Cascade Sputtering Research Articles

Atomic-level insight into process and mechanism of ion beam machining on aluminum optical surface

Polycrystalline Al workpieces go through a complex surface atom sputtering and surface/subsurface atom diffusion throughout the ion beam machining process that plays a critical role in determining machining quality of optical mirror surfaces. Here, we leverage molecular dynamics method for the first time to reveal the machining mechanism and the polycrystalline effect of aluminum during ion beam sputtering, in term of cascade collision, atom trajectory, sputtering yield, and surface characteristic. Polycrystalline Al presents a significant low potential energy state, leading to milder cascade collision and weaker sputtering effects than monocrystalline Al during ion beam machining. The atom trajectory demonstrates irregular variation as the atom diffusion is blocked by grain boundary (GB) in polycrystalline Al, leading to the relief microstructure and poor surface quality. With incident ions increasing, the GBs are broken and atom diffusion enhancing, as well as lower subsurface defects and better finishing surface quality. The microscopic morphology evolves into gravel microstructure. Simulation results also reveal that atom diffusion will benefit from high ion concentration and low ion energy. Sputtering yield variation is more sensitive to ion energy. To acquiring better finishing surface quality of Al without influencing machining efficiency, ion concentration must be increased while the ion energy needs to be decreased appropriately but a large number of cascading collisions need to be ensured.

Angular Dependences of Silicon Sputtering by Gallium Focused Ion Beam

Angular dependences of the surface layer composition and the sputtering yield of silicon upon irradiation of the surface with a focused beam of gallium ions with an energy of 30 keV are obtained. The surface composition is analyzed by scanning Auger electron spectroscopy (SAES) and secondary ion mass spectrometry (SIMS). The sputtering yields are determined by measuring the volume of sputtering craters and irradiation doses. It is found that the content of gallium in the surface layer is about 30 at % with incidence angles close to the normal. With incidence angles greater than 30°, the concentration of gallium decreases quite sharply. The angular dependence of the sputtering yield of silicon does not correlate with the content of gallium in the surface layer and is rather well described by the cascade sputtering mechanism proposed by P. Sigmund.

Modeling of the O+pickup ion sputtering efficiency dependence on solar wind conditions for the Martian atmosphere

AbstractSputtering of the Martian atmosphere by O+pickup ions has been proposed as a potentially important process in the early evolution of the Martian atmosphere. In preparation for the Mars Atmosphere and Volatile Evolution (MAVEN) mission, we performed a study using a Monte Carlo model coupled to a molecular dynamic calculation to investigate the cascade sputtering effects in the region of the Martian exobase. Pickup ion fluxes based on test particle simulations in an MHD model for three different solar wind conditions are used to examine the local and global sputtering efficiencies. The resultant sputtering escape rate is 2×1024 s−1at nominal solar wind condition and can be enhanced about 50 times when both the interplanetary magnetic field (IMF) strength and the solar wind pressure increase. It is found that when the IMF strength becomes stronger, both the pickup ion precipitation energies and the resultant sputtering efficiencies increase. The related escape flux, hot component, and atmospheric energy deposition deduced from the MAVEN measurements may reveal clues about the prominent enhanced sputtering effects. Significant hemispheric asymmetries can be observed related to the solar wind electric fields. The shielding by the crustal fields and the recycling onto the nightside due to different magnetic field draping features can also lead to regional variations of sputtering efficiencies. The results suggest that disturbed or enhanced solar wind conditions provide the best prospects for detecting sputtering effects for MAVEN mission.

Collision cascade temperature

Interaction of a projectile with a solid has been considered in detail. It has been found that any collision cascade generated by a projectile can be characterized by the average kinetic energy of cascade atoms that represents an “instantaneous temperature” of the cascade during its very short lifetime (10 −12 s). We refer to this value as the “dynamic temperature” in order to emphasize the fact that cascade atoms are in a dynamic equilibrium and have a definite energy distribution. The dynamic temperature defines the electron distribution in the cascade area and, hence, the ionization probability of sputtered atoms. The energy distribution of cascade atoms and, as a consequence, the dynamic temperature can be found experimentally by measuring the energy distribution of sputtered atoms. The calculated dynamic temperature has been found to be in good agreement with the experimental data on ion formation in the case of cesium and oxygen ion sputtering of silicon. Based on the developed model we suggest an experimental technique for a radical improvement of the existing cascade sputtering models.

Influence of the negative oxygen ions on the structure evolution of transition metal oxide thin films

The energy distributions of O− ions of magnetron sputtered Nb, Ta, Zr, and Hf in an Ar∕O2 atmosphere were measured as a function of the oxygen partial pressure. Three ion populations were detected in the plasma: high, medium, and low energy ions, with energies corresponding to the target potential, half of the target potential, and <150eV, respectively. The ion energy distribution functions were compared to distributions obtained based on Sigmund’s linear collision cascade sputtering theory. If the surface binding energy is assumed to be equal to the heat of formation, good agreement between the experiment and theory was achieved. From correlating the measured ion energy distributions with previously published phase stability data [Ngaruiya et al., Appl. Phys. Lett. 85, 748 (2004)], it can be deduced that large fluxes of medium and high energy O− ions comparable to the fluxes of the low energy O− ions enable formation of crystalline transition metal oxide thin films during low temperature growth. The presented data here may be of general relevance for understanding the structure evolution of thin oxide films.

Kinetic energy distributions of neutral In and In 2 sputtered by polyatomic ion bombardment

Kinetic energy distributions of neutral In monomers and In 2 dimers sputtered from a polycrystalline indium surface under bombardment with 5 keV/atom Au 1 − and Au 2 − projectiles have been investigated by means of laser postionization time-of-flight mass spectrometry. Results show that 5 keV Au 1 bombardment leads to results in full compliance with linear cascade sputtering theory. For polyatomic ion bombardment, we find a clear transition to a collisional spike dominated emission process. The spike contribution appears as a low-energy part in the sputtered flux which increases with increasing projectile nuclearity and energy. We show that, the velocity spectrum associated with the low-energy contribution is virtually identical for sputtered monomers and dimers. This finding has important implications with respect to the particle emission mechanism under polyatomic projectile bombardment.

Atomic excitations during ion beam sputtering of YBa2Cu3O7 targets

YBa2Cu3O7 targets were bombarded with 1–10keV Ar+ beam, and excitation of the sputtered particles was studied with the help of optical spectroscopy in 300–700nm wavelength range. Intense emission from several YI, BaI and CuI transitions was observed, but essentially no emission from sputtered excited atomic and molecular oxygen. Photon yields of various emission lines were measured as a function of the ion energy and results were compared with a model based on the collision cascade sputtering mechanism and an excitation process that depend on velocity of the ejected particle. Moreover, we observed decrease of the photon emission with ion fluence that may be due to depletion of the oxygen in surface region.

Sputtering of indium usingAumprojectiles: Transition from linear cascade to spike regime

We have investigated the yields and emission velocity distributions of neutral In atoms and In2 dimers sputtered from a pure indium surface under bombardment with Aum − m=1,2 ,3 projectile ions. It is shown that 5-keV Au1 bombardment leads to results in full compliance with linear cascade sputtering theory. All polyatomic projectiles are found to generate an additional low-energy contribution to the sputtered flux, which increases with increasing projectile nuclearity and energy and completely dominates the spectra under 10-keV Au3 bombardment. Analysis shows that this contribution cannot be explained in terms of thermal spike sputtering models. Instead, the results indicate a spike emission mechanism, which closely resembles a free expansion of a supercritically heated subsurface volume.

Cluster formation at metal surfaces under bombardment with SFm+ (m=1,…,5) and Ar+ projectiles

We investigate the formation of ionic and neutral clusters emitted from a polycrystalline indium surface under bombardment with SF m + ( m = 1, …, 5) and Ar + projectile ions at 10 keV impact energy. Mass spectra of secondary ions and sputtered neutral particles are recorded under otherwise identical conditions. The neutral species are post-ionized prior to mass analysis by means of single photon-ionization using an intense UV laser at a wavelength of 193 nm. It is found that the measured secondary ion signals increase much more than those of the corresponding neutral particles if SF m + projectiles are used instead of Ar + ions, indicating that the ionization probability under bombardment with SF m + is enhanced by a chemical matrix effect induced by fluorine incorporation into the surface. Interestingly, the largest values of the ionization probability are observed for SF 3 + projectiles. The total sputtering yield is found to be larger for SF m + compared to Ar + projectiles and to increase linearly with increasing m . Both findings are shown to be understandable in the framework of linear cascade sputtering theory. The partial sputtering yields of In n clusters exhibit a stronger enhancement than the sputtered monomers, the magnitude of the effect increasing with increasing cluster size and projectile nuclearity.

Cluster formation at metal surfaces under bombardment with [formula omitted] (m = 1, …, 5) and Ar+ projectiles

We investigate the formation of ionic and neutral clusters emitted from a polycrystalline indium surface under bombardment with SFm+ (m=1,…,5) and Ar+ projectile ions at 10 keV impact energy. Mass spectra of secondary ions and sputtered neutral particles are recorded under otherwise identical conditions. The neutral species are post-ionized prior to mass analysis by means of single photon-ionization using an intense UV laser at a wavelength of 193 nm. It is found that the measured secondary ion signals increase much more than those of the corresponding neutral particles if SFm+ projectiles are used instead of Ar+ ions, indicating that the ionization probability under bombardment with SFm+ is enhanced by a chemical matrix effect induced by fluorine incorporation into the surface. Interestingly, the largest values of the ionization probability are observed for SF3+ projectiles. The total sputtering yield is found to be larger for SFm+ compared to Ar+ projectiles and to increase linearly with increasing m. Both findings are shown to be understandable in the framework of linear cascade sputtering theory. The partial sputtering yields of Inn clusters exhibit a stronger enhancement than the sputtered monomers, the magnitude of the effect increasing with increasing cluster size and projectile nuclearity.

Angular distribution of atoms sputtered from silicon by 1–10 keV Ar ions

The angular distribution of atoms sputtered from silicon under 1–10 keV Ar ion bombardment (normal incidence) has been studied experimentally and using computer simulation. It has been found that the angular distribution is overcosine in the whole energy range studied. This is connected with a high contribution of secondary knock-on atoms to the sputter flux (cascade sputtering). The simulated angular distributions are shown to be quite sensitive to the variation of the interatomic potential. The results obtained are compared with data from the literature.

Non-additivity of multiply charged ion emission from Si and Al produced by molecular projectiles

Experiments are reported on the non-additivity of a multiply charged ion emission with various atomic and molecular projectiles. A distinction has been found in the non-additivity of the Si 2+ emission from Si. It manifests itself as a change of the non-additivity factor K 2,1 from K 2,1<1 (sub-linear effect) to K 2,1>1 (non-additive enhancement effect) in going from Au 2 − to Cu 2 − projectiles. It was shown that the sub-linear behaviour of K 2,1 is caused by the correlated move of projectiles gold atoms in a target matter (the clearing-the-way effect) and the origin of the non-additive enhancement is not determined by non-linear cascade sputtering. On the basis of the results obtained a new effect in an ion-induced atomic-like Auger electron emission has been predicted.

Ionized sputter deposition using an extremely high plasma density pulsed magnetron discharge

Time resolved plasma probe measurements of a novel high power density pulsed plasma discharge are presented. Extreme peak power densities in the pulse (on the order of several kW cm−2) result in a very dense plasma with substrate ionic flux densities of up to 1 A cm−2 at source-to-substrate distances of several cm and at a pressure of 0.13 Pa (1 mTorr). The pulse duration was ∼100 μs with a pulse repetition frequency of 50 Hz. The plasma consists of metallic and inert gas ions, as determined from time resolved Langmuir probe measurements and in situ optical emission spectroscopy data. It was found that the plasma composition at the beginning of the pulse was dominated by Ar ions. As time elapsed metal ions were detected and finally dominated the ion composition. The effect of the process parameters on the temporal development of the ionic fluxes is discussed. The ionized portion of the sputtered metal flux was found to have an average velocity of 2500 m s−1 at 6 cm distance from the source, which conforms to the collisional cascade sputtering theory. The degree of ionization of the sputtered metal flux at a pressure of 0.13 Pa was found to be 40%±20% by comparing the total flux of deposited atoms with the charge measured for the metal ions in the pulse.

Trends in sputter yield data in the film deposition regime

Knowing the fundamental trends of sputter yield in the sputtering regime used for film deposition is crucial for understanding the mechanisms of sputter deposition, interpreting yield data, and predicting the results of experiments. One source of experimental yield data is the comprehensive compilation [N. Matsunami, Y. Yamamura, Y. Itikawa, N. Itoh, Y. Kazomata, S. Miyagawa, K. Morita, R. Shimizu, and H. Tawara, At. Data Nucl. Data Tables 31, 1 (1984)]. They published graphical yield-energy curves based on empirical parameters which were best fits to nearly all the projectile-target combinations available in 1983. We interpret this experimental data using theoretical results from a ``simplified collisional model'' of sputtering. For 1-keV noble gas projectiles, several trends show forth (some old and some new): First, there is a repeating pattern in the dependence of yield on target atomic number, with the period being each row of the Periodic Table. Second, surface binding energy is the single most important target parameter; the yield varies roughly as ${1/U}_{\mathrm{sb}}^{1.3}$ for the empirical data, but as ${1/U}_{\mathrm{sb}}^{0.5}$ for the pure linear cascade sputtering mechanism. Third, while the principal mechanism is the linear cascade, the nonlinear cascade makes a detectable contribution to the experimental yield for yield values above \ensuremath{\sim}1. Fourth, for target atomic numbers above \ensuremath{\sim}35, the yield increases monotonically with projectile mass; for lighter targets the yield exhibits a maximum at an intermediate projectile mass. Fifth, the energy dependence of yield for a given projectile-target combination from \ensuremath{\sim}0.5 to 2 keV is, to a good approximation, $Y(E)\ensuremath{\propto}{E}^{0.5}.$ Scatter due to experimental error is evident in the data: Matsunami et al.'s approach of combining the results of different experiments, resulting in their empirical yield curves, is useful for compensating for this.

Influence of adsorbates, crystal structure, and target temperature on the sputtering yield and kinetic-energy distribution of excited Ni atoms

We present results for the emission of secondary Ni atoms in the electronic ground state and some excited metastable states under ion bombardment. In particular, we address some of the most urgent questions resulting from previous measurements of the Ni system: the influence of adsorbates, in particular oxygen, on the excitation, the influence of the crystal structure and orientation, and possible temperature effects. A direct comparison between sputtering under identical experimental conditions from polycrystalline and single crystal targets has been performed. Inelastic processes have been found to play an important role in determining the population of the different electronic levels of sputtered atoms. The density of states of the solid and the overlap of the electronic wave functions of the atom and the solid (bulk or surface) determine the efficiency of the population of individual atomic levels and the energy distribution of the atoms in these electronic states. Analyzing the experimental data, the contribution of inelastic processes to elastic cascade sputtering for clean, oxidized, and carbonized surfaces, at room temperature and up to $800\ifmmode^\circ\else\textdegree\fi{}\mathrm{C}$ as well as for various azimuthal angles could be determined and has finally led to a coherent description of the sputtering-excitation process for low-lying excited Ni atoms: the emission of an ionic core and subsequent electron capture into excited states.

Effects of cascade mixing, sputtering and diffusion by high dose light ion irradiation of boron nitride

The change of the BN surface composition resulting from sputtering, cascade mixing and diffusion is studied for large fluences of 0.5–20 keV D + and He +. The calculations are carried out with both the dynamic Monte-Carlo CASNEW-D code and a diffusion model. A system of equations for bound and unbound N atoms is solved taking into account the generation of N interstitials in cascade, recombination with vacancies and capture of that with traps. The boron enrichment of the BN surface layer observed experimentally could be obtained, if the easy evaporation of unbound N from the surface is assumed. The calculated N atom surface layer depletion depends on the kind and energy of ion, the exposure dose, ion current density as well as on the recombination rate with vacancies and the target temperature.

Io's Sodium Corona and Spatially Extended Cloud: A Consistent Flux Speed Distribution

A data set composed of different groundbased observations for Io's sodium corona and spatially extended sodium cloud and covering the spatial range from Io's nominal exobase of 1.4 satellite radii to east–west distances from Io of ±100 satellite radii ( R Io) is used to investigate the velocity distribution of sodium at the exobase. The data set is composed of the novel 1985 eclipse measurements of Schneider et al.(1991, Astrophys. J.368, 298–315) acquired from ∼1.4 to ∼10 R Io, the 1985 east–west emission data of Schneider et al.acquired from ∼4 to ∼40 R Io, and sodium cloud image data acquired near Io's orbital plane from ∼10 to ∼100 R Ioby a number of different observers in the 1976 to 1983 time frame. A one-dimensional east–west profile that contains Io is constructed from the data set and is analyzed using the sodium cloud model of Smyth and Combi (1988, Astrophys. J. Supp.66, 397–411; 1988, Astrophys. J.328, 888–918). When the directional feature in the trailing cloud is either north or south of this east–west line (i.e., not at the null condition), an isotropic modified [incomplete (α = 7/3) collisional cascade] sputtering flux speed distribution at the satellite exobase with a peak at 0.5 km sec −1provides an excellent fit to the data set for a sodium source of 1.7 × 10 26atoms sec −1. In particular, the model calculation reproduces (1) the essentially symmetric column density distributions exhibited by the eclipse measurements about Io within the Lagrange sphere radius (5.85 R Io, i.e., the gravitational grasp of the satellite), (2) the change in the slope of the column density observed just beyond the Lagrange sphere radius in the east–west profile of the forward cloud, but not in the trailing cloud, and (3) the distinctly different east–west brightness profiles exhibited by the forward and trailing clouds in the emission data at the more distant (∼ ±20–100 R Io) portions of the cloud. In contrast, the speed dispersion at the exobase for either an isotropic Maxwell–Boltzmann flux speed distribution or an isotropic classical (α = 3) sputtering flux speed distribution (which has a higher velocity-tail population than the Maxwell–Boltzmann, but not as high as the incomplete collisional cascade sputtering distribution) is shown to be inadequate to fit the data set. To fit the enhanced trailing east–west profile observed when the directional feature is at the null condition, an additional enhanced high-speed (∼15–20 km sec −1) sodium population is required which is nonisotropically ejected from the satellite exobase so as to preferentially populate the trailing cloud. The need for such a nonisotropic high-speed population of sodium has also been recognized in the earlier modeling analysis of the directional features (Pilcher et al., 1984, Astrophys. J.287, 427–444), in the more recent lower-velocity component required in modeling the sodium zenocorona (Smyth and Combi, 1991, J. Geophys. Res. 96, 22711–22727; Flynn et al., 1992, Icarus99, 115–130), and in the very recent modeling of the directional feature reported by Wilson and Schneider (1995, Bull. Am. Astron. Soc.27, 1154). A complete sodium source rate speed distribution function at Io's exobase from 0–100 km sec −1is then constructed by combining the isotropic modified [incomplete (α = 7/3) collisional cascade] sputtering flux speed distribution, the nonisotropic directional feature (lower-velocity zenocorona) source (∼15–20 km sec −1) , and the higher-speed (∼20–100 km sec −1) charge-exchange source required to simulate the sodium zenocorona far from Jupiter.

Binding energy effects in cascade evolution and sputtering

The marlowe model was extended to include a binding energy dependent on the local crystalline order: atoms are bound less strongly to lattice sites near surfaces or damage. Sputtering and cascade evolution were studied in self-ion irradiations of Cu and Au monocrystals. In cascades, the mean binding energy is reduced ∼8% in Cu with little dependence on the initial energy; in Au, it is reduced ∼9% at 1 keV and ∼15% at 100 keV. In sputtering, the mean binding energy is reduced ∼8% in Cu and ∼15% in Au with little energy dependence; the yields are increased less, averaging about half as much. Most sites from which atoms are sputtered are isolated in both metals. Small clusters of sites occur in Cu, but there are some large clusters in Au, especially in {111} targets. There are always more large clusters with damage-dependent binding than with a constant binding energy, but only a few clusters are compact enough to be regarded as pits.

Sputtering of compound semiconductor surfaces. I. Ion-solid interactions and sputtering yields

Abstract Several phenomena occur on the surface of a solid when being bombarded by energetic ions. A short general review is given of the major ion-solid interactions on compound semiconductor surfaces. An in-depth discussion is presented of the total sputtering yields of component semiconductors. For this discussion, GaAs is assumed to be the prototype compound semiconductor because most experimental measurements exist for GaAs. To exclude any chemical effects in the sputter yields, only the total sputtering yield data for argon ion bombardment of GaAs are compared with the predictions of the major sputtering theories, with particular attention to the Sigmund theory for linear cascade sputtering. Different proposals of each of the parameters in this theory are presented and compared with the GaAs data. These parameters are the surface binding energy, the nuclear stopping power, and the factor α, which represents the fraction of energy available for sputtering. Use of the different parameters results in a...

Sputtering yield increase with target temperature for Ag

Collision cascade sputtering may be influenced by the target temperature due to a modification of the collision cascade as well as by changes of the surface binding energy or heat of sublimation. Calculations with the computer simulation program TRIM.SP have shown that already a small decrease of the surface binding energy has a strong nonlinear effect on the sputtering yield, especially at energies near the threshold energy for sputtering. For silver the heat of sublimation near the melting point is about 2.7% lower than at room temperature. However, an increase of the sputtering yield of about 36% for 60 eV deuterium ion bombardment at normal incidence was found. Due to the large sublimation rate of Ag at temperatures above about 850 K sputtering yields measured at these temperatures show large uncertainties, and it will be difficult to experimentally confirm the increase of the collisional sputtering yield with temperature.