some theoretical interpretations. We then use a Green function model to derive explicitly the two-level band-anti-crossing model describing the interaction between the resonant states and the conduction band edge in ordered Ga(In)N_xAs_1−x. We extend the Green function model to show that the conventional k·p model must be modified to include two extra spin-degenerate nitrogen states, giving a 10-band k·p model to describe the band structure of GaNAs/GaAs and related heterostructures. We describe how this 10-band model provides excellent quantitative agreement with a wide range of experimental data and finally discuss briefly the effects of disorder on the electronic structure in dilute nitride alloys.

Research to realize long-wavelength, GaInNAs quantum well lasers has been intense in the past three years. The results have been very promising considering the relative immaturity and challenges of this new materials system. This paper reviews both the materials challenges and progress in growth of the metastable GaInNAs alloys required to reach the 1.3–1.55 μm communication wavelengths and the challenges and progress in device design for both vertical-cavity surface-emitting lasers and higher power edge-emitting lasers.

We review the status of InGaAsN-based vertical-cavity surface-emitting lasers (VCSELs) emitting in the wavelength range 1.2–1.3 μm and compare them with similar devices that have been realized using other approaches. To prove the potential of InGaAsN-based VCSELs, we present our results for monolithically MBE- and MOVPE-grown and electrically pumped VCSELs on GaAs substrates. Our MBE-grown devices emit at a wavelength of up to 1305 nm with cw output power at room temperature exceeding 1 mW and a threshold current of 2.2 mA. With an oxide-confined current aperture of about 5 μm diameter, they emit up to 700 μW in single-mode operation at room temperature. Bit-error rates of less than 10⁻¹¹ are achieved for transmission over 20.5 km of standard single-mode fibre at 2.5 Gbit s⁻¹. Our MOVPE-grown VCSELs with a similar device structure emit single mode at a wavelength of 1293 nm with a cw output power of 1.4 mW and a threshold current of 1.25 mA at room temperature. In back-to-back transmission, we reach a data rate of 10 Gbit s⁻¹, proving the feasibility of high-speed data transmission using InGaAsN VCSELs.

Use of GaInNAs in the base of heterojunction bipolar transistors (HBTs) on GaAs substrates allows a reduction of the turn-on voltage, V_be,on, of the devices, facilitating their use in applications with low power supply voltage (particularly battery operated power amplifiers for mobile communications). Using GaInNAs with N content below 2% and In content of 1–20%, HBTs have been demonstrated with V_be,on values lower by 25–400 mV than those of conventional GaAs-based HBTs. The GaInNAs base regions exhibit lower diffusion length than conventional GaAs bases, which reduces current gain and detracts from high-frequency performance, as well as higher base sheet resistance. These adverse effects can be mitigated by proper design tradeoffs of base thickness and nitrogen composition, as well as by compositional grading in the base to provide a built-in quasi-electric field to assist electron transport.