M. H. Ludwig

BRIGHT VISIBLE PHOTOLUMINESCENCE OF SPARK-PROCESSED GE, GAAS, AND SI

High‐frequency spark discharges were applied to single‐crystalline wafers of Ge, GaAs, and Si. The spark‐processed (sp‐) samples were characterized by photoluminescence (PL) and Raman measurements. Strong and stable luminescence with wavelengths centered at 416 and 525 nm was observed in sp‐Ge and sp‐Si layers, respectively, when excited with a 325 nm laser beam. A considerable blue shift of the PL (compared to the unsparked specimen) was also detected for sp‐GaAs with an average peak wavelength around 500 nm. The Raman shifts of the spark‐processed materials indicate that nanocrystals were formed, having diameters of 3.5–4 nm for Si and about 6 nm for Ge. A correlation between the PL wavelengths, the nanocrystal sizes, and the different semiconductor materials has been established based on the effective‐mass approximation. Making use of this model the nanocrystallite sizes have been found to range between ∼3 nm for Si and ∼5 nm for Ge. The related wavelengths for optical transitions confirm the PL result...

On the kinetics of growth of highly defective GaN epilayers and the origin of the deep trap responsible for yellow-band luminescence

The kinetics of growth of GaN/(0001) sapphire heteroepitaxial films have been examined in the relatively low substrate temperature range, 560–640 °C, using the reflection high energy electron diffraction (RHEED) specular reflection intensity monitoring technique. In particular, an alternate element exposure method of growth was employed in which Ga and N atoms were supplied separately (rather than simultaneously, as in conventional molecular beam epitaxy) to the substrate with the inclusion of a time delay between successive Ga flux and N flux exposures. We interpret the observed time dependent recovery of the RHEED specular reflection intensity during the time delay phases to be associated with Ga–N surface molecule migration on Ga-terminated surfaces and the activation energy for this migration process was determined to be 1.45±0.25 eV.

Optical properties of silicon-based materials : A comparison of porous and spark-processed silicon

Abstract Crystalline silicon is the principal semiconducting material and will doubtless continue to be the mainstay in the electronics industry for the foreseeable future. However, silicon is lacking the properties necessary to emit light efficiently and thus cannot be employed for optically active or optoelectronic applications. The poor optoelectronic behavior is caused by the indirect band gap that obstructs radiative transitions and restricts them to the infrared part of the spectrum. The discovery of an intense, room-temperature luminescence in the visible spectral range from porous silicon layers, prepared by electrochemical etching, thus attracted enormous attention. In addition to chemical etching, spark processing was also shown to generate a silicon-based substance that strongly photoluminesces in several visible bands at room temperature. Despite all research efforts, a vigorous debate is still ongoing about the best description for the complex phenomena of light emission in these materials, m...

Does the blue/violet photoluminescence of spark-processed silicon originate from hydroxyl groups?

Photoluminescence (PL) and Fourier transform infrared studies on annealed, spark-processed silicon have been performed. It was found that heat treatments at 600°C and higher temperatures substantially reduce the presence of hydroxyl groups as expected. Nevertheless, the PL of spark-processed Si remains intensive after heat treatments up to at least 900°C. It is therefore concluded that hydroxyl groups (silanol) are not the major cause for PL in spark-processed Si.

Spark-processing — A novel technique to prepare light-emitting, nanocrystalline silicon

The properties of spark-processed Si appear quite promising for applications like optically active components. Spark-processed Si photoluminesces strongly in the violet and in the green spectral range (dependent on processing). The response is fast (nanoseconds), and its stability against HF etching, intense laser irradiation, and annealing is remarkable. The microstructure reveals nanocrystalline Si particles which are imbedded in an amorphous matrix. An unambiguous link between these nanocrystals and photoluminescence has not been established. However, several other proposed mechanisms can be excluded based on experimental evidence.

Raman study of the relationship between nanoparticles and photoluminescence in spark-processed silicon

Spark-processed silicon (sp-Si) that strongly photoluminesces (PL) in the blue and green regions of the visible spectrum has been analyzed by Raman scattering, in order to probe a possible relationship between the presence and size of Si nanoparticles and PL emission wavelengths. Spatially resolved Raman spectra were measured across spark-processed regions, scanning areas which vary in PL intensity but not wavelength. The observed small shifts and broadenings of the Raman signals indicate the presence of Si particles having diameters of about 15 nm in the central, photoluminescing section of the sp region. Slightly smaller Si crystallites with sizes of about 8–9 nm were found in the halo region which does not contribute to PL. Furthermore, the Raman peak shifts and broadenings are essentially identical for blue and green photoluminescing sp-Si. These results suggest that the PL of sp-Si is not caused by a quantum-size effect which is coupled to the presence of Si nanoparticles.

On the formation process of luminescing centers in spark‐processed silicon

Radiative and compositional properties of spark‐processed silicon are studied by photoluminescence and x‐ray photoelectron spectroscopy measurements. Spark processing of silicon is performed in different atmospheres composed of nitrogen and oxygen. As a result of the process, room‐temperature radiative transitions occur at 2.35 eV and vary in intensity over five orders of magnitude depending on the N2/O2 ratio. After processing in pure nitrogen or pure oxygen, however, the green photoluminescence (PL) is wiped out and weak blue (2.7 eV) or orange (1.9 eV) PL bands, respectively, are discernable. The temperature‐dependent features of the 2.35 eV emission are characterized by an intensity increase in conjunction with a red shift of the peak position at lowered temperatures. A cross‐sectional study reveals that the green PL is mainly generated in a near‐surface layer having a chemical composition close to SiO2 and a nitrogen concentration below 1 at. %. Nearly no PL was observed from a deeper SiO2 layer enri...

Ferromagnetic Properties of Spark-Processed Photoluminescing Silicon

Magnetic properties of photoluminescing spark-processed silicon (sp-Si) have been investigated for the first time. Contrary to the diamagnetic signal known for bulk silicon, sp-Si displays a paramagnetic resonance as well as a ferromagnetic hysterisis loop. The paramagnetic resonance was studied using an EPR system and showed a high concentration of at least two distinct paramagnetic centers. One center can be eliminated by annealing in Ultra-High Purity nitrogen for 30 minutes at 600 °C. Measurements utilizing a SQUID magnetometer revealed that sp-Si displays ferromagnetic ordering with a saturization magnetization occuring at low fields. This is attributed to the high density of paramagnetic centers. Temperature dependent measurements were performed to establish possible links between magnetic properties and the luminescence of sp-Si.

Impact of surface stoichiometry control during the initial stages of growth on the stacking fault concentration in ZnSe epilayers grown by molecular beam epitaxy

We have determined that the surface stoichiometry maintained during the first five monolayers of ZnSe epitaxial growth can have a significant influence on the stacking fault concentration in 2 μm thick epilayers. In particular, we have been able to minimize the stacking fault concentration to a level in the 104 cm−2 range (comparable to the stacking fault concentration in the ZnSe substrates used for epitaxy) by appropriate selection of a delay time (∼30 s for a substrate temperature of 300 °C) employed during an alternate element (Zn and Se) exposure phase of growth. The delay time in question is the time elapsed between closing the Se shutter and opening the Zn shutter. We show that the surface stoichiometry (Zn to Se atomic ratio) can be tailored during the delay phase since Se thermal desorption occurs at the growth temperature in a controlled fashion from an initially Se‐terminated surface, and, it is postulated that selection of an optimum delay time corresponding to the attainment of a near‐stoichi...

Colour-switching effect of photoluminescent silicon after spark-processing in oxygen

We report on a new colour-switching effect of photoluminescing spark-processed silicon which was prepared in pure oxygen. Whereas as-prepared specimens display an orange PL peaking at 655 nm (1.89 eV), the PL changes to a more intense blue emission centred at 475 nm (2.61 eV) when the sample is subjected to a lower pressure. The same blue PL band is observed after heat treatments up to . Moreover, at annealing temperatures above a third radiative transition emerges in the infrared and eventually dominates the spectrum. A model is suggested that links ozone molecules, generated during spark-processing and incorporated into the sp-matrix, to the orange PL band. The results of heat treatments and temperature-dependent PL measurements allow for the relationship of the 2.61 eV band to oxygen deficiency centres, created by out-driven ozone molecules. The IR band is discussed in the light of luminescing properties of porous silicon and silicon oxides. The experiments demonstrate that the UV/blue (3.25 eV) and green (2.36 eV) PL bands for Si spark-processed in air cannot be related to radiative transitions in silicon oxides.