Search

The temperature dependences of the specific contact resistance of silicon ρc with a doping step are measured experimentally and described theoretically. The measurements are performed in the temperature range from 4.2 to 380 K. It is established that the contacts of the studied Au–Ti–Pd–n+–n-Si structures are ohmic. It is shown that minimal ρc is implemented at T = 75 K. Its value rises both with a decrease in temperature (due to the freezing effect) and with an increase in temperature (due to the electron-enriched layer at the boundary with the bulk material). It is established that the bulk electron concentration strongly decreases in the near-contact region in a layer with a thickness on the order of one micron due to the compensation of silicon by deep acceptors appearing because of the formation of a rather high vacancy concentration during stress relaxation and the appearance of a high dislocation density, as well as due to their diffusion from the contact after heating to 450°C. The data on the occurrence of vacancy-type defects are confirmed by X-ray measurements. The dislocation density in the studied structures is also estimated from X-ray measurements. [ABSTRACT FROM AUTHOR]

A theory is developed, which describes the experimental external quantum efficiency, EQE(λ) as a function of photon wavelength for structured Si-based solar cells. Short-circuit current density as a function of base thickness, d, is calculated for the high-efficiency solar cells with the photoconversion efficiency η ≥ 25% The procedure allows one to carry out a complete optimization of such solar cells, in particular, to find the optimal base thickness, dopt. [ABSTRACT FROM AUTHOR]

The time dependences of the key characteristic of a-Si:H solar cells over daylight hours are theoretically simulated. The model is used to calculate the time dependences for an arbitrary geographic latitude in the interval 30°–60° and arbitrary day of the year. The calculated results are illustrated for a geographic latitude of 45° and equinox. The relative variations in the characteristics of the a-Si:H solar cells are valid with a relatively high accuracy for the solar cells based on alternative semiconductors provided that their efficiency ranges form 7 to 20%. [ABSTRACT FROM AUTHOR]

The annual dependences of the powers and energies generated by the unit area of a solar cell (SC) are calculated for a-Si:H-based SCs operating at latitudes of 45°N, 50°N, 55°N, and 60°N and in some geographical localities of Russia. Normalization of these dependences gives an idea about the corresponding annual dependences for SCs based on other semiconductors. Combined with the data on the average number of sunny days in a year (or the total duration of sunshine per year) for a specific region in Russia, this information makes it possible, in particular, to judge about the prospects for constructing solar power plants in these regions. As a result, the regions in Russia for which the excess over the average values of electrical energy generated by solar power plants may reach 24% are determined. [ABSTRACT FROM AUTHOR]

The temperature dependences of the contact resistance ρ c( T) of ohmic Pd-Ti-Pd-Au contacts to n-GaN and n-AlN wide-gap semiconductors with a high dislocation density are studied. The dependences ρ c( T) for both contacts contain portions of exponential decrease ρ c( T) and very weak dependence ρ c( T) at higher temperatures. Furthermore, a plateau portion ρ c( T) is observed in the low-temperature region for the Au-Pd-Ti-Pd- n-GaN contact. This portion appears only after rapid thermal annealing (RTA). In principle, the appearance of the plateau portion can be associated with preliminary heavy doping of the near-contact region with a shallow donor impurity and with doping during contact fabrication as a result of RTA, if the contact-forming layer contains a material that is a shallow donor in III–N. The dependences obtained are not explained by existing charge-transport mechanisms. Probable mechanisms explaining the experimental dependences ρ c( T) for ohmic contacts to n-GaN and n-AlN are proposed. [ABSTRACT FROM AUTHOR]

Layers grown by magnetron deposition of Si and SiO on a p-type silicon substrate and containing silicon nanocrystals in the oxide matrix have been studied by the method of temperature dependences of the capacitive photovoltage. The effect of the substrate orientation and natural oxidation preceding high-temperature annealing that results in the formation of Si nanocrystals in the SiO matrix on the layer-substrate interface characteristics is studied. The density of fast interface states trapping majority carriers was estimated. It is found that structural changes occur at the layer-substrate interface in the case of a (111) substrate and are caused by stresses appearing upon cooling. It was shown that natural oxidation of the deposited layer, preceding high-temperature annealing, causes an increase in the charge trapped in the oxide. [ABSTRACT FROM AUTHOR]

Mechanism of charge transport in a diode of a silicon carbide’s Schottky barrier formed by a quasi-amorphous interstitial phase TiB x on the surface of n-6 H-SiC (0001) single crystals with an uncompensated donor (nitrogen) concentration of ∼1018 cm−3 and dislocation density of ∼(106–108) cm−2 has been studied. It is demonstrated that, at temperatures T ≲ 400 K, the charge transport is governed by the tunneling current along dislocations intersecting the space charge region. At T > 400 K, the mechanism of charge transport changes to a thermionic mechanism with a barrier height of ∼0.64 eV and ideality factor close to 1.3. [ABSTRACT FROM AUTHOR]

In quite realistic approximations, the theory of photoconversion in Si solar cells with the concentrated illumination is developed for the rear and standard arrangement of current-collecting contacts. The difference between the effect of heating on the efficiency of solar cells with the rear and standard geometries is taken into account. A comparative analysis of the results obtained for two mentioned geometries is carried out. It is shown that the Si solar cells with the rear metallization, in principle, can possess higher photoconversion efficiency. The agreement of the developed theory with the experiment is obtained. [ABSTRACT FROM AUTHOR]

The coefficient of nonradiative excitonic recombination by the Auger mechanism involving deep-level centers in n-Si was determined by comparing the theoretical dependence of the effective bulk lifetime on the doping level with the experimental dependence. It is shown that this mechanism controls the bulk lifetime in silicon at doping levels on the order of or above 1016 cm−3. This mechanism is more pronounced at shorter bulk lifetimes τ v0 and low doping levels. The dependences of the bipolar diffusion length in n-Si on the doping level (using the parameter τ v0) were calculated. [ABSTRACT FROM AUTHOR]

The results of the analysis of variations in the radiative recombination coefficient with varying doping level and concentration of excess electron-hole pairs are reported. It is shown that, along with the effect of narrowing of the band gap calculated in the many-electron approximation, the effect of screening of the Coulomb interaction responsible for the decrease in the excition binding energy should be taken into account. Both effects produce similar trends and decrease the radiative recombination coefficient with increasing levels of doping or injection. The contributions of excitonic radiative recombination and band-to-band radiative recombination to the total radiative recombination coefficient are separated from each other. It is shown that, in the region of room temperature, both contributions are comparable, while at liquid-nitrogen temperature, the excitonic component dominates over the band-to-band component. The results obtained by refined calculations of the limiting value of the internal quantum yield of electroluminescence for the silicon diodes and p-i-n structures are presented. It is shown that the internal quantum yield of electroluminescence can be as high as 14%. However, this values sharply decreases with increasing surface recombination rate and decreasing lifetime of excess charge carriers in the bulk. [ABSTRACT FROM AUTHOR]

The exciton binding energy and the energies of radiative excitonic transitions in the separate SiOx-Si-SiOx quantum wells are calculated in the effective-mass approximation with the quadratic dispersion relation. Along with the real finite offsets of the bands in such quantum well structures, the effect of dielectric enhancement of the exciton binding energy due polarization of the heterointerfaces is taken into account. In addition, the dependence of the zero-phonon radiative excitonic recombination time on the width of the SiOx-Si-SiOx quantum well is calculated. This dependence exhibits unsteady (oscillating) behavior, which is caused by the indirect band gap of the silicon material. It is shown that the theoretically calculated energies of the radiative excitonic transitions in the SiO2-Si-SiO2 quantum wells match the experimentally determined energies for the quantum wells whose widths are larger than 1.5 nm. Good agreement between the theoretically calculated and experimental spectral dependences of photoluminescence in the SiO2-Si-SiO2 quantum wells is attained. [ABSTRACT FROM AUTHOR]

Junction-photovoltage pleochroism is studied in crystalline silicon under the conditions of a conductivity anisotropy induced by a uniaxial compressive strain. Polarization modulation of light has been used: the samples are excited by linearly polarized light, with the polarizations periodically alternating with respect to the optical axis. The spectral characteristics obtained in such a way represent the polarization difference of the photovoltages, which depends on the light absorbance. A heavy dependence of the spectrum shape on the type of p–n junctions, which differ in relation to base parameters and emitter technologies, is detected. An analysis of the spectra shows that the condition of physical differentiation with respect to absorbance is satisfied only in p–n junctions with a negligible space-charge thickness. © 2005 Pleiades Publishing, Inc. [ABSTRACT FROM AUTHOR]

A new concept is suggested to explain the effect of dislocations on electroluminescence in silicon diodes. This concept is based on consideration of the spatial correlation between injected electrons and holes that recombine inside a dislocation core. This correlation leads to an increase in the probability of radiative recombination for electron–hole pairs. Two cases are analyzed. In the first case, the resulting current is controlled mainly by tunneling of electrons and holes along dislocations, which is followed by electron–hole recombination under the conditions of barrier lowering. In this situation, electroluminescence is not related to the fundamental absorption edge and the energy position of the electroluminescence band shifts to shorter wavelengths as either the temperature decreases or the applied voltage increases. In the second case, the diffusion-related component is prevalent in the total current. The radiative recombination of electron–hole pairs occurs in quasi-neutral regions and the electroluminescence exhibits an edge-related character. It is shown that the suggested mechanism may be responsible for a substantial enhancement of both the electroluminescence intensity and the quantum efficiency in silicon diodes with dislocations if the Shockley–Read–Hall lifetime is shorter than 10–3 s. © 2005 Pleiades Publishing, Inc. [ABSTRACT FROM AUTHOR]

The ultimate quantum efficiency of electroluminescence in silicon diodes and p–i–n structures at room temperature is calculated. It is shown that the internal quantum yield of electroluminescence is about 10% and is implemented at optimal doping levels for the n- and p-type regions of silicon diodes, ∼1015 and 5 × 1016 cm–3, respectively. With a decrease in the Shockley–Read–Hall lifetimes of electrons and holes, the internal quantum yield of electroluminescence in silicon barrier structures drops. The physical processes related to the effect of excitons in silicon has much in common with those in electroluminescence, photoluminescence, and photoconversion. It is shown that only electroluminescent p–i–n structures are promising for use in silicon integrated circuits. © 2004 MAIK “Nauka / Interperiodica”. [ABSTRACT FROM AUTHOR]

Time-resolved photoluminescence (PL) spectra have been measured at 90-300 K in the visible spectral range for porous nanocrystalline silicon films fabricated by laser deposition. The energy and time ranges in which the spectra were taken were 1.4-3.2 eV and 50 ns-10 gs, respectively. The correlation between PL characteristics (intensity, emission spectrum, relaxation times and their temperature dependence), structure, and dielectric properties (size and shape of Si nanocrystals, oxide phase of their coatings, porosity of films) has been studied. A model ofphotoluminescence is adopted, in which the absorption and emission of photons occur in quantum-size nanocrystals, and in which kinetically coupled subsystems of electron-hole pairs and excitons are involved in the radiative recombination. Possible mechanisms of the exciton Auger recombination in lowdimensional silicon structures are proposed. [ABSTRACT FROM AUTHOR]

The distribution of the transverse photovoltage has been investigated in samples of a Ge-GaAs heterostructure with an oxide layer (∼15 Å) at the interface. A method of simultaneous excitation with modulated and unmodulated radiation is used to observe the long-term relaxation of the photovoltage in these heterostructures. The energy barriers responsible for this effect are localized at the surface of the Ge film and the GaAs substrate, adjacent to the interlying oxide. It is shown that the gradients of the dark carrier concentration in the film and in the substrate are in opposite directions. A transition layer is formed during the growth process and not as a result of the classical heterodiffusion of the components in the film and the substrate. It is noted that the photoconductivity signal observed at the modulation frequency in the field formed at the sample as a result of the unmmodulated illumination must be taken into account in the recorded voltage signal. [ABSTRACT FROM AUTHOR]