Hydrogen related defects in high purity n-type float zone silicon samples have been studied by means of Deep Level Transient Spectroscopy. They were introduced, as well as the characteristic vacancy-oxygen (VO) and divacancy (${\rm V}_2$) centers, by MeV proton implantation. Two hydrogen related defect levels were resolved at 0.32 eV and 0.45 eV below the conduction band edge (Ec). Careful annealing studies indicate strongly that a third hydrogen related level, overlapping with the singly negative charge state level of ${\rm V}_2$, is also present in the implanted samples. The annealing behavior of the hydrogen related defects has been compared with literature data leading to a rather firm identification. The $E_{\rm c}-0.32$ eV level originates from a VO center partly saturated with hydrogen (a VOH complex) while the $E_{\rm c}-0.45$ eV level may be ascribed to a complex involving a monovacancy and a hydrogen atom (a VH complex). The third hydrogen related defect is tentatively ascribed to a complex involving a hydrogen atom and a divacancy (a ${\rm V}_2{\rm H}$ complex).

Results of the comprehensive study of deuterium-implanted hexagonal SiC (4H and 6H) using optical absorption and infrared measurements, elastic recoil detection analysis, thermal desorption and positron annihilation spectroscopies are reported. It is shown that implanted deuterium mainly forms bonds with lattice atoms. The amount of deuterium in the form of interstitial molecules and in vacancies is considerably smaller. Ion implantations with fluences exceeding 1015 D+/cm−2 create point defects in concentrations sufficiently high for complete positron trapping. Recrystallisation of the amorphised SiC does not remove the positron traps.

The number of scattering centres measured using channeled particles passing through a cavity layer formed by He implantation in silicon is shown to depend on the probe energy. This unusual result is explained in terms of interaction of channeled particles with atoms of cavity walls and dislocation lines. The Rutherford backscattering analysis in channeling incidence (RBS-C) is then used to study the time evolution of cavity and dislocation populations during annealing. It is shown that cavities are actually bubbles in equilibrium with the silicon matrix. The time to reach this equilibrium is related to the gas desorption. For our experimental conditions (40 keV, 5 × 1016 cm−2 and 800 °C), this equilibration time is about 10 minutes. Strain and dislocation build-up occur under shorter time scales, during which bubbles are probably overpressurized. The pinning of bubbles at dislocations must be taken into account to explain the evolution of these extended defects over a longer period.

Standard [001]-oriented p- (12–20 Ω cm) and n-type (1.8–2.6 Ω cm) Czochralski (Cz) silicon wafers were treated by a RF (13.56 MHz) hydrogen plasma at a substrate temperature of 250 °C. After the plasma hydrogenation subsequent annealing was applied up to 600 °C in air. The formation of H2 molecules in voids/platelets was investigated by Raman spectroscopy. The Raman intensities of the H2 vibration modes at ~4150 cm−1 exhibited significant intensity modulations in dependence on the annealing temperature. The intensities of the H2 Raman lines indirectly monitor the evolution of the voids/platelets upon annealing. This assumption was verified by cross-sectional transmission electron microscopy (XTEM), which was applied for comparison. The intensity modulations of the H2 Raman signal can be explained by the evolution of $\{111\}$ and $\{001\}$ platelets. At lower annealing temperatures (<500 °C) platelets laying in $\{111\}$ planes are dominant, while at elevated temperatures (≥500 °C) [001]-oriented platelets become more and more important. $\{110\}$ platelets were also observed using XTEM measurements in p-type material. In case of p-type substrates the Raman intensities were significant higher than for n-type material. The higher H2 Raman intensities in p-type Cz Si can be explained by the amphoteric character of hydrogen in silicon.

In this work we show that the cooling rate following a platinum diffusion strongly influences the electrical conductivity in weakly phosphorus doped silicon. Diffusions were performed at the temperature of 910 °C in the range of 8–32 hours in 0.6, 30, and 60 Ω cm phosphorus doped silicon samples. Spreading resistance profile analyses clearly show an n-type to p-type conversion under the surface when samples are cooled slowly. On the other hand, a compensation of the phosphorus donors can only be observed when samples are quenched. One Pt related acceptor deep level at 0.43 eV from the valence band is assumed to be at the origin of the type conversion mechanism. Its concentration increases by lowering the applied cooling rate. A complex formation with fast species such as interstitial Pt atoms or intrinsic point defects is expected. In 0.6 Ω cm phosphorus doped silicon, no acceptor deep level in the lower band gap is detected by DLTS measurement. This removes the opportunity of a pairing between phosphorus and platinum and suggests the possibility of a Fermi level controlled complex formation.

We report an in situ transmission electron microscopy (TEM) study of nanocavity evolution in amorphous Si (a-Si) under ion beam irradiation. The size evolution of the nanocavities was monitored during ion irradiation with Si or As at various temperatures between 300 and 600 K. A linear decrease of the nanocavity diameter was found as the ion fluence increased; it was much faster than its counterpart in crystalline Si (c-Si). Here, the shrinkage rate depended on the irradiation-induced atomic displacement rate. No significant temperature dependence was observed, confirming that the irradiation-induced nanocavity shrinkage in a-Si is essentially due to ballistic interactions, i.e., differs radically from that in c-Si. 


The use of gettering techniques that can be precisely localized in the component, such as neutral element implantation, has become crucial for power device manufacturing. High-dose helium implantation in silicon with a subsequent annealing induces defects (cavities and dislocations) which are able to getter metallic impurities and interact with dopants. These interactions can become a drawback of this technique. In this work, Si wafers $\langle 111\rangle$-oriented, p-doped with a concentration in between 6 × 1014 and 8 × 1018 B/cm3, were used. Iron contamination was performed by backside implantation (100 keV and 1012 Fe cm−2). Helium implantation was performed at 40 keV and 5 × 1016 at cm−2 at room temperature and followed by a Furnace Annealing (FA) stage. Secondary Ion Mass Spectrometry (SIMS) was used to study the dopant/metal/cavity layer system. In this paper, we will focus on the cavity layer interactions with both metals and dopant impurities. Our experimental results exhibit a large segregation of both species within the cavity layer that occurs during the annealing stage. Both species segregate simultaneously in helium-implanted samples without interacting one to the other. The number of gettering sites does not limit the trapping reactions.

In this paper, we report the results on the influence of a top oxide layer on the He-cavity formation in silicon samples. Si samples with top oxide layers of different thickness together with the pure silicon (used as reference) were implanted at room temperature with 1 MeV 3He at a dose of 5 × 1016 He/cm2. After implantation, the oxide layer was removed by using HF solution. Cross-sectional transmission electron microscopy (XTEM) was used to study the induced cavities followed by annealing. The results show that He implantation induces a well-defined-damaged layer with a thickness of about 180 nm in the reference Si. The defects are mainly made up of big cavities in the middle of the projected range and a large population of smaller ones both towards the surface and into the bulk. Very few stacking faults and dislocations loops can be seen around this band. The creation of cavities in Si with a 2.3 μm oxide layer, however, is a little different. Regions containing a chain of the biggest cavities surrounded by few smaller ones have also been found. Unexpected results are obtained in Si with a 1.2 μm oxide layer. In this case, only a monolayer of big cavities is observed at the depth corresponding to the projected range. The results are discussed in combination with SIMS measurements and SRIM simulations.

Nanostructured ferrites have been prepared by milling starting from commercial ferrite powders. Subsequent compaction and sintering were applied to obtain cylindrical pieces. The magnetization as a function of the grain size follows the non magnetic grain boundary (NMGB) model. The boundary thickness is estimated as about 1.5 nm, which is consistent with that calculated from the Mössbauer absorption areas. Due to the high lattice distortion, the small exchange length impedes the averaging of the anisotropy.

This paper describes how to forecast copper losses in a bus bar, composed of thin copper layers and several access points. The main difficulties are, first to determine current path and associated electrical characteristics, and second to propose electrical equivalent circuits of the bus bar, allowing copper losses representation. For bus bar electrical modeling, the choice of an integral method, PEEC method, is justified. Then, several equivalent circuits are proposed and their validity for copper losses computation discussed.

In the frame of the thermal step method (TSM) used to characterize the space charge in dielectric materials, we present an original numerical technique for determining the electric field distribution in the bulk of a dielectric plate or cable. The first stage of our technique is the application of the finite element method (FEM) in order to find instantaneous distributions of temperature profiles. The calculation of the electric field distribution is based on these obtained profiles. During the stage of the determination, our mathematical treatment is based on the TSM charge q(t) in order to avoid numerical instabilities on the temperature derivatives. Therefore, we have transformed the integral equation for the TSM current I(t) used in previous works to an integral equation for the TSM charge q(t). The control of the instantaneous propagation of the thermal wave front, produced by submitting one face of the dielectric to a thermal step, and the application of Simpson's method to the integral equation of the TSM charge q(t), allow to determine the electric field distribution. The results of the electric field distribution are validated with those obtained in previous works. A good agreement and an improvement near the dielectric thickness boundaries are observed on these results. The numerical space charge density within the material is obtained by numerical derivation of the field according to Poisson's equation.