Si/Gex Si1−x heterojunction bipolar transistors formed by Ge ion implantation in Si. Narrowing of band gap and base width

Si/Ge_xSi_1−x heterojunction n-p-n bipolar transistors (HBT) with a double polycrystalline silicon (polysilicon) self-aligned structure were fabricated by using high dose Ge implantation for the formation of the Si/Ge_xSi_1−x heterostructure and As and BF₂ implantation for emitter and base doping. Improvements in electrical characteristics compared to reference Si transistors are demonstrated and related to a band gap narrowing in the base region and to a reduction of B diffusion.     

The production of non linear damage under molecular ion implantation

Germanium atomic (Ge₁) and molecular ions (Ge₂) of equivalent energy are implanted in silicon at an elevated temperature. The ion induced damage has been characterized by RBS channeling (RBS/C) and positron annihilation spectroscopy. The RBS/C studies indicate that the molecular ion implantation has produced more defects in the near surface regions compared to the atomic ion implantation. This paper reports a first time observation of an enhanced production of vacancy related defects in silicon implanted with molecular ions.     

Relaxation of strain during solid phase epitaxial growth of Ge+ ion implanted layers in silicon

Formation of Si_1−xGe_x-alloy layers by solid phase epitaxial growth (SPEG) of Ge⁺ ion implanted silicon has been studied. The ion implantations were performed with 40, 100, 150, 200 and 300 keV ⁷⁴Ge⁺ ions and various ion doses. The SPEG of the ion implanted layers was carried out in a conventional furnace at 850°C for 20 min under a flow of nitrogen gas. The Si_1−xGe_x-alloy layers were characterised by Rutherford backscattering spectrometry and transmission electron microscopy (TEM). For a given ion energy, a Si_1−xGe_x-alloy layer with no observable extended defects can be manufactured if the ion dose is below a critical value and strain-induced defects are formed in the alloy layer when the ion dose is equal to or above this value. The critical Ge⁺ ion dose increases with ion energy, while the critical maximum Ge concentration decreases. For ion energies ⩽150 keV, the defects observed in the alloy layers are mostly stacking faults parallel to the {1 1 1} planes. For higher ion energies, 200 keV and above, the majority of defects in the alloy layer are hairpin dislocations. In the whole ion energy range, the critical ion dose and the depth position of the nucleation site for the stacking faults obtained from the measurements are in good agreement with theoretical predictions. Extended defects are formed in the alloy layer during the SPEG when the regrowth of the crystalline/amorphous interface has reached the depth position in the crystal where the accumulated strain energy density is equal to the critical value of 235 mJ/m².     

Integrity degradation of UO2 pellets subjected to thermal shock

Thermal shock behavior of UO₂ pellets has been investigated by means of out-of-pile experiments and a theoretical analysis which particularly emphasized the porosity effect on the thermal shock damage. In the experiments, specimens of porosity range 0.05–0.15 were thermal-shocked by heating and then quenching in a water bath at various quenching temperature differences (ΔT). Results showed that with increasing porosity, ΔT values cause a first damage (ΔT_c) and bring about destructive failure of the specimens increased, while the strength loss at ΔT_c was reduced. These findings suggested that the higher the porosity, the higher the pellet integrity during rise to power. Theoretical equations expressing the thermal shock damage were introduced. Good agreements were obtained between observed and predicted values.     

Critical current density changes in irradiated Nb3Sn

Söll recently modeled changes in the current-carrying capacity of superconducting Nb₃Sn after irradiation. As the dose increases, the critical current density (J_c) generally increases, reaches a maximum, and decreases. The model relates the maximum J_c for different types of irradiations to the integrated damage energy (E_D) that the irradiating particles transfer to the lattice. Earlier, Söll et al. related cirtical-temperature (T_c) decreases in irradiated Nb₃Sn to E_d, which appears more reasonable since T_c is a measure of disorder (or replacements) accompanying defect production, and the final defect configuration (or displacements) are less important. The annealing temperature for the disorder (~700°C) exceeds any of the irradiation temperatures (T_IRR); therefore, T_IRR is unimportant in the T_c experiments. However, different defect structures exhibit considerably different flux pinning and the J_c model does not consider different spatial variations of the defects during production or migration and agglomeration of the defects during high-T_IRR experiments, both of which affect flux pinning. The lack of the model to take into account the physics of the damage is the subject of this paper. Arguments are presented why the defect configurations should be considered, and recently published data is presented that conflict with the conclusions of this damage-energy model.     

Defects in carbon and oxygen implanted p-type silicon

Both oxygen and carbon ion implantation are frequently used to form either insulating buried SiO₂ or SiC layer for various purposes. This creates a renewal of the interest in defects produced during such implantation processes. In the present paper we report on deep level transient spectroscopy studies of defect states occurring in boron-doped p-type silicon after high dose C⁺ and CO⁺ ion implantation and subsequent thermal annealing. It is shown that the predominant defect created during the implantation is in both cases related to silicon selfinterstitial clusters, which upon annealing at higher temperatures evolve to extended structural defects.     

Microstructure evolution of Ge implanted silicon oxide thin films upon annealing treatments

The structural evolution of silicon oxide films with Ge⁺ implantation was traced with a positron beam equipped with positron annihilation Doppler broadening and lifetime spectrometers. Results indicate that the film structure change as a function of the annealing temperature could be divided into four stages: (I) T < 300 °C; (II) 300 °C ? T ? 500 °C; (III) 600 °C ? T ? 800 °C; (IV) T ? 900 °C. In comparison with stage I, the increased positron annihilation Doppler broadening S values during stage II is ascribed to the annealing out of point defects and coalescence of intrinsic open volumes in silicon oxides. The obtained long positron lifetime and high S values without much fluctuation in stage III suggest a rather stable film structure. Further annealing above 900 °C brings about dramatic change of the film structure with Ge precipitation. Positron annihilation spectroscopy is thereby a sensitive probe for the diagnosis of microstructure variation of silicon oxide thin films with nano-precipitation.     

Radiation damage in ZnO ion implanted at 15 K

Commercial O-face (0 0 0 1) ZnO single crystals were implanted with 200 keV Ar ions. The ion fluences applied cover a wide range from 5 × 10¹¹ to 7 × 10¹⁶ cm⁻². The implantation and the subsequent damage analysis by Rutherford backscattering spectrometry (RBS) in channelling geometry were performed in a special target chamber at 15 K without changing the target temperature of the sample. To analyse the measured channelling spectra the computer code DICADA was used to calculate the relative concentration of displaced lattice atoms.Four stages of the damage evolution can be identified. At low ion fluences up to about 2 × 10¹³ cm⁻² the defect concentration increases nearly linearly with rising fluence (stage I). There are strong indications that only point defects are produced, the absolute concentration of which is reasonably given by SRIM calculations using displacement energies of E_d(Zn) = 65 eV and E_d(O) = 50 eV. In a second stage the defect concentration remains almost constant at a value of about 0.02, which can be interpreted by a balance between production and recombination of point defects. For ion fluences around 5 × 10¹⁵ cm⁻² a second significant increase of the defect concentration is observed (stage III). Within stage IV at fluences above 10¹⁶ cm⁻² the defect concentration tends again to saturate at a level of about 0.5 which is well below amorphisation. Within stages III and IV the damage formation is strongly governed by the implanted ions and it is appropriate to conclude that the damage consists of a mixture of point defects and dislocation loops.     

Effects of hydrogen implantation into GaN

Proton implantation in GaN is found to reduce the free carrier density through two mechanisms – first, by creating electron and hole traps at around E_C − 0.8 eV and E_V + 0.9 eV that lead to compensation in both n- and p-type material, and second, by leading to formation of (AH)° complexes, where A is any acceptor (Mg, Ca, Zn, Be, Cd). The former mechanism is useful in creating high resistivity regions for device isolation, whereas the latter produces unintentional acceptor passivation that is detrimental to device performance. The strong affinity of hydrogen for acceptors leads to markedly different redistribution behavior for implanted H⁺ in n- and p-GaN due to the chemical reaction to form neutral complexes in the latter. The acceptors may be reactivated by simple annealing at ⩾600°C, or by electron injection at 25–150°C that produces debonding of the (AH)° centers. Implanted hydrogen is also strongly attracted to regions of strain in heterostructure samples during annealing, leading to pile-up at epi–epi and epi–substrate interfaces. IR spectroscopy shows that implanted hydrogen also decorates V_Ga defects in undoped and n-GaN.     

Comparative study of damage production in ion implanted III–V-compounds at temperatures from 20 to 420 K

The damage evolution in ion implanted InP, GaAs, GaP and InAs is studied as a function of the ion fluence in the temperature range 20–420 K using various ion masses. It is shown that the macroscopic behaviour can be described in terms of critical temperatures T_c which depend on the ion mass and on the dose rate for a given material. At temperatures T_I< T_c amorphization is obtained by direct impact amorphization and the growing of the amorphous zones. However, athermal in-cascade annealing is observed even at 20 K, which is the more pronounced the lighter the ion is. This indicates the influence of the density of the primary cascades on defect recombination. Around T_c intrinsic defects are mobile leading to an equilibrium between defect production and annealing over a broad dose region. Amorphization at very large ion fluences is the consequence of complex processes which are influenced by the high ion concentration and the formation of dislocation bands near the end of range. Using an empirical formula which describes the ion mass and dose rate dependence of the critical temperature, the damage evolution for certain implantation conditions becomes predictable.     

An ERD/RBS/PIXE apparatus for surface analysis and channeling

A system for surface analysis, lattice localisation of impurities, defect studies in crystalline solids and routine ERD (elastic recoil detection) is described. The apparatus features a high vacuum chamber, a computer controlled precision goniometer and an energy plus time-of-flight mass discrimination system. First results of channeling/ERD experiments using 15–30 MeV ³⁵Cl beams on silicon crystals implanted with B⁺ and BF₂⁺ are presented, as well as data on the effect of beam induced damage on the boron distribution.     

Populations of np terms in deuterium atoms emergent from carbon foils bombarded with D+, D2+ and D3+ ions

We have measured relative beam-foil populations of 2p, 3p, and 4p terms in D⁰ as a function of the projectile energy $\text{(20 ?}\text{E}\text{M}\text{? 500}\text{keV}\text{amu}\text{)}$ for D⁺, D₂⁺, and D₃⁺ ions impinging on carbon foils of various thicknesses (? 2–20 $\text{μg}\text{cm}{}^2$).With D⁺ projectiles, the np populations reach their equilibrium values even in the thinnest foils used. We compare the dependence on energy of these populations to the equilibrium neutral fraction variation for hydrogen (deuterium) beams emerging from a carbon foil and deduce some information concerning beam-foil populations.When molecular projectiles pass through very thin foils, well known molecular effects appear which depend on the dwell time, t, i.e., the time spent by the projectile in the foil. In this work we consider only the long-dwell-time region t > 2 × 10^?15s. We study the variation of R_α = I^molec/I^atom (I^molec and I^atom are the Ly-α intensities per incident deuteron (proton) observed with molecular and atomic projectiles of the same velocity, respectively) with the projectile energy per nucleon ($\text{E}\text{M}$) and the thickness (T) of the foil. For a foil of given thickness, R_α increases with $\text{E}\text{M>}$ and reaches a saturation value R_∞ which decreases when T increases. These results, in agreement with our previous measurements using hydrogen projectiles, indicate that t is not the only parameter relevant to molecular effects. Comparisons are reported between $\text{R>}{}_{\mathrm{\alpha }}\text{(E}\text{M>}$) values obtained (a) with H₂⁺ and D₂⁺ projectiles and (b) with D₂⁺ and D₃⁺ projectiles, using foils of various given thicknesses. Ratios $\text{R}{}_{\mathrm{\beta }}\text{(E}\text{M}$) and $\text{R}{}_{\mathrm{\gamma }}\text{(E}\text{M}$) are also measured using Ly-β and Ly-γ radiations and compared to $\text{R}{}_{\mathrm{\alpha }}\text{(E}\text{M}$) values. An interpretation for some of our results is proposed.     

Optical and electrical characterizations of Mn doped p-type β-FeSi2

β-FeSi₂ has attracted increasing attention as a promising material for optoelectronic and thermoelectronic devices due to a high optical absorption coefficient (α) of about 10⁵ cm⁻¹ near 1.0 eV and its chemical stability at higher temperatures. For the future practical use of this material in devices, the control of each electrical conductivity type and the improvement of the material quality are highly required. Although unintentionally doped β-FeSi₂ layers formed on n-type Si(1 0 0) by the conventional electron-beam deposition (EBD) have typically shown n-type conductivity, the p-type β-FeSi₂ layers were formed by the introduction of Mn impurity using ion-implantation at room temperature (RT) and subsequent annealing procedures. In this study, we aimed to make p-type β-FeSi₂ by implantation of ⁵⁵Mn⁺ ions into EBD-grown n-type β-FeSi₂ layers/n-Si, where ⁵⁵Mn⁺ ions were implanted at two different temperatures (T_sub) of RT and 250°C using an energy and a dose of 300 keV and 2.68 × 10¹⁵ cm⁻², respectively. Their optical and electrical properties, which ought to be affected by implantation and annealing temperatures (T_a2), were investigated by Raman scattering, optical transmittance, reflectance and van der Pauw measurements. The results showed that the ⁵⁵Mn⁺ doping with T_sub=RT and higher thermal annealing at T_a2=900°C produced p-type layers of good quality with maximum hole mobility of 454.5 cm²/Vs at about 65 K.     

Energy distributions of neutral atoms sputtered by very low energy heavy ions

Polycrystalline Cu was sputtered by normally incident, very low energy Ar⁺ ions (E₀ = 40–1000 eV). The kinetic energy (E) distributions of the neutral Cu atoms sputtered normally from the Cu surface were measured, using secondary neutral mass spectrometry. For values of E₀ above approximately 600 eV, the observed energy distributions agreed closely with the Thompson-Sigmund theory. For values of E₀ less than about 600 eV the distributions fell off faster than predicted by the Thompson-Sigmund theory, and the peak value of the distribution shifted to somewhat lower energies. Both these effects were exaggerated as E₀ was further lowered. The average kinetic energy of the sputtered neutral Cu atoms increased with increasing E₀. The rate of this increase was less at higher values of E₀.     

On the emission of MoCs molecular ions from Cs irradiated molybdenum surface

The present paper deals with the emission of atomic and molecular ions from elemental molybdenum surface under Cs⁺ bombardment to explore the MCs⁺ formation mechanism with changing Cs surface coverage. Integrated count of MoCs⁺ shows a monotonic increase with increasing primary ion energy (1-5 keV). Change in MoCs⁺ intensity is attributed to the variation of surface work function ? and cesium surface concentration c_Cs due to varying impact energies. Variation of c_Cs has been obtained from the expression, c_Cs ∝ 1/(1 + Y) where Y is the elemental sputtering yield estimated from TRIM calculations. Systematic study of the energy distributions of all species emerging from Mo target has been done to measure the changes in surface work function. Changing slopes of the leading parts of Cs⁺ energy distributions suggest a substantial depletion in surface work function ? with decreasing primary ion energies. Δ? shows a linear dependence on c_Cs. The maximum reduction in surface work function Δ?_max = 0.69 eV corresponds to the highest value of c_Cs = 0.5. A phenomenological model, based on the linear dependence of ? on c_Cs, has been employed to explain the MoCs⁺ data.     

Nanometer-thick SGOI structures produced by Ge+ ion implantation of SiO2 films and subsequent hydrogen transfer of Si layers

Nanometer-thick silicon-germanium-on-insulator (SGOI) structures have been produced by the implantation of Ge⁺ ions into thermally grown SiO₂ layer and subsequent hydrogen transfer of silicon film on the Ge⁺ ion implanted substrate. The intermediate nanometer-thick Ge layer has been formed as a result of the germanium atom segregation at the Si/SiO₂ bonding interface during annealing at temperatures 800–1100 ^оС. From a thermodynamic analysis of Si/Ge/SiO₂ system, it has been suggested that the growth of the epitaxial Ge layer is provided by the formation of a molten layer at the Si/SiO₂ interface due to the Ge accumulation. The effect of germanium on the hole mobility in modulation-doped heterostructures grown over the 3–20 nm thick SGOI layers was studied. An increase in the Hall hole mobility in SGOI-based structures by a factor of 3–5 was obtained in comparison with that in respective Ge-free SOI structures.     

Post-annealing effects on the shallow-junction characteristics caused by high-fluence 77 keV BSi molecular ion implantations at room and liquid nitrogen temperatures

In this study, n-type <1 0 0> silicon specimens were liquid nitrogen temperature (LT) and room temperature (RT) implanted with 2 × 10¹⁵ cm⁻² 77 keV BSi molecular ions to produce shallow junctions. Post-annealing methods under investigation included furnace annealing (FA) at 550 °C for 0.5, 1, 2, 3 and 5 h and rapid thermal annealing (RTA) at 1050 °C for 25 s. Post-annealing effects on the shallow-junction characteristics were examined using one-step (FA) and two-step (FA + RTA) post-annealing treatments. Secondary ion mass spectrometry (SIMS), cross-sectional transmission electron microscopy (XTEM), a four-point probe and Raman scattering spectroscopy (RSS) were employed to analyze junction depths (x_j), damage microstructures, sheet resistance (R_s) and damage characteristics, respectively. The results revealed that the shallow-junction characteristics of the LT implant are better than those of the RT one when post-annealing time in FA exceeds 1 h. A post-annealing time of 3 h in FA is needed in order to obtain the optimal one- or two-step post-annealing effects on the shallow-junction characteristics in both the LT and RT implants.     

Computer simulation of primary damage creation in displacement cascades in copper. I. Defect creation and cluster statistics

Atomic-scale computer simulation has been used to investigate the primary damage created by displacement cascades in copper over a wide range of temperature (100 K ? T ? 900 K) and primary knock-on atom energy (5 keV ? E_PKA ? 25 keV). A technique was introduced to improve computational efficiency and at least 20 cascades for each (E_PKA, T) pair were simulated in order to provide statistical reliability of the results. The total of almost 450 simulated cascades is the largest yet reported for this metal. The mean number of surviving point defects per cascade is only 15-20% of the NRT model value. It decreases with increasing T at fixed E_PKA and is proportional to (E_PKA)^1.1 at fixed T. A high proportion (60-80%) of self-interstitial atoms (SIAs) form clusters during the cascade process. The proportion of clustered vacancies is smaller and sensitive to T, falling from 30% to 60% for T ? 600 K to less than 20% when T = 900 K. The structure of clusters has been examined in detail. Vacancies cluster predominantly in stacking-fault-tetrahedron-type configurations. SIAs tend to form either glissile dislocation loops with Burgers vector b = 1/2<1 1 0> or sessile faulted Frank loops with b = 1/3<1 1 1>. Despite the fact that cascades at a given E_PKA and T exhibit a wide range of defect numbers and clustered fractions, there appears to be a correlation in the formation of vacancy clusters and SIA clusters in the same cascade. The size and spatial aspects of this are analysed in detail in part II [unpublished], where the stability of clusters when another cascade overlaps them is also investigated.     

Formation of germanium nanoparticles in silica glass studied by optical absorption and X-ray absorption fine structure analysis

We examined the relation between the 3.1 eV emission band and local structure for Ge⁺ implanted silica glass by means of photoluminescence, optical and X-ray absorption spectroscopies. In the 2 × 10¹⁵ cm^?2 implanted sample, a new emission band around 2.7 eV was observed, the origin of which was assigned to the B_2α oxygen deficient center and/or small Si clusters in silica. When the Ge⁺ fluence exceeded 2 × 10¹⁶ cm^?2, a sharp and intense 3.1 eV emission band replaced the 2.7 eV band. We found that the intense 3.1 eV PL occurred by the prolonged X-ray irradiation onto the 2 × 10¹⁵ cm^?2 implanted sample. UV–vis absorption and XAFS spectroscopies suggested that the aggregation of atomically dispersed tetravalent (Ge(IV)) atoms into Ge(0) clusters of ～1 nm exhibit strongly correlation with the generation of the 3.1 eV PL. Such nano- and/or subnano-size Ge(0) clusters formed by the X-ray radiation were oxidized and decomposed again to the isolated Ge(IV) atoms, while those produced by the higher Ge⁺ fluence were stable against the exposure to air.     

Growth of silicon nitride films by bombarding amorphous silicon with N ions: MD simulation

In this study, the molecular dynamics simulation method was employed to investigate the growth of silicon nitride films by using N⁺ ions, with energies of 50, 100, 150 and 200 eV, to bombard an amorphous silicon surface at 300 K. After an initial period of N⁺ bombardment, saturation of the number of N atoms deposited on the surface is observed, which is in agreement with experiments. During subsequent steady state deposition, a balance between uptake of N by the surface and sputtering of previously deposited N is established. The Si(N_x) (x = 1-4) and N(Si_y) (y = 1-3) bond configurations in the grown films are analyzed.