T. R. Guilinger

Ion-irradiation control of photoluminescence from porous silicon

Ion irradiation was used to pattern a region of red‐light emitting porous silicon by eliminating visible‐light photoluminescence (PL). The PL peak wavelength is approximately 735 nm and shows little dependence on the excitation‐light wavelength. The ratio of PL intensities for different excitation wavelengths was shown to be proportional to the ratio of the absorption coefficients. Below saturation, the integrated PL intensity increased linearly with excitation‐light power density.

Selective porous silicon formation in buried p+ layers

We report a systematic microstructural study of enhanced lateral porous silicon formation in the buried p+ layers of n/p+/p− and p−/p+/p− structures. We find, surprisingly, extremely selective porous silicon formation due to the thin p+ layer in both structures, despite the absence of a p‐n junction in the p−/p+/p− structure. The interface between the isolated island and the buried porous silicon layer was always located at the depth where the net p‐type dopant concentration was 1–8×1015/cm3. The observed microstructure can largely be understood in terms of a recent model for porous Si formation in uniformly doped Si, proposed by Beale et al. [J. Cryst. Growth 73, 622 (1985)]. However, we also observe, for the first time, important effects unique to a nonuniform dopant concentration.

Porous Silicon Formation in N−/N+/N− Doped Structures

This paper examines how dopant profile and anodization conditions affect the formation of buried porous silicon layers in n{sup {minus}}/n{sup +}/n{sup {minus}} doped wafers. Wafers with peak n{sup +} donor concentration {le}10{sup 18}/cm{sup 3} exhibit stray dendritic pores propagating from the n{sup +} layer into the n{sup {minus}} layers. depending on the anodization conditions these larger diameter dendritic pores can even penetrate the entire upper n{sup {minus}} layer, making it unusable for silicon-on-insulator device applications. Lower anodization voltages produce shorter dendrite lengths. Wafers with peak n{sup +} donor concentration {ge}3 {times} 10{sup 18}/cm{sup 3} exhibit negligible stray dendritic pores. In these wafers the buried porous silicon layer is confined only to areas with doping level {ge}1-2 {times} 10{sup 17}/cm{sup 3}. These results should help in optimizing n{sup {minus}}/n{sup +}/n{sup {minus}} doping profiles and anodization conditions for silicon-on-insulator device applications.

Control of photoluminescence from porous silicon

A description of ion-irradiation-induced reduction in the photoluminescence (PL) signal from porous silicon is given and a simple model which is consistent with a nanocrystalline Si structure is presented. Ion irradiation with 250 keV Ne is used to controllably reduce the integrated PL signal by 20% after a fluence of 4*1012 Ne cm-2 and completely eliminate the PL signal after a fluence of 4*1013 Ne cm-2. The use of vacuum and air annealing to recover ion-induced damage is also described, but the high temperatures for annealing cause elimination of the PL signal.

Porous silicon photoluminescence: Implications from in situ studies

Photoluminescence and Raman measurements have been performed on an anodized silicon surface in an HF/ethanol anodization solution and after replacement of this solution with water. Immediately after anodization and while resident in HF/ethanol, the porous silicon produced does not exhibit intense photoluminescence. Intense photoluminescence develops spontaneously in HF/ethanol after 18–24 h or with replacement of the HF/ethanol with water. The results are consistent with a quantum confinement mechanism in which electron‐hole pair migration to traps and nonradiative recombination dominates the de‐excitation pathways until silicon nanocrystals are physically separated and energetically decoupled by hydrofluoric acid etching or surface oxidation. Raman spectra show the development of nanometer‐size silicon crystals concurrent with intense photoluminescence. Illumination of the silicon surface during anodization tends to inhibit the formation of nanocrystalline silicon, but even very thin layers (tens of nano...

Porous silicon oxynitrides formed by ammonia heat treatment

Porous silicon and its oxide can be converted into porous silicon oxynitrides by ammonia heat treatment. For example, ammonia treatment at 1000 °C for 1 h following 850 °C, 30‐min steam oxidation of porous silicon can result in up to 40 at. % nitrogen in the porous oxynitrides. These porous silicon oxynitrides are compositionally more uniform than ammonia‐nitrided thermal oxides which exhibit nitrogen buildup at the oxide layer interfaces. However, the order of the oxidation and nitridation treatment matters: nitrided oxidized porous silicon exhibits higher electrical breakdown strength than nitrided porous silicon or oxidized nitrided porous silicon.

Photoluminescence and passivation of silicon nanostructures

A new method was used to fabricate nanometer‐scale structures in Si for photoluminescence studies. Helium ions were implanted to form a dense subsurface layer of small cavities (1–16 nm diameter). Implanted specimens subjected to annealing in a variety of atmospheres yielded no detectable photoluminescence. However, implantation combined with electrochemical anodization produced a substantial blueshift relative to anodization alone. This blueshift is consistent with the quantum confinement model of photoluminescence in porous silicon.

Multilevel Porous Silicon Formation

Oxidized porous silicon is the basis for one of the frontrunning silicon-on-insulator (SOI) fabrication techniques. Recently, it has also been demonstrated that porous silicon can be metallized to form silicon-on-conductor (SOC) structures. If a method for forming multilevel stacks of porous silicon layers (PSLs) can be developed, it should also be possible to combine the SOI and SOC techniques to form a buried, insulated conductor under single crystal silicon. In this communication, the authors report such a method for multilevel PSL formation.

Microstructure of pores in n+ silicon

Abstract The structure of pores in n−1/n+/n− silicon structures has been studied by cross-section transmission electron microscopy. Under the experimental conditions examined, the pore directions in the n+ layer follow the current path and do not show crystallographic preference. Stray pores were observed in the n− layer and they appear to grow along 〈100〉 directions. By using cross sections transverse to the pore length, we have obtained end-on views that show that the pore walls tend to facet along {111} planes. We have also observed wafer surface faceting on {113} planes as a result of the anodization process.

Thin-foil electrochemical cells: High-sensitivity fusion tests andin-situ ion beam measurements of deuterium loading

Electrochemical cells constructed with a thin Pd or Ti foil electrode mounted at one wall of the cell have been used both to test for the existence of “cold fusion” and to measure directly D∶Pd loading ratios in an operating cell. The first type of experiment used a surface-barrier particle detector positioned a few millimeters from the foil to provide a very sensitive monitor for possible fusion-generated protons at 3.02 MeV. The detection limit for this arrangement is estimated to be 10−24 fusions/deuterium/s, assuming a bulk fusion effect. These experiments included cells with 5- and 25-μm-thick Pd foils, 10-μm Ti foils, parallel experiments with 0.1M LiOD (heavy water) in one cell and LiOH (light water) in another, current densities up to 0.5 A/cm2, and run times as long as 22 days. No evidence for fusion products was seen. The second type of experiment using these cells, both as an adjunct to the fusion tests and to provide new information, was the use of external beam nuclear reaction analysis to monitor directly the loading and unloading of deuterium in the foil of an operating cell. Using a 1.5-MeV3He ion beam in air, the deuterium in the outer 2 μm of the exposed Pd foil was measured for the first time using the D(3He,p) nuclear reaction. The maximum D∶Pd ratios observed using this technique were 0.8–0.9.