Precipitation Of Atoms Research Articles

Cleaning Efficiencies of Various Chemical Solutions for Noble Metals such as Cu, Ag, and Au on Si Wafer Surfaces

In this study, the cleaning efficiency of various chemical solutions for noble metals such as Cu, Ag, and Au on Si wafer surfaces is investigated and the cleaning mechanism is interpreted based on the result and the metal adsorption mechanism. Cu adsorption behavior can be interpreted by both electrochemical metal particle deposition and inclusion in the metal‐induced oxide grown during contamination in solution. Based on the Cu deposition mechanism, the required characteristics of a cleaning solution for removal of Cu on Si wafer surfaces are the dissolution of Cu, etching ability for both Si and , and surface passivation with or surfactant for preventing the readhesion of dissolved Cu ions. Ag adsorption behavior is interpreted only by the electrochemical metal particle deposition and not by film inclusion. The oxide formation enthalpy of Ag is so high that Ag cannot be removed by just the dissolution process of the cleaning solution, which has a higher redox potential than that of Ag. In contrast, the Au adsorption mechanism can be described as an atomic Au precipitation on hydrophobic Si surfaces with a strong electrostatic force. It is very difficult to remove Au on Si wafer surfaces by wet cleaning technology and it is suggested that Au contamination unintentionally induced during processing be removed by a gettering process. As a result, it was concluded that the metal deposition behavior and the removal mechanisms for Cu, Ag, and Au in a solution are partly similar but fundamentally different, even though Cu, Ag, and Au are categorized in the same noble metal group which depends on their redox potential, electronegativity, and formation enthalpy of metal‐oxide. © 1999 The Electrochemical Society. All rights reserved.

Silver cluster formation in ion-exchanged waveguides: processing technique and phenomenological model

Silver-containing light waveguides, formed by ion-exchange process in glasses containing alkali oxides, were annealed in hydrogen atmosphere at temperatures in the range 150–250°C. The processing technique causes silver atom precipitation to form nanometer-size clusters with ∼2–10 nm distribution in sizes and a spatial distribution. The composites were measured by transmission electron microscopy, optical spectroscopies, m-lines technique, Rutherford backscattering energies, elastic recoil detection analysis and secondary ion mass spectrometry. The nano-sized cluster formation, accompanied by silver and sodium interdiffusion processes, after hydrogen permeation in the matrix, is not simply ascribed to a direct interaction with hydrogen, but to a more complex process involving a charge balancing mechanism during ion transport phenomena. Such a process is discussed in terms of a phenomenological model. Non-linear refractive indices of some samples were measured.

Ion-induced crystal damage during plasma-assisted MBE growth of GaN layers

Gallium nitride layers were grown by plasma-assisted molecular-beam epitaxy on (0001)-oriented sapphire substrates using an electron cyclotron resonance (ECR) and a radio frequency (rf) plasma source. An applied substrate bias was varied from -200 to +250 V, resulting in a change of the density and energy of nitrogen ions impinging the growth surface. The layers were investigated by high-resolution x-ray diffractometry and high-resolution transmission electron microscopy (HRTEM). Applying a negative bias during growth has a marked detrimental effect on the crystal perfection of the layers grown with an ECR plasma source. This is indicated by a change in shape and width of (0002) and (202\ifmmode\bar\else\textasciimacron\fi{}5) reciprocal lattice points as monitored by triple axis x-ray measurements. In HRTEM images, isolated basal plane stacking faults were found, which probably result from precipitation of interstitial atoms. The crystal damage in layers grown with a highly negative substrate bias is comparable to that observed for ion implantation processes at orders of magnitude larger ion energies. This is attributed to the impact of ions on the growing surface. None of the described phenomena was observed for the samples grown with the rf plasma source.

Differences in iron, manganese, and phosphorus binding in freshwater sediment vegetated withLittorella uniflora and benthic microalgae

Undisturbed sediment cores from an oligotrophic lake were percolated with artificial porewater to examine the effects of isoetid macrophytes, Littorella uniflora, and benthic microalgae on daily dynamics of sediment retention of phosphorus (P) by either iron (Fe) or manganese (Mn). Retention of Fe and Mn was observed due to oxidation processes mediated by oxygen release from L. uniflora roots and benthic microalgae. Therefore increased retention of P was observed because of P precipitation with oxidized Fe- and Mn-compounds. During light periods, the ratio between Fe and P precipitation in the sediment was positively correlated with the P uptake by L. uniflora (p < 0.001, r2 = 0.984). The atomic precipitation ratio between Fe and P was between 1 and 2. The ratio between oxidized Fe-compounds and Fe-bound phosphate in the sediment was positively correlated with the root density of L. uniflora (p < 0.001, r2 = 0.995). The ratio between Mn and P precipitation was higher (≥6) than the ratio between Fe and P. The role of benthic primary producers on P retention in freshwater littoral sediments is discussed.

The effect of second phase on the creep deformation of 6061Al matrix composites.

Creep behaviour of two types of 6061Al matrix composites, which were reinforced with 10 and 20 vol.% of alumina particles or 11 and 18 vol.% of alumina short fibers, have been investigated at temperatures ranging from 573 to 773 K. Creep curves, stress dependences of the minimum creep rate and the microstructure after creep deformation were examined. A quantitative analysis of creep curves was also carried out by the θ-projection method. The difference in the volume fraction of the reinforcing second phases does not affect the creep strength in both types of composites. Creep strength of the fiber reinforced composites is higher than that of the particle reinforced ones at 673 and 773 K. Parameters determined by θ-projection method are able to describe the creep curves of the composites quantitatively. It has been shown that the precipitation of the solute atoms influences the creep behaviour at 573 K in both composites. The existence of the soft regions around the second phase is suggested in the particle reinforced composite crept under low stress at 773 K.

Microstructure control in semiconductor metallization

The microstructure of semiconductor metallization is becoming increasingly important as linewidths decrease below 0.5 μm. At these dimensions, reliability and performance are greatly influenced by specific microstructural features rather than only by the average material properties. In this article, we address the prospects for controlling the microstructure of thin film interconnection metals as linewidths are predicted to decrease below 0.1 μm by the year 2010. First, we evaluate the sources of energy available to drive microstructure changes in thin films, both during and after deposition. The internal energy sources considered are grain boundaries, interfaces, surfaces, strain, solidification, crystallization, solute precipitation, and phase transformations, with energy densities ranging from less than 1 meV/atom to greater than 100 meV/atom. The external energy sources considered are particle bombardment during deposition, mechanical deformation, and radiation damage, which may deliver energies greater than 100 eV/atom. Second, we review examples of microstructure changes in terms of these energy sources. These examples include the dependence of Al–Cu and Ti fiber texture on the roughness of SiO2, orientation change and abnormal Cu grain growth coupled to the precipitation of Co in Cu–Co alloys, and in-plane orientation selection during phase transformation of TiSi2 in very narrow lines. A substantial degree of microstructure control is also achieved in films deposited with off-normal incidence energetic particle bombardment, which has been used to produce both in-plane and out-of-plane crystallographic orientations in metals (Mo, Nb), nitrides (AlN), and oxides (ZrO2). Drawing on these examples, we discuss the prospects for microstructure control in future semiconductor metallization with respect to the list of energy sources, the decreasing dimensions, and the changing fabrication processes. One mechanism in particular, discontinuous precipitation of supersaturated solute atoms, is highlighted as having a substantial amount of stored energy available to drive microstructure evolution, and may provide a means to more fully control the microstructure of semiconductor metallization.

Electro-Discharge Consolidation of Atomized High Strength Aluminum Powders

Electro-discharge consolidation (EDC) was carried out for atomized Al-based alloy powders produced at YKK. Input energies of 2.5 to 3.8 kJ/g produced compacts of 88 to 99% in theoretical density. The densification can be described by a simple relation of Δρ=11.8[ERs⁄(Rs+2)]0.36 in % where Rs is the specimen resistance, and it predominantly depends on input energy, E. A comparison of XRD spectra before and after EDC with respect to the peaks’ position and intensity indicates that no microstructural change occurred. Oxide film removal by EDC was confirmed by TEM observations. Compression tests of EDC compacts at room temperature produced under the present conditions (specimen resistance Rs=9.4–15.8 mΩ) yielded 800 MPa and 24% elongation for true ultimate compressive stress and true compressive strain, respectively. The annealing treatment at 573 K for 1.2 ks raiseds compressive strength by at expense of ductility due to precipitation of solute atoms, indicating that a super-saturated solid solution is maintained in the EDC state.

Effect of Cr implantation on the morphology of Al:Si films

Al:Si 1% films subjected to irradiation with Cr + ions are investigated for their morphology and appearance of hillocks. The oxidized silicon wafers were coated with 1 μm thick metallic layers by magnetron sputtering. Prior to sintering at 450°C for 30 min the films were implanted with chromium ions at an energy of 130 keV and to doses 4 × 10 15–9 × 10 15 cm −2. Investigations of the processed surface using SEM, alpha-step together with a procedure for digital evaluation of the surface morphology examined by an optical microscope show that chromium implantation leads to a significant reduction of the hillock population, in particular those of the largest size. The effect is observed for Al:Si and Al:Si:Cu alloys but not for pure Al. SIMS profiling gives evidence of measurable mobility of Cr atoms during the annealing. TEM analysis displays a noticeable reduction of the material grain size and reveals in Cr + implanted Al:Si layers an array of fine precipitates located within the grain volume. From the fact that the modification of grain-growth kinetics occurs only when Si is present in the film one may conclude that Cr atoms precipitate in Si-rich phases (binary or ternary). Other effects of Cr implantation into Al-based films are: reduction of the lattice constant and an increase in sheet resistance (by 10%). Formation of Cr-containing precipitates seems beneficial in suppressing overall diffusivity in the layer preventing thus the grains from excessive vertical growth.

Pseudomorphic Si1−xSnx alloy films grown by molecular beam epitaxy on Si

Strained Si1−xSnx (0.001≤x≤0.052) alloys were synthesized on (001) Si substrates by molecular beam epitaxy at 220 and 280 °C. The as-grown alloys were found to be pseudomorphic to Si with no indication of extended defects and tin precipitates. Within the accuracy of our studies the compressive strain in the alloys corresponds to that deduced from Vegard’s linear interpolation between the lattice parameters of Si and α-Sn. The annealing experiments show that the alloys are thermally unstable at a higher temperature (1000 °C) and that the transition of the Si1−xSnx/Si system to a lower energy state occurs through two channels: (i) alloy decomposition through precipitation of tin atoms into metallic β-Sn, and (ii) introduction of 60° misfit dislocations.

Structure and electrical transport in microcrystalline composite Si-NiSi 2 thin films

The thermally activated conductivity behavior in microcrystalline composite Si-NiSi 2 thin films depends significantly on their microstructure. Incorporated Ni atoms with concentrations from 0.2 to 1 at.% in as-deposited amorphous silicon thin films induce a crystallization process during thermal annealing. For concentrations below 0.5 at.% the Ni atoms precipitate in NiSi 2 crystals which induce a needle-like growth of Si crystallites in the amorphous matrix. In the following process strain-induced effects lead to a transition from the amorphous to a microcrystalline matrix. For Ni concentrations of 0.5 at.% and above the crystallization leads to a homogeneous microcrystalline phase. In both cases the electric transport is found to be due to combined grain boundary states.

Middle‐ and low‐latitude emissions from energetic neutral atom precipitation seen from ATLAS 1 under quiet magnetic conditions

During the ATLAS 1 mission spectral observations were made at middle and low latitudes of features expected from the precipitation of energetic neutral atoms. The Imaging Spectrometric Observatory was used at night in the UV and visible with maximum gain. The tangent ray heights of the look directions ranged from near 100 km to near 200 km, and the geomagnetic conditions were quiet during the observations, which were made March 28 to April 3, 1992. The N2+1N 391.4‐nm and O I 130.4 and 135.6‐nm emissions were observed at all latitudes, with lower emission rates at lower magnetic dip latitudes, except that enhancements in the O I lines were seen within 30° of the dip equator due to radiative recombination of ionospheric plasma. The latitude profile observed for the N2+1N emission did not show an equatorial or midlatitude peak. This implies that the source of energetic neutrals is more consistent with prompt charge exchange loss of freshly injected trapped ions with relatively low mirror heights (i.e., ions on higher L shells with equatorial pitch angle distributions nearly isotropic to the loss cone) than loss of highly eroded populations of particles with high mirror heights (i.e., ions on lower L shells with pancake equatorial pitch angle distributions). The N2+1N emission rates have been compared with models of atmospheric emission due to fluxes of O/O+ and H/H+ in the thermosphere, as produced by energetic neutral oxygen or hydrogen atom precipitation. Energy deposition rates are inferred that range from about 12 (±3) × 10−4 erg cm−2 s−1 near 45° geomagnetic latitude to about 3 (±1) × 10−4 erg cm−2 s−1 near the geomagnetic equator. These can be compared with energy deposition rates a factor of 10² higher during magnetic storms.

Relationship between energy flux Q and mean energy 〈E〉 of auroral electron spectra based on radar data from the 1987 CEDAR Campaign at Sondre Stromfjord, Greenland

The incoherent scatter radar at Sondre Stromfjord, Greenland, measured electron density profiles from 90 to 500 km during four auroral events over a 3‐hour period on February 28, 1987. The profiles were obtained with the radar pointed along the magnetic field near zenith at 15‐s intervals. Under the assumption that proton/H atom precipitation was unimportant during these events a representation of the incident electron flux was obtained by fitting calculated profiles with measured profiles in the vicinity of their peaks (lower E region). Maxwellian and Gaussian electron distributions with high‐ and low‐energy tails were used to generate the calculated profiles. The distributions were specified in terms of average energy 〈E〉 and energy flux Q. We find that we can clearly distinguish between profiles that result from a Maxwellian incident electron spectrum and those that result from a Gaussian spectrum. Interpreting Gaussian and Maxwellian spectra as representative of discrete and diffuse aurora, respectively, the measurements indicated good correlation between 〈E〉 and Q for discrete aurora, while essentially no correlation was observed for diffuse aurora. This is consistent with current understanding that discrete auroras are produced by electrons accelerated by magnetic field‐aligned potential drops whereas diffuse auroras are produced by pitch angle diffusion of plasma sheet electrons into the loss cone.

Dislocation structure formation due to ion implantation

A mechanism of dislocation structure formation and its dynamics is presented. Coalescence of the interstitial atoms leads to the formation of dislocation loops. The formed loops grow due to precipitation of the interstitial atoms. Interaction between loops results in their motion deep into material, where loops form a developed dislocation structure. 1. There are many experimental evidence of an influence of ion bombardment on the material structure and properties in the region disposed much deeper than the ion penetration depth. One of this long range effects is the dislocation structure formation at a depth up to about 100 microns.'-4 A number of mechanisms explaining that phenomenon has been suggested. However, a general theory of dislocation structure formation has not been found yet. A mechanism of dislocation structure formation and its dynamics in the implanted specimen is considered in the present paper. It is suggested that the coalescence of the interstitial atoms emerging due to ion bombardment results in the creation of dislocation loops. Then the loops move deep into the specimen because of interaction with each other. Collisions between the loops result in pinning and formation of immobile dislocation structure. 2. Under irradiation a lot of interstitial atoms and vacancies appear in the material. They are created in a narrow region at the ion penetration depth and then diffuse deep into the sample. The diffusion coefficient of interstitial atoms, D, is much greater than that of vacancies, D,. Therefore, at the depth z > (D\ty5 the concentration of interstitial atoms, c, exceeds that of vacancies, c,. This is the condition for the preferable formation of interstitial dislocation loops and their growth. The loops are formed by means of fluctuation. The critical number of atoms in the loop is equal to7

Misfit Accommodation in a Quenched and Aged Al-Cu Alloy with Silicon Particles

Quenched and aged specimens of the Al-1.3 at. pct Cu-19.1 at. pct Si alloy were studied by X-ray diffraction. Since this alloy contains a high volume percentage (~20 vol pct) of second- phase Si particles, it is regarded as a model for a metal matrix composite (MMC). During isothermal aging of the solid-quenched Al-1.3 at. pct Cu-19.1 at. pct Si alloy, the Cu and Si atoms precipitate. This causes the Al-rich phase lattice parameter to increase from a value lower to a value higher than the lattice parameter of pure unstrained aluminum. Due to thermal misfit after quenching from heat-treatment temperatures, all lattice parameters are influenced by re- sidual stresses. A model describing the elastic/plastic accommodation of a misfitting spherical inclusion in an infinite matrix is adapted for the case of misfitting inclusions in a finite matrix. This model describes the measured lattice parameter shifts of the Si phase reasonably well. Comparison of the model for elastic accommodation and the model for elastic/plastic accom- modation with measured stresses shows significant discrepancies for the low-temperature range (ΔT < 200 K). These discrepancies may be related to the volume effect of defects (dislocations, vacancies) created in the plastic zone.

Optically thin H Lyman alpha production on outer planets: Low‐energy proton Acceleration in parallel electric fields and neutral H atom precipitation from ring current

A Monte Carlo model has been constructed to describe the energization and energy degradation of low‐energy protons in an H2 atmosphere in the presence of parallel electric field. The energy spectrum of Ly α collision events is employed to determine the volume emission rate (VER) of H Ly α emissions. Numerical experiments have been performed to study the effect of initial proton energy, electric field, neutral number density, initial pitch angle, and cutoff limit on H Ly α VER. The present study demonstrates that energization of solar EUV‐generated low‐energy protons by parallel electric fields is incapable of producing optically thin Ly α emissions on Uranus through direct collision with H2. However, nonthermal H atoms, produced through acceleration of protons in parallel electric fields, play an important role in enhancing the Ly α intensity through resonant scattering of solar Ly α flux: resulting emissions are consistent with the broadening of the Ly α line observed on Jupiter by IUE. The Monte Carlo model is also applied to the problem of Doppler‐shifted H Ly α emissions from the auroral atmosphere of Jupiter. Energetic neutral H precipitation from the ring current is discussed as a possible source of optically thin H Ly α emissions on outer planets. Estimates are also given for Hα, Hβ, and H2 band emissions resulting from the impact of fast H (of ring current origin) on H2. The input energies and fluxes of precipitating ring current particles are, however, too uncertain to permit definite conclusions.

Implants of aluminum into silicon

The electrical behaviour of ion implanted aluminum into silicon was investigated by varying the beam energy in the 80 keV-6 MeV range, the dose in the 1 × 10 13–1 × 10 14/cm 2 range and the annealing procedure. Aluminum atoms precipitate into exten defects at the end of range damage and where the concentration exceeds the solid solubility value (about 2 × 10 19/cm 3 at 1200°C Escape of Al atoms occurs very easily as soon as they reach the external surface during the thermal diffusion. Using high energy implants, 6 MeV, it was possible to follow in detail the broadening of the diffused profiles. The measured trends between the retained dose and the junction depth and between the outdiffused dose and the annealing time are quite well predicted by the solution of the diffusion equation with the surface acting as a perfect sink for the dopant.

Diamond hexagonal silicon phase and {113} defects Energy calculations and new defect models

The hypothesis that the precipitation of self-interstitial Si atoms leads to the formation of the diamond hexagonal Si phase, inside the structure of {113} stacking faults and rod-like defects, is ...

Enhanced diffusion of liquid-coated gas bubbles in solids

Abstract Gas bubble diffusion in solids is typically due to surface diffusion of matrix atoms along the gas-solid interface. Insoluble impurities that segregate to the interface tend to suppress surface diffusion by chemically combining with the surface atoms, as in the creation of a metal oxide layer, or by simply presenting a physical barrier to the kinetic jump process. We propose a mechanism by which surface self-diffusion (and hence gas bubble diffusion) may in fact be enhanced during high-temperature anneals, for a special class of matrix-impurity couple. A liquid layer at the gas-solid interface, comprised of the impurity species and an equilibrium concentration of matrix atoms, acts as a conduit for rapid mass transport, as equilibrium is maintained by continuous dissolution and precipitation of matrix atoms at the solid surface. This liquid dissolution diffusion process thus requires that the impurity species melt at a temperature lower than the annealing temperature, that the liquid lower the ga...

A study by ion-induced electron emission of silicon surfaces irradiated by keV Li+ ions

We present experimental results of the secondary electron emission induced by energetic lithium ions from Si(111) surfaces. The ion dose dependence of the electron emission coefficient, γ, and of the different structures which appear on the electron energy distribution were studied from samples heated up to 250°C. The measurements were performed for normal and oblique incidence and for different ion energies. All the results demonstrate that first implanted Li atoms precipitate partially at the silicon surface. The quantity of Li particles trapped at the surface increases with the ion dose until a saturation value which essentially depends on the target temperature and to a lesser degree on the angle of incidence and the ion energy.

Co diffusion and precipitation in Si/SiGe heterostructures

Several Si/Si1−xGex multilayer structures, grown by molecular-beam epitaxy (MBE) or synthesized by high-dose Ge implantation into Si substrates, have been implanted with Co atoms at doses of typically 1.0×1015 and 1.0×1016 ions/cm2. These samples are annealed at elevated temperatures (typically 850 °C) to study the diffusion and precipitation of the Co atoms. Depth distributions of Co atoms are measured with secondary-ion mass spectrometry. X-ray diffraction is used to measure the strain evolution of the Si/Si1−xGex MBE multilayers during processing. Channeling Rutherford backscattering spectrometry is used to study the lattice position of the Co atoms after heat treatment. Upon annealing, Co atoms are observed to diffuse out of the Si1−xGex layers. The efficiency of this outdiffusion process strongly depends on the implantation dose. The released Co atoms are gettered by the Si layers adjacent to the Si1−xGex layers and, in case of the MBE samples, at the interface between the substrate and MBE-grown bufferlayer. A physical model to explain these observations is presented.