The in-plane stress in silicon wafers during amorphization by ion-bombardment was determined from wafer curvature measurements using an in-situ laser scanning technique. Measurements were made during room temperature bombardment with 2 MeV Ne, Si, Ar, Kr, and Xe ions. In all experiments, compressive stress was built-up in the bombarded region as a function of the fluence, until a maximum was reached at the dose required to form amorphous silicon. During further amorphization by bombardment, the stress decreased and eventually stabilized. If ion bombardment was interrupted during amorphization, a stress increase was observed over a period of several minutes; when the beam was turned on again, the stress returned immediately to the value measured before interruption. Step height measurements were performed on implanted wafers to determine the out-of-plane strain, and RBS was used to determine the damage profiles. A model is proposed that describes the behavior in terms of the expansion of crystalline silicon by the creation of defects and the flow of amorphous silicon under the ion beam.

Channeled implants have been performed with lOOkeV 28Si+ into p-type Si(100) to obtain a buried amorphous layer. Before and after recrystallization of the a-Si layer, Cu was implanted at an energy of 15 keV and a dose ranging from 5E13 to 1E15 I cm2- to obtain a high concentration of Cu in the near surface region. Also, Cu implants were performed into virgin Si for comparison. After Cu implantation, thermal annealing was performed at temperatures between 490 °C and 900 °C for 10 min. to 320 min. Cu profiles before and after annealing were studied with Rutherford Backscattering Spectrometry and channeling analysis. For the case where Cu was implanted after recrystallization of the buried amorphous layer, Cu was gettered at the position where the ale interfaces met during recrystallization. For the case where Cu was implanted before recrystallization, Cu diffused towards the buried a-Si region upon annealing and was trapped inside the recrystallizing buried amorphous layer. The results show that buried damage layers can effectively getter Cufrom the Si surface layer and gettering is most efficient at 600 °C.

Microstructural changes occurring during the early stages of rapid thermal annealing of polycrystalline silicon bipolar emitters crucially affect the final dopant distribution and hence the performance of these devices. The first stage of annealing is epitaxial regrowth in the solid phase of the layer amorphised by the implantation. In-situ studies using time-resolved reflectivity measurements, combined with cross-sectional transmission electron microscopy of partly annealed structures, have determined the effects of initial grain size, annealing temperature and amorphising species (Si or As) on the rate of regrowth and the microstructural changes which occur during annealing. As the grain size is reduced, the regrowth rate decreases and the interface roughness increases. Arsenic implantation alters the rate of regrowth in such a manner as to produce a smoother interface than that in silicon implanted material.

This work was concerned with comparing the relative effects of boron and phosphorus impurities on the solid phase epitaxial (SPE) regrowth rate of self-ion amorphized layers in silicon wafers with (100) orientation. We used previously reported data measured by in situ, high precision, cw laser interferometry during isothermal annealing for temperatures from 450°C to 590°C, and concentrations in the range from 7.8×1018 cm-3 to 5×l020 cm-3 for boron (NB), and from 5×l017 cm-3 to 3×1020 cm-3 for phosphorus (Np) impurities. The basis for the comparison was a recently developed model that extends the Spaepen-Turnbull model for silicon recrystallization to include ionization enhanced processes.

The experimental data for bom boron and phosphorus exhibited the linear variation in regrowth rate expected for low concentrations of implanted hydrogenic impurities having a concentration-independent fractional ionization in amorphous silicon. In the linear range the relative enhanced regrowth rate produced by these impurities can be expressed as a product of their, relative fractional ionizations, and the relative amount the rate constant for reconstruction is altered by localizing an electron, or a hole, at the reconstruction site. Assuming that a localized hole and electron equally softened the potential barrier for reconstruction, the experimental results indicated that boron had an ?40 meV lower barrier to ionization in amorphous silicon than phosphorus.

The variations in the SPE regrowth rates with higher concentrations of both implanted boron and phosphorus were well fit by quadratic equations, but with different curvatures (+ and - for B and P respectively). This result was interpreted to indicate that SPE regrowth was further enhanced by localized hole pairs, but retarded by localized electron pairs.

1.5 MeV Kr+ irradiation of polycrystalline Ge at room temperature, and subsequent annealing were carried out with in situ TEM observations. After a Kr+ dose of 1.2x1014 ions/cm2, Ge in the electron transparent region was completely amorphized. Continuous irradiation of the amorphized Ge resulted in a high density of small cavities. These cavities, first observed after 7x1014 ions/cm2 with an average diameter of ∼3 nm, grew into large (∼50 nm) irregular-shaped holes, transforming the irradiated Ge into a sponge-like material after 8.5x1015 ions/cm2. The crystallization temperature and the morphology of the crystallized Ge after annealing were found to be dependent on the Kr+ dose. The sponge-like structure was retained after crystallization at ∼600°C.

The electrical activation characteristics of implanted Be, Mg, Si and S in AlxGa1–xAs (x = 0-1) were investigated as a function of ion dose for rapid annealing in the range 600-950°C. The apparent activation energy for electrical activity of these species increases with increasing AlAs mole fraction - for Be, the activation energy is 0.35eV for GaAs and 0.49eV for Al0.54Ga0.46 As. There is no evidence for pairing of Be and O in AlGaAs, in contrast to the situation for GaAs. Self-compensation is the predominant limiting mechanism for Si activation in AlGaAs as determined by the relative photoluminescence intensities of the SiGa-to-SiAs related transitions. No significant redistribution of implanted Si is observed for any AlAs mole fraction for rapid annealing (5 sec) up to 900°C, whereas S shows motion into the AlGaAs and no tendency to outdiffuse. By contrast, both Be and Mg display loss of the dopant to the surface, and little redistribution toward the bulk. Minimal damage is observed by transmission electron microscopy in as-implanted AlGaAs for Be or Si doses below the amorphization threshold. Upon annealing at the conditions for optimum activation, a high density of small dislocation loops is observed near the end of the ion range.

Previous studies have indicated that dual implantation can efficiently introduce group IV dopant onto selected sub-lattice sites in m-V compound semiconductors, thus enhancing electrical activation. We studied this phenomenon in GaP using Rutherford Backscattering Spectroscopy (RBS) to determine the lattice location of Sn atoms. We used single crystals of GaP (100) which had been implanted at 400° C with120 Sn+ following previously implanted69 Ga+ or 31P+. Energies were selected for euivalent projected ranges, and all species were implanted with doses of 1 × 1015 atoms/cm2 . Asymmetry in the angular scan of the {111} planar channel was then used to determine the sub-lattice location of the implanted Sn. RBS results indicated that for all implants Sn atoms were substituting Ga and P sites eually. However, Hall effect measurements gave p type conduction for GaP implanted with Sn alone, while those with prior implants of Ga or P resulted in n-type conduction. RBS and Hall effect results are explained by a vacancy complex model.

(100) GaAs was implanted with 29Si+ at energies of 10keV or 20keV and 119Sn+ at an energy of 60keV. The doses used for the Si+ implants were from 1 × 1013 crrr2 to 1 × 1016 cm-2, whereas the Sn+ doses ranged from 1.5×1012 cm-2 to 1.5×1015 cm-2. Implanted samples were rapidly thermal annealed (RTA) in either an argon or arsine ambient at 950°C or 900°C for 12 seconds. Transmission electron microscopy (TEM) results indicate that the annealing ambient has a dramatic effect on the defect annealing kinetics for 10keV implants but not the 20keV Si+ implants. For the 10keV implants reducing arsenic loss via use of an arsine ambient or a Si3N4 cap results in a dramatic decrease in the dislocation loop density. The defect morphology is not related to the as implanted morphology (amorphous or crystalline) but is related to the species. Hall effect measurements revealed an increase in the mobility by a factor of 1.5 times for high dose Si+ and Sn+ implanted samples when RTA was carried out in an arsine instead of an argon ambient. This is despite little or no change in the extended defect morphology or carrier concentration, indicating that the point defect concentration influences the mobility.

Thermodynamic and structural properties of amorphous Si (a-Si), prepared by MeV 28Si+-ion implantation are investigated by differential scanning calorimetry, Raman spectroscopy and X-ray diffraction. The influence of thermal annealing below 500 °C on a-Si is investigated with these different probes. The observed changes result from structural relaxation. Raman spectroscopy and X-ray diffraction show that structural relaxation is accompanied by changes in the average atomic structure.

The density of amorphous Si was measured. Continuous and buried amorphous Si films were produced by 0.5-8 MeV Si implantation through a steel contact mask. Surface steps of amorphous Si stripes with initial thicknesses from 0.9 to ∼ 5.0 μm were measured using a surface profilometer. For implants up to 5 MeV, the amorphous Si is constrained laterally by the surrounding crystal and deforms plastically. The density of amorphous Si deduced from the surface step heights is 4.91 × 1022 cm-3, 1.7 ± 0.1 % less than the density of crystal Si (5.00 × 1022 cm-3).

The diffusion of boron ions implanted in Si(100) during high temperature treatment has been studied by means of Secondary Ion Mass Spectrometry. The boron ions were implanted at energies of 5 and 10 keV along random direction and along [100]. The implantation dose amounted to 1×1013/cm2. During annealing an enhanced diffusion of boron is observed in the tail region of the implantation profiles. In all cases the enhanced diffusion is completed after Rapid Thermal Annealing at 900°Cfor 10”. The enhanced diffusion is largest for the channeled ion implantations. Channeling analysis shows that the transient diffusion is not correlated with the annihilation of defects near the projected range. Therefore, the observed diffusion tail is attributed to rapidly diffusing interstitial boron. Post-amorphization of boron implanted silicon with Si+ ions is used to prevent transient diffusion. Thus a very sharp pn-junction is obtained at a depth of 03 jim and the effective number of charge carriers equals the implanted dose after annealing at 550°Cfor 5hrs and 1000°Cfor 10”.

The effect of electron beam irradiation on the transport properties of HEMT structures has been studied. The data is found to agree with a model containing no free parameters which predicts the spatial extent of the damage. As grown HEMT material was fabricated into Hall bars. The material was then characterized by measurement of 2D EG number density and drift mobility. The material was then irradiated at constant dose with electron energies between 2.5 and 20 keV. Damage was assessed by changes in the number density and mobility. The number density was essentially unchanged at all energies. No change in mobility occurred for 2.5, 15, and 20 keV, however, a decrease in the mobility occurred for energies from 5.0 to 12.5 keV. These results agree with our model of the electrons energy loss as a function of its path length. The defects are shown to be charged, and possible sources of the defects are discussed.

Beryllium (Be) depth profiles, carrier profiles, ideality factors (deviation from thermionic emission and diffusion theory) and Schottky barrier heights of Al/p+-GaAs Schottky diodes were investigated by using focused ion beam (FIB) and de-focused ion beam (DFIB) of 40keV Be+, with dose range from 5.0×1012ions/cm2 to 3.4×1014ions/cm2. Differences of the depth profiles between FIB implantation and DFIB implantation were observed, and it was confirmed that depth profiles of the ion implantations were affected by current densities of the ion probes. The other side, it was recognized that the ideality factors and the Schottky barrier heights of the Schottky diodes were depended on effect of thermionic field emission.

Ion beams of keV and MeV energies have been used to bombard amorphous Si (a-Si), which had previously been annealed (‘relaxed’). Analysis by Raman spectroscopy and differential scanning calorimetry shows that when 1 out of every 20 Si atoms is displaced by a nuclear collision, the a-Si returns to its unrelaxed state and cannot be distinguished from as implanted a-Si. Moreover, the kinetics of the heat release on annealing of similarly bombarded crystalline Si (c-Si) are qualitatively identical to those of structural relaxation in a-Si. This implies that the population of ion beam induced defects in a-Si is very similar to that in c-Si. It also shows that defect annihilation is an important ingredient in the mechanism of structural relaxation of a-Si.

150keV Si* ions and 160keV P* ions were implanted at 200°C with doses ranging from 5x1013 to 1x1015/cm2 to study the effect of dual implantations on the electrical properties of Fe doped InP. Samples encapsulated with Si3N4 films of about 1000Å were annealed in a halogen tungsten lamp RTA system under flowing N2 at different temperatures from 700 to 900°C for 5s. It has been found that Si*+P* dual implantations into InP can result in an enhanced activation, particularly significant at high dose of implantation. The maximum dopant activation and average electron mobility for Si*+P* dual implants at a dose of 1×1015/cm2 are 70% and 750cm2/vs, which corresponds to a peak electron concentration of 5×1019/cm3 while that for Si* single implant at the same dose are 29% and 870cm2/vs, which corresponds to a peak electron concentration of 1.2×10 19/cm3. The improvement of the electrical properties is discussed in terms of amphoteric behavior of silicon in InP.

We have prepared GexSi1-x/Si structures by implanting germanium into silicon <100> surfaces and thermally annealing. Implant energies were 100 and 150 keV and the total ion doses were between 1 and 10 × 1015 ions/cm2. Ion channeling studies of these structures indicates that after annealing, the germanium atoms substitute into the silicon lattice and create a pseu-domorphic strained layer. In addition, channeling results suggest a residual band of dislocations at or beyond the ion range in the as-annealed samples.

Double-crystal x-ray diffraction and topography, along with Raman spectroscopy and topography are used to study lattice reconstruction and carrier activation for pulsed laser annealed Si implanted GaAs. Although lattice strain is essentially eliminated, along with the production of carrier concentrations to about 3×1019 cm.3 at the center of laser annealed spots, incomplete removal of implant induced disorder and little dopant activation are observed in surrounding areas. Correlations of Raman and x-ray topographs suggest that concentric regions of partial melting, but without epitaxial regrowth, occur in the periphery of the laser annealed spots.

Recently, an enhanced leakage associated with ion implantation at polysilicon gate edges has been reported (1). In this study, TEM and SIMS characterization have been done to supplement electrical measurements in order to better understand the degradation. SIMS, XTEM, and TRIM analyses of arsenic implants through 7 nm gate oxides show considerable ion mixing. A roughening of the underlying silicon substrate leads to asperities which are believed to play an important role in enhancing leakage (2). XTEM analysis of a polysilicon gate edge also reveals the effects of volume expansion associated with the incorporation of both dopant and knock-on oxygen.

The damage behavior of <100>-Si implanted with P2+ and P+ ions at equivalent energies were investigated by 2MeV He* backscattering and channeling analysis. Different incident energies (25-90keV/atom) and intermediate doses (1013–1014/cm2) were used for the implantation with sample holder being kept at temperatures ranging from 77k to 483K. It has been shown that the damage created by P2* implants is always greater than that of P2+ implants when the dosage is below the threshold fluence at which amorphization takes place. This damage enhancement phenomenon is strongly related to implantation temperature. A striking damage enhancement induced by 90 keV/atom P2+ implants in the surface region of the sample was observed, and it has been attributed to the multiple collision effect between diatomic ions and host atoms.

In this work we measured the functional dependence of the solid phase epitaxial (SPE) regrowth rate, V, of amorphous silicon on the concentration of implanted arsenic (n-type) impurity, NAs. The SPE regrowth rates of self-ion amorphized layers in silicon wafers with (100) substrate orientation were measured by in situ, high precision, cw laser interferometry during isothermal annealing for temperatures from 470 °C to 580 °C, and concentrations in the range 1.5×1018cm−3 ≤NAs≤3.5×1020 cm−3.

In the concentration range 7×1018 cm−3≤NAs≤2.2×1019 cm3, selected from the medium dose sample, the SPE regrowth data satisfied a linear equation; V/Vi=1+NAS/Ni, where Ni(T) was fit to an Arrhenius form obtained from the temperature dependent intersections of the SPE regrowth rate data with the concentration axis and Vi(T) was the temperature dependent apparent intrinsic SPE regrowth rate at zero impurity concentration. A similar linear dependence was obtained earlier for boron (B) and phosphorus (P).

However, unlike B and P, an enhancement of SPE regrowth was observed for samples implanted with As in the concentration range 3×1018 cm−3 ≤NAS≤ 1.3×1019 cm−3, selected from the low dose sample. This result indicates that arsenic implanted at low dose levels has a higher fractional ionization in amorphous silicon than either boron or phosphorus implanted at the same dose.

In the high dose samples with arsenic concentrations ≤NAs 2.2×1019 cm−3, the SPE regrowth rate varied nonlinearly with NAS. The nonlinear function had a negative curvature similar to that observed previously for P.