Transient Sputtering Research Articles

Depth profiling and standardization from the back side of a sample for accurate analyses: Emphasis on quantifying low-fluence, shallow implants in diamond-like carbon.

Back-side thinning of wafers is used to eliminate issues with transient sputtering when analyzing near-surface element distributions. Precise and accurate calibrated implants are created by including a standard reference material during the implantation. Combining these methods allows accurate analysis of low-fluence, shallow features even if matrix effects are a concern. Implanted Na (<2.0 × 1011 ions/cm2 , peaking <50 nm) in diamond-like carbon (DLC) film on silicon (solar wind returned by NASA's Genesis mission) was prepared for measurement as follows. Implanted surfaces of samples were epoxied to wafers and back-side-thinned using physical or chemical methods. Thinned samples were then implanted with reference ions for accurate quantification of the solar wind implant. Analyses used a CAMECA IMS 7f-GEO SIMS in depth-profiling mode. Back-side-implanted reference ions reduced the need to change sample mounts or stage position and could be spatially separated from the solar wind implant even when measuring monoisotopic ions. Matrix effects in DLC were mitigated and the need to find an identical piece of DLC for a reference implant was eliminated. Accuracy was only limited by the back-side technique itself. Combining back-side depth profiling with back-side-implanted internal standards aides quantification of shallow mono- and polyisotopic implants. This technique helps mitigate matrix effects and keeps measurement conditions consistent. Depth profile acquisition times are longer, but if sample matrices are homogeneous, procedural changes can decrease measurement times.

Fluence-dependent sputtering yield of micro-architectured materials

We present an experimental examination of the relationship between the surface morphology of Mo and its instantaneous sputtering rate as function of low-energy plasma ion fluence. We quantify the dynamic evolution of nano/micro features of surfaces with built-in architecture, and the corresponding variation in the sputtering yield. Ballistic deposition of sputtered atoms as a result of geometric re-trapping is observed, and re-growth of surface layers is confirmed. This provides a self-healing mechanism of micro-architectured surfaces during plasma exposure. A variety of material characterization techniques are used to show that the sputtering yield is not a fundamental property, but that it is quantitatively related to the initial surface architecture and to its subsequent evolution. The sputtering yield of textured molybdenum samples exposed to 300eV Ar plasma is roughly 1/2 of the corresponding value for flat samples, and increases with ion fluence. Mo samples exhibited a sputtering yield initially as low as 0.22±5%, converging to 0.4±5% at high fluence. The sputtering yield exhibits a transient behavior as function of the integrated ion fluence, reaching a steady-state value that is independent of initial surface conditions. A phenomenological model is proposed to explain the observed transient sputtering phenomenon, and to show that the saturation fluence is solely determined by the initial surface roughness.

Measurement of transient sputtering of Mo and W by Al and B ions injected by laser ablation in PISCES-A

The development of a transient sputtering technique in PISCES-A plasmas is reported. A Nd:YAG laser is used to ablate impurities from a sample placed outside the plasma column, delivering a short (∼10μs) pulse of impurities into the plasma. The injected impurity ions travel down the plasma column and cause a brief (∼10–100μs) pulse of line emission from sputtered target material. This technique offers some advantages over steady-state sputtering experiments: a wide range of impurity ions can be injected, and incoming impurity ion and sputtered atom velocities can be inferred from time of flight measurements. Measurements of Mo and W sputtering by Al2+ and B+ ions are presented. The Mo sputtering yields are found to be about 5–10times lower than expected, consistent with steady state measurements of sputtering of Mo by He+ under the same conditions. W sputtering is also lower than predicted.

Molecular Dynamics Simulations of Ar+–Organic Polymer Interactions

AbstractMD simulations of ion–organic styrene‐containing polymer interactions are reviewed and compared to experiment. We report results for argon ion bombardment of PS, PαMS and P4MS. All three polymers exhibit the formation of a similar, highly cross‐linking, dehydrogenated near‐surface damaged layer at steady state, but small changes in the structure of the polymer (P4MS and PαMS are isomers) can lead to drastic changes in the initial transient sputtering of the material. We correlate this behavior to differences in radiation chemistry (P4MS and PS are cross‐linking while PaMS is a chain scission polymer), and examine how the behavior in MD may relate to larger‐scale experimental results, such as roughness formation.magnified image

Incident angle dependence in the transient sputtering of an amorphous Si surface by sub‐keV O2+ ion bombardment

AbstractThe transient sputtering of an amorphous Si surface with sub‐keV oxygen ion bombardment was studied with medium energy ion scattering (MEIS) spectroscopy and dynamic Monte Carlo simulation. The in‐depth composition profiles of the incorporated oxygen and sputtering yield change in Si were obtained from in situ medium energy ion scattering spectroscopy. As ion doses increase, the Si surface is thickly oxidized, and the sputtering yield is rapidly reduced. At the initial stage of the sputtering, the surface is continuously swelled owing to an incorporation rate of oxygen higher than the sputtering rate of Si. A dynamic Monte Carlo calculation, which takes into account swelling and diffusion effects, describes the experimental results for the in‐depth composition profiles and the transient Si sputtering yield very well. At normal incidence, the surface is swelled rather than etched due to incorporation rate of oxygen atoms higher than surface etching rate, but at grazing incidence, the transient effect can be minimized. Copyright © 2006 John Wiley &amp; Sons, Ltd.

Transient sputtering of an amorphous Si surface under low energy O2+ ion bombardment

The transient sputtering of an amorphous Si surface with 500eV O2+ ion bombardment at surface normal incidence was studied via the in-depth composition profiles for the incorporated oxygen and the sputtering yield change in Si, which were obtained from in situ medium energy ion scattering spectroscopy. As ion doses are increased, the Si surface is thickly oxidized, and the sputtering yield is rapidly reduced. At the initial stage of the sputtering, owing to an incorporation rate of oxygen higher than the sputtering rate of Si, the surface is gradually swelled by the ion dose of ∼1.5×1016 O2+ions∕cm2. A dynamic Monte Carlo calculation, which takes into account both swelling and diffusion effects, has also been carried out to understand the underlying mechanism of the transient processes. In particular, we calculated the sputtered depth as a function of ion dose, and the secondary ion mass spectroscopy depth profiling showed a transient shifting of 2.7nm to the thicker depth.

Response function during oxygen sputter profiling for deconvolution of boron spatial distribution

We have studied the spatial distribution of a surface deposited boron layer by secondary ion mass spectrometry (SIMS) using oxygen beams of various incident energies. A boron layer around 0.4 nm thick was deposited by using electron-gun evaporation combined with liquid nitrogen cooling of the target. The SIMS boron profiles can be approximated closely by exponential-like distributions with decay lengths that are linear functions of the incident energy of the oxygen beam. Furthermore, our studies show that corrections to the depth and yield scales due to transient sputtering near the Si surface region must be considered for a plausible deconvolution of the original boron distribution.

Response function during oxygen sputter profiling and its application to deconvolution of ultrashallow B depth profiles in Si

The secondary ion mass spectrometry (SIMS) response function to a B “δ surface layer” has been investigated. Using electron-gun evaporation combined with liquid nitrogen cooling of target, we are able to deposit an ultrathin B layer without detectable island formation. The B spatial distribution obtained from SIMS is exponentially decaying with a decay length approximately a linear function of the incident energy of the oxygen during the SIMS analysis. Deconvolution with the response function has been applied to reconstruct the spatial distribution of ultra-low-energy B implants. A correction to depth and yield scales due to transient sputtering near the Si surface region was also applied. Transient erosion shifts the profile shallower, but beam mixing shifts it deeper. These mutually compensating effects make the adjusted distribution almost the same as original data. The one significant difference is a buried B peak observed near the surface region.

Transient sputtering of silicon by argon studied by molecular dynamics simulations

Molecular-dynamics simulations of Si(0 0 1) targets subject to uninterrupted Ar bombardment of various energies and angles of incidence were performed, in order to study the atomic processes that take place during the transient and steady-state sputtering regimes. Silicon interactions were described with the modified embedded atom method potential and the Stillinger–Weber potential. It is found that, although these potentials both give a reasonably good description of Si, they lead to significant differences in sputter behavior. Briefly said, for MEAM the sputter yield and the fraction of sputtered dimers are higher, frequent sputter bursts occur, the ‘amorphized’ zone is in fact a patchwork of semi-crystalline regions and voids, the average coordination number in amorphized zone is lower, the surface is rougher, and the transient stage lasts longer. All these issues are discussed and possible explanations are proposed.

Study of a pulsed glow discharge ion source for time-of-flight mass spectrometry

This paper describes a new type of glow discharge (GD) ion source coupled to a time-of-flight mass spectrometer (TOFMS). The GD is operated in the microsecond pulse (μs-pulse) mode. The operational parameters of the μs-pulse GD were optimized against the ion signals, giving 180 Pa for the discharge pressure, 3 A for the transient discharge current, 1.75 kHz for the discharge frequency and 2 μs for the discharge pulse duration. Experimental results show that the discharge current in the μs-pulse mode can be one order of magnitude higher than that obtained in the d.c. mode. The structure of the interface between the μs-pulse GD and the mass spectrometer was found to be critical, and a Macor disc must be applied in front of the sampling orifice in order to shield the sampling plate from the anode of the GD to achieve both a good vacuum and the best sputtering. A transient sputtering rate of 24.4 μs s −1 mm −2 was reached in the μs-pulse mode and was significantly higher than that for the d.c.-GD. Typical mass spectra of brass and nickel samples were studied and are discussed. © 1997 Elsevier Science B.V.

Theoretical and experimental studies of the broadening of dilute delta‐doped Si spikes in GaAs during SIMS depth profiling

AbstractDelta‐doped structures represent a powerful class of test structures to investigate the experimental and fundamental factors limiting the depth resolution obtainable in SIMS sputter depth profiling.In this work, theoretical studies of the effects on the broadening of an Si delta spike in GaAs as a function of the energy (1.4–4.4 keV) and angle of incidence (2°, 45° and 60° off‐normal) of the O2 sputter probe beam have been compared with recent experimental data. The theoretical calculations were carried out using the newly developed IMPETUS computer code, which simulates the depth profiling process by taking into account the combined effects of ballistic mixing (treating collisional mixing as a diffusion process), projectile incorporation into the matrix and sputtering. All of these are processes that always occur in any practical sputter depth profiling situation.The IMPETUS model can reproduce low‐energy Si depth profiles with great accuracy by using the well‐established TRIM calculated range, energy deposition and sputtering data and by making reasonable assumptions for the threshold energy for diffusion in addition to assuming a beam‐ and sputter statistics‐induced surface microtopography, which is described by a Gaussian area versus height distribution having a standard deviation σ = 0.8 nm. Significantly, it is shown that the effects of these parameters on the shape of the sputter profile are largely independent, with σ (accounting for microroughness) mainly affecting the leading edge and the threshold energy (determining mixing processes) the trailing edge of the sputter profile. Good agreement on the energy dependence of the broadening is also obtained. The expected improvement in depth resolution with increasing off‐normal bombardment angle is confirmed and can be quantified. The error in the experimental depth scale calibration based on a constant sputter rate, ignoring transient sputtering, is evaluated. Finally, the sputter depth profile observed for an Si delta spike in GaAs subjected to thermal annealing during growth by molecular beam epitaxy (MBE) can be reproduced accurately by considering diffusion broadening of the initial spike followed by a sputter profiling simulation.

Sputtering of metal alloys containing second-phase precipitates

Glow-discharge optical spectroscopy has been used as a real time monitor of the transient sputtering characteristics of alloy samples containing various amounts of second-phase precipitates. All alloys investigated reached a steady-state sputtering condition in which the chemical composition of the sputtered target material was equal to that of the bulk target. The transient sputtering time required to reach steady state was directly related to the time required to attain a steady-state rate of target surface cone formation and annihilation. Scanning Auger and electron microprobe analyses were used to show that cone formation in these alloys was due to masking of the matrix by slow sputtering second-phase precipitates and surface agglomerates. At steady state, the cone number density was found to be a function of bulk-alloy precipitate density, Ar sputtering pressure, and the sample sputtering rate. Increasing the Ar pressure resulted in increased scattering of sputtered atoms back to the target surface, while changes in overall target sputtering rate affected the surface migration and, hence, surface distribution of low-yield species. The average cone base diameter for Inconel 718, Incoloy 800, and Monel K-500 alloys was approximately 8 μm. The average depths removed during transient dc sputtering at an Ar pressure of 40 mTorr (5.3 Pa) and 2 kV were 1.5 μm for Inconel 718, 1.3 μm for Incoloy 800, and 0.1 μm for Monel K-500. For comparison, high-purity GaAs single-crystal samples sputtered under the same conditions reached steady-state sputtering after the removal of less than 100 Å. In all cases, films deposited on glass substrates at ambient temperature during steady-state sputtering had the same chemical composition as the bulk target. Inconel 718 and Incoloy 800 films also exhibited second-phase precipitates similar to those in the bulk targets.