K. P. Homewood

A silicon/iron-disilicide light-emitting diode operating at a wavelength of 1.5 μm

Although silicon has long been the material of choice for most microelectronic applications, it is a poor emitter of light (a consequence of having an ‘indirect’ bandgap), so hampering the development of integrated silicon optoelectronic devices. This problem has motivated numerous attempts to develop silicon-based structures with good light-emission characteristics, particularly at wavelengths (∼1.5 μm) relevant to optical fibre communication. For example, silicon–germanium superlattice structures can result in a material with a pseudo-direct bandgap that emits at ∼1.5 μm, and doping silicon with erbium introduces an internal optical transition having a similar emission wavelength, although neither approach has led to practical devices. In this context, β-iron disilicide has attracted recent interest as an optically active, direct-bandgap material th might be compatible with existing silicon processing technology. Here we report the realization of a light-emitting device operating at 1.5 μm that incorporates β-FeSi2 into a conventional silicon bipolar junction. We argue that this result demonstrates the potential of β-FeSi2 as an important candidate for a silicon-based optoelectronic technology.

Prediction of amorphous phase stability in the metal–silicon systems

The stability of the amorphous phase with respect to the liquid phase in metal–silicon systems is modeled thermodynamically as a second-order phase transformation. The glass transition temperature of amorphous silicon is estimated according to the experimentally determined heat of crystallization and the Third Law of Thermodynamics. The feasibility of the model has been demonstrated using the Pd–Si, Co–Si, and Au–Si systems as examples. The predicted glass transition temperatures and heat of formation of the amorphous phase are consistent with available experimental data. The predicted amorphization stabilization at low temperatures in the Co–Si systems agrees with experimental observations.

Optical absorption study of ion beam synthesized polycrystalline semiconducting FeSi2

Ion beam synthesized polycrystalline semiconducting FeSi2 on Si(001) has been investigated by transmission measurements at temperatures between 10 and 300 K. The existence of a minimum direct band gap was demonstrated and its variation with the temperature was studied by means of a three‐parameter thermodynamic model and the Einstein model. Band tail states and states on a shallow impurity level were found to give rise to the absorption below the fundamental edge. The presence of an Urbach exponential edge was shown and the temperature dependence of the Urbach tail width was also studied based on the Einstein model. A strong structural disorder associated with grain boundaries between and within the FeSi2 grains and their related defects was found to be the dominant contribution at room temperature.

Thermal quenching of the photoluminescence of InGaAs/GaAs and InGaAs/AlGaAs strained‐layer quantum wells

Photoluminescence in InGaAs/GaAs strained‐layer quantum wells is strongly quenched by temperatures above 10–100 K, depending on the well width. Analysis of this dependence shows that the quenching mechanism is thermal activation of electron‐hole pairs from the wells into the GaAs barriers, followed by nonradiative recombination through a loss mechanism in bulk GaAs. The addition of Al to the barriers to improve confinement eliminates loss through this route but introduces another loss mechanism, characterized by an activation energy independent of well width and with a smaller pre‐exponential factor.

Interdiffusion in InGaAs/GaAs quantum well structures as a function of depth

Interdiffusion in InGaAs/GaAs quantum wells has been studied using photoluminescence to follow the development of the diffusion with time in a single sample. Two distinct regimes are seen; a fast initial diffusion and a second steady‐state diffusion. The steady‐state diffusion was found to be dependent on the depth of the quantum well from the surface and to correlate with published data on the indiffusion of gallium vacancies into gallium arsenide.

Visible photoluminescence at room temperature from microcrystalline silicon precipitates in SiO2 formed by ion implantation

We have investigated the photoluminescence of microcrystalline silicon formed in SiO2 layers by ion beam synthesis. 28Si+ ions over the dose range 1 × 1017 to 6 × 1017 cm−2 at energies of 150 keV and 200 keV were implanted into thermal oxide. Samples were annealed in a halogen lamp furnace at temperatures of 900°C, 1100°C and 1300°C for times between 15 and 120 min. The implanted layers were analysed by Rutherford Backscattering Spectroscopy (RBS), Cross-Sectional Transmission Electron Microscope (XTEM) and Photoluminescence (PL) (80 K to 300 K) using an Ar laser of 488 nm wavelength. Room temperature (300 K) visible photoluminescence has been observed from all the samples. XTEM confirms the existence of Si microcrystals (within the SiO2 layers), which typically have a diameter within the range of 2–15 nm. The luminescence peak wavelength was about 600 nm or 800 nm, depending upon processing. Changes in the peak wavelength and intensity from these samples and other samples in which the crystallites were reduced in size by thermal oxidation, show trends which are generally consistent with quantum confinement, however, other mechanisms cannot be ruled out.

On the origin of the 1.5 μm luminescence in ion beam synthesized β-FeSi2

In this letter we present photoluminescence results on β‐FeSi2/Si using excitation energies above and below the silicon band gap. These results show that the luminescence emission observed at 1.5 μm can be firmly attributed to band edge related emission from the β‐FeSi2. This result confirms the potential of β‐FeSi2 as a strong contender for a silicon compatible optoelectronics technology that matches the conventional optical fiber transmission wavelength at 1.5 μm.

Amorphous-iron disilicide: A promising semiconductor

We report here the synthesis and the measurements of the microstructural and optical properties of a promising semiconductor, amorphous-iron disilicide. The material was obtained by ion-beam mixing of Fe layers on Si, with Ar8+ ions, at 300 °C. Optical absorption measurements indicate a semiconductor with a direct band gap of 0.88 eV. The significance of this discovery is that it demonstrates the existence of such a material. It should be possible to synthesize by other techniques and could be applied in large-area electronics.

Optical properties and phase transformations in α and β iron disilicide layers

Abstract Ion beam synthesis (IBS) has been used to fabricate semiconducting β-FeSi2 and metallic α-Fe0.82Si2. For all of the doses studied a photoluminescence (PL) signal is observed at 1.54 μm. This signal is first seen after annealing at 800°C and increases in intensity with a commensurate decrease in full width half maximum (FWHM) as the anneal temperature is increased up to 920°C. Likewise the intensity increases and FWHM decreases as the anneal time at 920°C is increased up to 18 h. Optical absorption measurements reveal a linear relationship between the square of the absorption coefficient and the incident photon energy, indicating a direct allowed transition from a semiconductor (β-FeSi2) with a band gap of about 0.87 eV. After annealing at 1000°C no PL or absorption is observed in this spectral region; this is because a thicker, conducting layer of α-Fe0.82Si2, containing ~ 18% Fe vacancies has then been formed. If an α-Fe0.82Si2 layer is subsequently annealed below the phase transition temperature (~ 950°C) then the PL signal reappears as the layer is largely reconverted back to the s-phase.

THE EFFECTS OF ION-IMPLANTATION ON THE INTERDIFFUSION COEFFICIENTS IN INXGA1-XAS/GAAS QUANTUM-WELL STRUCTURES

Photoluminescence coupled with repetitive thermal annealing has been used to determine the diffusion coefficients for intermixing in InxGa1−xAs/GaAs quantum wells and to study the subsequent effects of ion implantation on the intermixing. It is shown that following ion implantation there is a very fast interdiffusion process, which is independent of the implanted ion and that is thought to be due to the rapid diffusion of interstitials created during the implantation. Following this rapid process, it was found that neither gallium nor krypton ions had any effect on the subsequent interdiffusion coefficient. Following arsenic implantation in addition to the initial damage related process, an enhanced region of interdiffusion was observed with a diffusion coefficient that was an order of magnitude greater than that of an unimplanted control wafer. This enhanced process is thought to be due to the creation of group III vacancies by the arsenic atoms moving onto group V lattice sites. This fast process was pr...