Thermally oxidized porous silicon shows stable visible luminescence under laser irradiation. The photoluminescence decay has been measured in the temporal range from microsecond to millisecond. The decay curve is well fitted to a stretched exponential function. The mean lifetime of the luminescence increases with increasing the excitation pulse width. It is also found that the higher energy emission exhibits a faster decay compared with the lower energy emission at 293K, while at 18K it is reversed. The results are explained by a model in which photogenerated carriers are captured by localized radiative centers through a thermally assisted tunneling mechanism

Radiative and non-radiative recombination in porous silicon (PSi) is measured under a wide injection range by detecting of photoluminescence (PL) and free-carrier absorption (FCA) decay. We used 2.34 eV and 140 ns light pulses for carrier excitation in PSi flakes. The excited carriers were probed by focused cw IR light and the carrier concentration was calculated by using the free-carrier absorption cross section of bulk (c-Si). The results demonstrate a scaling between the total PL yield in the S-band, integrated over wavelengths, and the free carrier concentration. At lower injections the observed free carrier and PL decay follow a similar, density-independent stretched-exponential law. At injections approaching 1018 cm−3, a fast recombination component appears reducing the lifetime by a factor of ten. This component is attributed to Auger recombination of separated e-h pairs on a set of Si subclusters.

We present a detailed study of the time-resolved photo-luminescence of porous Silicon samples with different porosities providing clear evidence of anomalous relaxation behaviour of the luminescence, which follows stretched exponential decay for a variety of experimental conditions. In addition, a numerical study of the underlying transport behaviour in these disordered materials by means of Monte-Carlo simulations has been performed. Nanometer sized particles, characterised by a distribution of radiative and non-radiative recombination times, are randomly placed at the sites of a cubic lattice forming a single three dimensional percolation cluster. Charge carriers are allowed to hop between nearest-neighbour occupied sites. The competing effect between radiative and non-radiative transitions in a single nanometer particle, as well as the effects of geometrical constraints on transport due to the complex topology, are discussed and compared to experiments.

The mechanical stress induced in as-prepared porous silicon is generally compressive. Upon thermal annealing, the stress shows hysteresis. Isothermal annealing reveals second-order kinetic processes for both the stress and hydrogen effusion. The correlations between the changes in stress and porous silicon composition and structure are discussed.

The thermal conductivity of porous silicon is measured as prepared and after oxidation. The measurement method uses thermal wave propagation in the porous film. We investigate three types of porous silicon: Nanoporous p-type silicon, nanoporous n-type silicon and mesoporous p+-type silicon. The nanoporous material shows a thermal conductivity in the region of 1.2 W/mK to 1.8 W/mK as prepared and after oxidation. This value is close to silicon oxide. The mesoporous material shows a high thermal conductivity of 80 W/mK as prepared which drops to 2.7 W/mK after oxidation.

We have fabricated diode structures containing clusters of nanoscale silicon particles. The modelling of the observed multiple current steps from resonant tunnelling through the quantum states of these particles is presented. Whenever the applied voltage aligns the Fermi surface of the contact to a discrete state, conduction results. Since the clusters are connected in parallel, successive connections of parallel current paths can cause multiple current steps. These current steps can be drastically amplified by avalanche multiplication when the substrate is deeply depleted. The transport in this regime is quite nonlinear, and cannot be represented by a linear equivalent circuit. Apart from gaining a fundamental understanding of the measured I-V characteristics of our diodes, our results serve to promote new applications for quantum switches.

In order to clarify the role of the enlarged surface area of porous silicon on the electrical conductivity, we have studied the transport in mesoporous silicon layers, for which quantum confinement effects are negligible. We prepare free standing mesoporous films, from highly doped p-type Si wafers. The dark conductivity of the mesoporous layers is activated with an energy of 0.5 eV. Thermopower measurements show negative sign indicating electron conduction. The exposure of these layers to methanol vapor results in an increased conductivity and change of the thermopower magnitude. Photoconductivity measurements and the Steady-State Photocarrier Grating technique (SSPG) are used to evaluate the density of the surface states and the dynamics of the photo-excited carriers. All these results indicate that a large density of surface states exist, which results in a depletion of the free holes.

Electronic states in por-Si are investigated by the pulsed surface photovoltage technique using the por-Si/c-Si interface as probe. The values of the potentials at the por-Si/n-Si and at the por-Si/p-Si interface are -0.05 V and +0.47 V, respectively. Non-monotonous photovoltage transients related to trapping in por-Si are observed. The traps can be annealed at 150°C and generated in H20 atmosphere.

We report the first observation of the lateral photovoltaic effect in porous silicon. Contacts placed on either side of a porous silicon region develop a voltage up to several millivolts if the sample is asymmetrically illuminated. If the light spot is closer to one contact, the voltage will have one polarity; if it is closer to the other contact, the polarity will be opposite. In the case of n-type, the contact nearest the light spot is positive; for p-type, the contact nearest the light spot is negative In the region between the contacts, the photovoltage varies almost linearly with the position of the light spot, over a distance 4.5 cm across. The origin of our lateral photoeffect may be explained by the trapping of photoexcited carriers by a pair of dangling bond centers in porous silicon. In the case of p-type, the photogenerated electrons are trapped by the dangling bond states while holes diffuse away in the substrate. The situation for n-type is opposite; holes are trapped by the dangling bond states while electrons diffuse away in the substrate. This differs from the conventional lateral photoeffect, which arises under the nonuniform illumination of a junction between two layers of differing conductivities. Hamamatsu sells silicon-based position-sensitive detectors with a resolution down to 0.1 µm. The possibility of using this lateral photoeffect to characterize these dangling bond states in porous silicon as well as several possible device applications will be discussed.

Observation of light emission from porous Si has demonstrated that the optical properties of Si can be drastically altered by the quantum size effects. We have investigated the improvement of absorption properties of Si material by forming a porous Si layer. Shallow-junction commercial crystalline as well as polycrystalline Si solar cells without anti-reflective coatings have been processed into porous Si solar cells by a wet chemical etching technique. Our best results have demonstrated more than 15% improvement in short-circuit current with no change in open-circuit voltage. The performance of the porous Si solar cells has been found to be sensitive to the porous layer thickness. The efficiency can be reduced when the porous layer is relatively deep, presumably due to the penetration of pores through the shallow junction. We believe porous Si can be optimized for photovoltaic applications by properly controlling its porosity and thickness.

Porous silicon is known to be sensitive to moisture. Using an aluminium-porous p+ silicon junction, we have realized a sensor which dc current increases up to two orders of magnitude in the presence of ammoniac. We have tested a series of various gases and we show that if the dipole moment of the molecule is zero, there is no effect on the dc current. To interpret quantitatively this phenomenon, we assume that the conductivity is governed by the width of a channel resulting from the partial depletion of silicon located between two pores. This depleted region is due to the charges trapped on surface states associated with the Si-SiO2 interface where SiO2 is the native silicon oxide. When some gas is adsorbed, we propose there is a passivation of the interface states (mainly dangling bonds), leading to a decrease of the depleted region, i.e. an increase of the width of the channel and thus an increase of the current. The adsorbed gas gives a dipole layer at the surface of the pore. This layer has no influence on the depleted region. It stabilizes electrons or holes at the porous Si surface, allowing a stable charge state of the dangling bonds.

Porous silicon is investigated as an initiator for radical polymerization reactions. Surface radicals and photogenerated e−/h+ pairs in porous silicon can be used to initiate polymerization reactions. Porous silicon wafers were placed in ethanolic solutions of methyl methacrylate, N-vinylcarbazole, acrylonitrile, or styrene and either left in the dark or excited by light. Methyl methacrylate polymerized in the absence of light, presumably via silicon-based radical initiator sites at the porous silicon surface. The N-vinylcarbazole polymerized to a small extent in the dark, and to a much greater extent during white light irradiation of the porous silicon surface. Photopolymerization is thought to occur via excited state hole transfer from the porous silicon to the N-vinylcarbazole monomer. Very little to no polymerization of acrylonitrile or styrene could be detected either in the dark or under illumination.

Wide band-gap semiconductors such as TiO2, CdSe and CdS, forming the working electrode in photoelectrochemical cells, show quite different spectral responses depending on semiconductor morphology. On the basis of a theoretical model, which takes into account the collection of both majority and minority carriers, spectral responses have been simulated and compared with experimental data for both nanocrystalline and polycrystalline materials.

Experimental results showing two electrically-induced phenomena, namely the voltage-tunable electroluminescence (VTEL) and the voltage-induced quenching of porous silicon photoluminescence(QPL) are given. In both cases, a spectral shift as large as 300 nm can be recorded for an external bias variation of only 0.5V. This spectral shift is characterised by a blue-shift of the whole EL line in the case of the VTEL whereas it results from a progressive and selective quenching starting by the low-energy part of the luminescence line in the case of the QPL experiments. The origin of this spectral shift is discussed in relation with an electrically-induced selective carrier injection into the silicon nanocrystallites accompanied with an enhancement of the non-radiative recombination which might take place by an Auger relaxation process. Finally, it is shown that a partial oxidation of the porous silicon layer leads to a complete loss of the selectivity of these two phenomena. This result is qualitatively discussed by considering the voltage drop distribution between the substrate and the silicon nanocrystallites. The voltage drops are modified by the growth of the oxide layer on the nanocrystallite surface leading to a modification of the energy barriers at the crystallite boundaries.

In addition to photoluminescence and electroluminescence porous silicon is capable of emitting an ultraviolet line spectrum. This emission can be observed already at ambient conditions. We could identify this line emission as the spectrum of nitrogen. At a pressure of about 10 mbar the UV-intensity exceeds the intensity at ambient pressure about two orders of magnitude. Angular dependent spectroscopy and the light emission behaviour at lowered pressure led us to the conclusion that the silicon structures in the samples behave as sub-micrometer-sized electron guns. Dye covered glass substrates can be excited by the intense UV-light at about 5-20 mbar so that the red and green light of the dyes can easily be recognized under usual laboratory illumination. A luminous density of 240 Cd/m2 and 40 Cd/m2 could be achieved for the green luminescing ZnS:Cu,Al and the red luminescing YVO4:Eu respectively. Continuous UV-light emission could be observed for more than 1.5 hours at 2 mbar.

In applying porous Si (PS) to color display technology, it is important to fabricate light emitting devices with three primary colors. However, there have been few reports on blue and green electroluminescence (EL), and its mechanism (even the relationship between PL and EL spectra) is unclear. To obtain blue and green EL and to investigate its mechanism, we have formed PS anodized under UV illumination (UV-PS) with green photoluminescence (PL) and porous SiC with blue PL. Consequently, green and blue light emitting devices were successfully fabricated by using these materials. The observed spectra are from 350 to 750 nm with a peak of, 520 nm for ITO / UV-PS junctions and from 300 to 600 nm with a peak of 470 nm for ITO / porous SiC junctions. The EL mechanism is also discussed by reference to experimental results of comparing PL and EL spectra and of investigating the dependence of EL intensity on current.

Visible to near-infrared emission is produced from a porous silicon device under current injection. The porous silicon emitter is fabricated by selectively under-etching a p-n junction. The device is rectifying when biased across the junction, and exhibits a region of negative differential resistance (NDR) at the higher current levels. Bright red-orange emission is observed along the length of the junction in forward bias, but most of the light is emitted in the near-infrared with a peak near 1.3 µm at 30 mA. The intensity of the visible component has an exponential dependence on photon energy. The optical and electrical properties of this device are presented and possible mechanisms are discussed.

We have developed three classes of techniques to produce micron-size and submicron-size light emitting porous Si (LEPSi) patterns and to protect the rest of the wafer. In the 1st class, LEPSi lines down to 2 µm width have been made using a photoresist/silicon nitride trilayer mask, followed by anodization. PL mapping of the structures indicates that the protected regions have not been etched. Using electron beam lithography sub-0.5 micron porous Si lines have been generated. In the 2nd class, formation of porous Si is inhibited by amorphizing Si using ion implantation followed by anodization and annealing. The crystallinity and electrical properties of the implanted region have been fully characterized after annealing. Using focussed ion-beam implantation, LEPSi patterns of the order of 100 nm have been obtained. The 3rd class consists of enhancing the formation of porous Si by a low energy/low dose bombardment (ion-milling) with argon ions prior to anodization. Under appropriate conditions, we have observed a strong enhancement of the formation rate of LEPSi where bombardment took place, possibly due to the generation of a large number of defects on the wafer surface.

The quantum efficiency of electroluminescence (EL) in porous silicon structures (PSL) is practically limited by the mechanical instability of higher porosity layers. This can be overcome by adding a thin implanted p+ surface region before the formation of PSL. The resulting structure consisted of different morphologies depending on the local doping level. When the implantation doping level and diffusion depth was increased, reduced photoluminescence (PL) from the otherwise strongly emitting buried PSL of 80% porosity was obtained. Spectroscopic ellipsometry provided detailed analysis of optical transmission of PSL prepared on a boron implanted substrate. Since the spectral transmission of PSL is morphology dependent, in this novel structure the excitation wavelength can be absorbed in the upper mesoporous p+ region, which is much more transparent for the visible emission spectrum. The width and peak position as well as integral intensity of EL spectra during anodic oxidation of the novel, workable structure were not affected by the doping level in the thin surface region.

We demonstrate current induced visible light emission from Schottky junctions between aluminium electrodes and porous silicon formed by electrochemical etching of degenerate n+ -type silicon. HF concentration and anodizing current were chosen to yield preparation conditions in the transition region between electropolishing and porous silicon formation regimes. The light emitting diodes were formed by magnetron sputtering of aluminum on the porous silicon surface. Visible electroluminescence (EL) was recorded when dc or ac voltages larger than 4 V were applied between the aluminium electrodes. The visible EL appears in the dark, at the edge of the electrodes at a reverse bias of 5-6 V. The intensity of emitted light increases with applied voltage; at applied bias higher than 7 V the light emitted was observable by the naked eye at normal daylight. Compared to forward bias solid state contact porous silicon devices, the structure has an increased stability (after 100 hours of continuous operation under a 7 V reverse bias, no appreciable modification was observed in emission intensity). The main features of this electroluminescence are very similar to the ones observed under avalanche breakdown of silicon p-n junctions.