The model of point defects-induced inversion of conductivity, formation and movement of the inverted p-n junction during Ar+ ions irradiation of p-Si crystals is presented.

This review addresses the centennial of the identification of the electron as the unit charge particle as well as the golden anniversary of the first demonstration of transistor action in a semiconductor material. Surprising skepticism arose in the acceptance of the quantized nature of the electric fluid; and formidable hurdles were encountered by the early generations of transistors. These obstacles are difficult to understand today. The transistor demonstration initiated research leading to highly perfected semiconductors. These materials can today be utilized toward precision measurements of fundamental quantities, such as the elementary charge by the quantum Hall effect, or the electron's effective spin magnetic moments by spectroscopy of quantum beats. Electrons and holes represent particle-antiparticle pairs in semiconductors with strong photon interactions, much like the fundamental lepton particles, for which a speculation on an evolutionary origin is presented. The unexpectedly long life of injected excess minority carriers in germanium established lifetime as the essential quantitative measure for materials quality and initiated massive and successful research on semiconductor defects.

It has been found that deuterium (D) instead of hydrogen (H) can be used to greatly strengthen the resistance of metal oxide semiconductor (MOS) transistors against hot carrier induced degradation. We have applied the new deuterium sintering process to CMOS technology and have obtained significantly improved hot carrier reliability resulting from the isotope effect. We will present a summary of these lifetime improvements from five different transistor structures of five different manufacturers, as well as the physical and electrical characterizations of the deuterium sintering process.

Deep level spectra obtained on n-type silicon samples after copper diffusion and rapid quench give evidence of a positive charge state of the precipitates in p-type silicon. Non-exponential precipitation behavior of interstitial Cu is demonstrated and explained. The possibility of Coulomb interaction between copper ions and copper precipitates is suggested and its influence on Cu precipitation kinetics is disCussed.

Characteristic 1.54 μm Era3+ emission has been observed from Er-implanted and annealed, low-temperature grown GaAs:Be samples. Cross-sectional transmission electron microscopy (TEM) studies reveal very little structural damage for elevated temperature implants up to an Er total fluence of 1.36 × 1014 Er/cm2. No Er emission was observed from any of the as-implanted samples, while post-implantation annealing optimized the Er emission intensity near 650°C. The Er-emission appears on top of a broad background luminescence peaking near 1500 nm. Significant enhancement of the optically active Er incorporation was achieved when the implantation was carried out at 300TC. The Er emission intensity was found to scale linearly with the Er implantation fluence for samples with an Er concentration up to ∼1019 Er/cm3. The sample with the highest Er concentration (∼1020 Er/cm3) began to show a sublinear dependence. The beginning of Er precipitation was observed after 750°C annealing, but it could even be observed after a 650°C annealing for the highest Er concentration sample. These precipitates are likely ErAs.

The incorporation and optical activation of Er in III-N optically-pumped and electroluminescent structures has been studied. For the first time, strong Er3+-related photoluminescence (PL) was measured at 300K for GaN:Er doped during growth on c-plane Al2O3 and Si. Also, room temperature electroluminescence was observed for simple GaN:Er structures on Si. Experiments to evaluate the effects C and 0 on the optical activity of Er indicated that these impurities enhance Er PL and EL in GaN. GaN films doped with Er to a concentration of 3 ×1018 cm−3 with [0] ∼ 1020 cm−3 and [C] ∼1021 cm−3 luminesce at 1.54 pim with an intensity ∼2 orders of magnitude greater than films with oxygen and carbon backgrounds of less than 1019 cm−3. Implantation and activation annealing was also shown to increase Er3+ signal in p(Mg)-i(Er)-n(Si) structures where C and 0 implant profiles were made to coincide with the Er-doped layers. The thermal PL quenching behavior was also markedly different for samples of varying 0 and C content. Er3+ photoluminescence from Gan:Er/Al2O3 samples with high O and C concentrations quenched by only 10% between 15 K and 300 K while the integrated PL signal from samples with lower [O] and [C] quenched ∼85% over the same temperature range.

The effects of hydrogen passivation in undoped p-ZnTe single crystals were studied by photoluminescence (PL) and photoconductivity (PC) measurements. Samples were exposed to r.f hydrogen plasma at 250 °C for different durations. Before passivation PL peaks were observed at 2.06 eV, 1.47 eV, 1.33 eV and 1.06 eV. After 60 minutes exposure, samples showed strong band edge green luminescence at 2.37 eV due to an exciton bound to a Cu acceptor. Further exposure to plasma resulted in disappearance of 2.37eV and 2.34 eV peaks due to damage. In PC studies the dark current was found to decrease by a factor of 70 on 60 minutes passivation. From the temperature dependence of PC gain, the minority carrier lifetime τn, was found to go through a maximum of 4.5 × 10−7 sec at 220 K before passivation. After 60 minutes exposure, τn, remained constant at 4.5 × 10−7 sec for T > 220 K and decreased for T < 220 K. The activation energies of τn, were determined and show marked changes on passivation for T > 220 K. Comparison between PL and PC studies showed that the deep acceptor level OTe responsible for emission at 2.06 eV is passivated giving rise to strong band edge emission at 2.37 eV while emission due to the midgap impurity levels at 1.47, 1.33 and 1.05 eV remained unaffected. The thermal activation energies of the PL peaks have also been determined and allow the construction of a defect energy level diagram for ZnTe.

Using the radioactive isotope 111Ag, the incorporation of Ag into CdTe is investigated by photoluminescence spectroscopy (PL) and the perturbed yy-angular correlation technique (PAC). PL is used to demonstrate the incorporation as a substitutional acceptor (AgCd) by monitoring the intensity of the donor-acceptor transition, corresponding to an acceptor level of EV + 108 meV, which decreases just with the half life of the radioactive decay of the 111Ag isotope. Since the isotope 111Ag also serves as a probe atom for PAC experiments, it is possible to confirm the incorporation of Ag atoms on lattice sites with cubic symmetry. Additionally, in CdTe a defect complex is detected, which is assigned to a AgCd-VTe pair. A second defect complex is observed in CdTe doped with In, which is assigned to an InCd-AgCd pair. By PAC experiments with the probe 111In that are performed in Ag doped CdTe, the formation of the InCd-AgCd pair is confirmed besides the well known formation of the InCd-VCd pair. The migration energy of the cation vacancy is estimated on the basis of an isothermal annealing sequence.

We have investigated stability of a new kind of boron clusters in silicon crystal using the first-principles local density functional approach. In the calculation, we used three types of hydrogenterminated silicon clusters; one of these types included the four interstitial cluster I4 as a complex defect. According to our calculation, we found that a new kind of boron clustering of n-borons substituted I4 (n < 3) became most stable when the ratio r = N(Bi) / (N(Sii) + N(Bi) in a confined area was less than or equal to 1/2, where N(B) is a number of interstitial boron atoms and N(Sii) is that of interstitial silicon atoms. On the other hand, when r was greater than 1/2, the l4 would be destroyed to create the [100] split boron pair, the [110] split silicon pair, and/or the substitutional boron and interstitial silicon pair.

We have carried out electrical characterization of defects in heavily damaged silicon, where damage is created by MeV heavy ions at doses near but below amorphization threshold. Trapping kinetics over several orders of magnitude in time have been monitored using isothermal spectroscopy called Time Analyzed Transient Spectroscopy (TATS). Two distinct effects regarding the nature of changes in density of states in the gap have been demonstrated. Firstly, we show that charge redistribution among multiple traps occur such that only the occupancy of the deeper states increase at the cost of shallower ones for long time filling. Secondly, a novel defect relaxation mechanism is observed for samples with relatively lower damage. A trap is seen to exhibit progressive deepening in energy with increase in filling time, finally stabilizing for large filling times. From the athermal nature of associated TATS peaks, it is argued that the relaxation involves large entropic contribution to free energy. The necessity of using a time domain relaxation spectroscopy such as TATS in the study of different mechanisms of relaxation is demonstrated.

Studies of the thermal dissociation kinetics of Cu-pairs in p-type silicon revealed that the dissociation energy of the pairs is 1.02±0.07 eV, which is twice as large as the binding energy of a coulombically bound donor-acceptor pairs formed by singly charged ions placed on nearest neighbor <111> sites. The dependence of the hole emission rate from the center versus electric field in the depletion region was found to be much weaker than predicted by the Pool-Frenkel law. On the other hand, the polarization potential describing emission from a neutral impurity gave a satisfactory fit to the experimental data. It is concluded that the Cu-pair is a donor with either covalent or mixed type of bonding of interstitial and substitutional copper atoms.

We have directly measured the solubility of iron and nickel in high and low boron-doped silicon using Instrumental Neutron Activation Analysis. Boron doping levels were 1.5×1019 and 6.5× 1014 atoms/cm3. Iron and nickel impurity concentrations were measured after extended indiffusions at 800, 900, 1000 and 1100°C for iron and 600, 700 and 800°C for nickel. We have measured a significant enhancement of Fe and Ni concentrations in high boron-doped silicon as compared to low boron-doped silicon. Based on these measurements, we show the iron donor energy level shifts towards the valence band with increased temperature, e.g. at 900°C the donor level is 0.24eV above the valence band as opposed to 0.39eV at room temperature. These results demonstrate that the impurity energy level shift with temperature must be accounted for in any prediction of segregation gettering of metal impurities into heavily doped substrates and heavily implanted doping layers. Additionally, our results suggest that either Ni solubility is greatly enhanced and/or the Ni diffusivity is greatly decreased with high boron doping of silicon.

In the paper the model of interaction of the hydrogenated vacancy with silicon interstitial and different dopants (B, P and As) in crystalline silicon is considered. Quantum chemical calculations using the SCF MO LCAO technique in the NDDO valence approach show that the hydrogenation of the vacancy leads to the considerable decrease of the energy barrier height for the interstitial atom incorporation into the vacancy site of the crystalline lattice. The potential barriers for incorporation of the interstitial into the site and for leaving the atoms from the site have been calculated as a function of hydrogen localization in the vicinity of the vacancy (inside and outside of the vacancy), the charge state of hydrogen localized outside the vacancy (HO, H+ and H-) and the transport direction (<111>, <110> and <100>) of the atoms both to the vacancy and out from it. The theory is compared with the reported experimental results.

Multicrystalline Si (mc-Si) wafers, used for the commercial solar cell fabrication, have spatial nonuniformities in the material properties that cause strong variations in the minority carrier lifetime, τ. We present the results of two-dimensional modeling to show carrier generation, recombination and transport in such a material. These results are used to infer measurement conditions that can yield meaningful spatially weighted average value of τ.

The Smart-Cut® process is based on proton implantation and wafer bonding. Proton implantation enables delamination of a thin layer from a thick substrate to be achieved whereas the wafer bonding technique enables different multilayer structures to be achieved by transferring the delaminated layer onto a second substrate. One of the best known applications of Smart Cut® is the Silicon On Insulator structure. The physical mechanisms involved in the delamination process are discussed based on the study of Proton-induced microcavity formation during implantation and growth during annealing. The experimental results on the time and temperature required to achieve delamination lead to different activation energies depending on the implantation conditions and resistivity of the substrate. All the experiments indicate that growth of microcavities is mainly controlled by hydrogen diffusion. The growth of these microcavities and the pressure inside them induce delamination when the catastrophic radius of the microcavities is reached.

We experimentally demonstrate the annealing and electrical activation of doped silicon without the direct application of heat as in conventional thermal annealing or pulsed laser annealing. 25 × 25 × 3 mm samples of silicon doped by neutron transmutation were irradiated with a short pulse from a 1.06-micron laser. The few joule laser pulse was focused to mm-diameter surface spot which resulted in high power (∼ 1011W/cm2) capable of launching shocks into the entire sample. In a few instances the entire silicon slab, including regionsfar outside the illuminated spot, was annealed and electrically activated. In the annealed samples electrical activation of donors throughout the slab, measured with a four-point probe, was uniform and comparable to that of thermally annealed control samples. Far-infrared spectroscopy and Hall measurements also showed that the donor species was activated and Raman spectroscopy demonstrated marked improvement in the crystal structure. We conjecture that the annealing was caused by the mechanical energy that was launched into the slab by the laser pulse. Results of experiments on an ion-implanted silicon sample are also discussed.

We have studied electrically active defects created by MeV heavy ion implantation in n-silicon. The buried damaged layer, created by implanting Ar’ ions of energy 1.45 MeV and doses in the range 1013-1014 cm−2 at room temperature, is embedded within the depletion layer of a Schottky diode. The defects are characterized using capacitance-voltage (C-V), current-voltage (I-V) and deep level transient spectroscopy (DLTS). Large concentration of electrically active defects are found to occur in a region several microns beyond the ion range or the damage profile predicted by Monte Carlo simulations. The dominance of a single trap in the damaged region is established from hysteresis effect in C-V, space charge limited conduction in forward I-V characteristics and DLTS results. With annealing in the temperature range of 400-600C, the observed changes in defect charge profile indicate that the effective electrical interface moves progressively towards the surface. C-V characteristics have been simulated using model charge profiles which suggest presence of a compensated region and a sharp negatively charged defect profile at a distance much larger than that expected from ion range. Our results constitute experimental evidence of migration and clustering of interstitial related defects, even at room temperature in case of high dose irradiation.

Subthreshold damage in p-type Si implanted and annealed at elevated temperature is characterized using deep level transient spectroscopy (DLTS) and transmission electron microscopy (TEM). P-type Si is implanted with Si, Ge and Sn with energies in the range of 4 to 8.5 MeV, doses from 7 × 1012to 1×1014cm−2and all annealed at 800°C for 15 min. For each implanted specie, DLTS spectra show a transition dose called threshold dose above which point defects transform in to extended defects. DLTS measurements have shown for the doses below threshold, a sharp peak, corresponding to the signature of point defects and for doses above threshold a broad peak indicating the presence of extended defects. This is found to be consistent with TEM analyses where no defects are seen for the doses below threshold and the presence of extended defects for the doses above threshold. This suggests a defect transformation regime where point defects present below threshold are acting like nucleating sites for the extended defects. Also the mass dependence on the damage evolution has been observed, where rod-like defects are observed in the case of Si and (rod-like defects and loops) for Ge and Sn despite the fact that peak concentration of vacancies for Ge and Sn are normalized to the peak number of vacancies for Si.

Due to the lack of effective gettering, gate oxide on thin-film-silicon-on-insulator (TFSOI) substrates is much more sensitive than its bulk Si counterpart to process damage during device fabrication, especially prior to gate oxide growth. Presented in this paper as a typical example is the severe oxide degradation caused by PMOS threshold-voltage implant. Several approaches to circumvent this problem are explored, such as Vt implant without sacrificial oxide (sacox), low temperature anneal before sacox removal, or implementation of lateral gettering. As a result of these efforts, a significant improvement in gate oxide integrity is achieved with increased oxide breakdown voltages and charge-to-breakdowns, as well as a reduction in oxide charge trapping. This work also demonstrates the feasibility of achieving bulk-comparable gate oxide on TFSOI substrates.

P-type Czochralski (Cz) Si was implanted with H (180 keV, 2.7.1016 cm−2) or He (300 keV, 1.1016 cm−2) ions. The gettering of O and H atoms by the buried implantation damage layers during annealing up to 4 hours (1000°C in H2 or N2 ambient) was studied by secondary ion mass spectroscopy (SIMS) and spreading resistance probe (SRP) measurements. Buried defect layers act as good getter centers for O and H atoms at appropriate heat treatments. The enhanced gettering of O atoms in H implanted Cz Si (as compared to the gettering of O in He implanted samples) as well as the enhanced gettering of O during annealing in H2 flow (as compared to N2 ambient) can be explained by a hydrogen enhanced O diffusion towards the defect layers. According to a strong accumulation of O at the buried damage layers and near the surface some anomalies of the SRP profiles can be observed after post-implantation annealing.