Sputtering Of Silicon Research Articles

Precise Depth Control of Silicon Etching Using Chlorine Atomic Layer Etching

In this study, the atomic layer etching (ALE) of Si was carried out using Cl2 adsorption followed by Ar+ ion beam irradiation with a low energy Ar+ ion beam generated by an inductively coupled plasma ion gun. A saturated silicon etch rate due to chlorine ALE could be obtained when the Ar+ ion acceleration voltage of the ion gun was in the range of 70 to 90 V, as a result of the preferential etching of silicon chloride formed during the chlorine adsorption period by the Ar+ ions while the silicon sputter etch rate remains insignificant. This was attributed to the differences in the silicon-to-silicon and silicon-to-silicon chloride binding energies. The saturated silicon etch rate by ALE was dependent on the chlorine flow rate, i.e. the surface coverage of chlorine and the Ar+ ion irradiation time. In this experiment, a silicon etch rate of 1.36 Å/cycle, which is a (100) silicon monolayer per cycle, could be obtained by flowing more than 10 sccm chlorine gas followed by bombarding the surface by Ar+ ions with an acceleration voltage of 70 V for more than 40 seconds. Under this condition, when a 30 nm scale silicon etch profile was examined after 200 cycles, a silicon etch profile with no undercut could be obtained.

Raman analysis of Si–C–N films grown by reactive magnetron sputtering

Silicon carbon nitride thin films have been deposited by reactive magnetron sputtering of silicon and graphite targets in mixed Ar/N 2 atmosphere at substrate temperature of 300 °C. The substrate bias voltage varied from −50 up to +50 V and the nitrogen flow rate varied from 0 to 20 sccm. The as-deposited films were analyzed by Raman spectroscopy (RS) and X-ray photoelectron spectroscopy (XPS). The Raman analyses show that the film without nitrogen incorporation has mixed sp 2–sp 3-hybridized carbon structures while those with nitrogen introduction give rise to nitrogen-bound sp 1-, sp 2- and sp 3-coordinated carbon structures as well as Si–N phase. The change of the D band position (∼1360 cm −1), FWHM and its relative intensity with respect to the G band (∼1595 cm −1), I D/ I G, seem to be correlated with the formation of these phases and therefore to the deposition conditions. XPS analyses not only confirm the bonding natures revealed by Raman spectroscopy, but also give quantitatively the relative importance of the phases. It was shown that the area ratio of the nitrogen-bound sp 3- to sp 2-coordinated carbon bonds is: 1.41:1.38:1.8:3.19 and that of Si N bonds to (Si N+Si C) bonds is 0.4:0.5:0.9:1 for the Si–C–N films prepared with 5, 10, 15 and 20 sccm nitrogen flow rate, respectively.

Quantitative study of oxygen enhancement of sputtered ion yields. I. Argon ion bombardment of a silicon surface with O 2 flood

A novel quantification approach is applied to determine in situ the amount of surface oxygen within the sputtered particle escape depth during steady-state sputter depth profiling of silicon under simultaneous oxygenation with an oxygen flood gas or with an oxygen primary ion beam. Quantification is achieved by comparing the secondary ion intensities of 16O that is adsorbed or implanted at the Si surface with the measured peak intensities of a calibrated 18O ion implant used as a reference standard. Sputtered ion yields can thereby be related to surface oxygen levels. In the present work the dependences of the partial silicon sputter yield Y and of the positive and negative secondary ion useful yields UY(X ±) (X = B, O, Al, Si, P) on the oxygen/silicon ratio, O/Si, in the sputtered flux are studied for 40Ar + bombardment of Si with simultaneous O 2 flooding. The silicon sputter yield is found to decrease with increasing flood pressure and O/Si ratio by up to a factor of 3. Both positive and negative secondary ion yields are enhanced by the presence of oxygen at the silicon surface. The useful ion yield of Si + scales non-linearly with the atom fraction of surface oxygen; this behavior is shown to invalidate models that suggest that Si + ion yield enhancement is dominated either by isolated oxygen atoms or by formation of SiO 2 precipitates. In contrast a microscopic statistical model that assumes that local Si + ion formation depends only on the number of oxygen atoms coordinated to the Si atom to be ejected fits the ion yield data quantitatively.

Photosensitive structures based on single-crystal silicon and phthalocyanine CuPc: Fabrication and properties

Vacuum thermal deposition of phthalocyanine CuPc onto the surface of crystalline silicon and subsequent magnetron sputtering of ZnO:Al are used to form n-ZnO:Al-p-CuPc-n-Si photosensitive structures for the first time. The highest photosensitivity of these structures S ≈20 V/W is attained if the ZnO side of the structure is illuminated and is observed in the photon-energy range 1–3.2 eV at T=300 K. An induced photopleochroism is observed if the linearly polarized light is incident obliquely on the ZnO side; the magnitude of the photopleochroism oscillates as a result of the interference of linearly polarized light in the ZnO film. It is concluded that the suggested structures have prospects for use in broadband photoconverters of natural light and in rapidly tunable photoanalyzers of linearly polarized light.

Effect of sputtering input powers on CoSi2 thin films prepared by magnetron sputtering

CoSi2 thin films were prepared by radio frequency magnetron sputtering using CoSi2 alloy target. Effect of sputtering input powers on characteristics of CoSi2 thin films was researched by X-ray diffraction (XRD), transparent electron microscope (TEM), energy dispersive X-ray analysis and four points probe, etc. It was shown that the deposition rate increased lineally, the selective sputtering of silicon was strengthened, the (1 1 1) texture increased, and the resistivity decreased when the input powers were increased.

Stoichiometry dependence of hardness, elastic properties, and oxidation resistance in TiN/SiNx nanocomposites deposited by a hybrid process

TiN/SiN x nanocomposite layers with Si contents between 0 and 25 at. % were deposited by a reactive arc-magnetron sputtering hybrid process. The stoichiometry of the SiNx phase was found to be related to the silicon sputter target state, i.e., elemental or nitrided. TiN/SiNx layers with a Si:N ratio close to 0.75 (silicon nitride) show a hardness maximum at overall Si contents between 5 and 7 at. %. The hardness maximum is absent for nitrogen deficient stoichiometries of SiNx. The oxidation resistance of the composite layers is three to five times better than that of pure TiN. In contrast to the effect of the stoichiometry on hardness, the oxidation resistance depends on the overall silicon content only, regardless of stoichiometry.

Molecular dynamics simulation of silicon sputtering: sensitivity to the choice of potential

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. Detailed information is extracted and briefly discussed. It is found that these potentials lead to significant differences in sputter behavior. By testing these potentials (and two others) against newly calculated DFT data for Si energies in a wide variety of atomic environments, we demonstrate that they all have severe shortcomings. Therefore, the Modified Embedded Atom Method potential was re-fitted to the DFT data, leading to a new parametrization and a greatly improved fit. However, a spurious ‘amorphous phase’ energy minimum was detected during molecular dynamics simulations at room temperature. Roads to dealing with this problem are discussed.

Ultra-shallow arsenic implant depth profiling using low-energy nitrogen beams

Sputtering of silicon by low-energy nitrogen primary ion beams has been studied by a number of authors to characterize the altered layer, ripple formation and the sputtered yields of secondary ions [Surf. Sci. 424 (1999) 299; Appl. Phys. A: Mater. Sci. Process 53 (1991) 179; Appl. Phys. Lett. 73 (1998) 1287]. This study examines the application of low-energy nitrogen primary ion beams for the possible depth profiling of ultra-shallow arsenic implants into silicon. The emphasis of this work is on the matrix silicon signals in the pre-equilibrium surface region that are used for dose calibration. Problems with these aspects of the concentration depth profiling can give significant inconsistencies well outside the error limits of the quoted dose for the arsenic implantation as independently verified by CV profiling. This occurs during depth profiling with either oxygen primary ion beams (with and without oxygen leaks) or cesium primary ion beams.

Effect of silicon addition on surface morphology and structural properties of titanium nitride films grown by reactive unbalanced direct current-magnetron sputtering

Thin films of Ti1–x–y Six Ny were produced on unheated Si(100) substrates by reactive unbalanced dc-magnetron sputtering of titanium and silicon in an Ar–N2 gas mixture. The effects of silicon incorporation on surface morphology and structural properties of these films as well as the influence of postdeposition annealing have been studied. These films were characterized ex situ in terms of their core-level electron bonding configuration by x-ray photoelectron spectroscopy, their microstructure by cross-sectional transmission electron microscopy and x-ray diffraction, their hardness by nanoindentation measurements, and their roughening kinetics by atomic force microscopy (AFM) with the scaling analysis. It was found that a linear increase in the Si concentration of the films was observed with increasing Si target current up to 2 A whereas the reverse trend was seen for the Ti concentration. The films consisted of 15–20-nm-sized TiN crystallites embedded in an amorphous SiNx matrix. They had a hardness of about 32.8 GPa with silicon concentration x = 0.1. The improved mechanical properties of Ti1–x–y Six Ny films with the addition of Si into TiN were attributed to their densified microstructure with development of fine grain size and reduced surface roughness. The reduction in grain size has been supported by means of a Monte Carlo simulation that reveals that the average size of TiN grains decreases with the volume fraction of amorphous SiNx approximately according to a power law, showing a reasonable agreement with the experimental results. By applying the height–height correlation functions to the measured AFM images, a steady growth roughness exponent α = 0.89 ± 0.05 was determined for all the films with different Si additions. It was also found that the nanocomposite films were thermodynamically stable up to 800 °C. The effect of thin SiNx layer in stabilizing nanocrystalline TiN structure is also elucidated and explained on the basis of structural and thermodynamic stability.

Mechanical properties of hydrogen-free a-C:Si films

Hydrogen-free a-C:Si films with Si concentration between 3 and 50 at.% were prepared by magnetron sputtering of pure graphite and silicon. In contrast to a-C the a-C:Si films do not delaminate even from steel substrates and remain stable after vacuum annealing up to 780 °C. Mechanical properties (microhardness, intrinsic stress) and film structure (electron diffraction (ED), Raman spectra) were investigated in dependence on discharge power, Ar pressure and substrate bias. For lower silicon concentrations (below ∼15 at.%) both microhardness and intrinsic stress of films on biased substrates exceeds the value observed at grounded bias. In the range 38–43 at.% of Si the microhardness of a-C:Si films almost does not depend on the substrate bias and reaches values of 45–55 GPa corresponding to the hardest a-C films prepared by magnetron sputtering under the same conditions. The intrinsic stress is lower than that in a-C films. ED indicates a disappearing graphite-like layer structure and gradual transition to structure formed by the SiC (sp 3) bonds as the silicon concentration increases. These structural changes are reflected in Raman spectroscopy.

Nanocomposite Ti–Si–N films deposited by reactive unbalanced magnetron sputtering at room temperature

Thin films of Ti–Si–N were produced on unheated Si(1 0 0) substrates by reactive unbalanced magnetron sputtering of titanium and silicon in an Ar–N 2 gas mixture. The effects of Si incorporation ranging from 0 to 18 at.% on the structural and mechanical properties of these films have been studied. These films were characterized ex situ in terms of their atomic concentrations and core-electron bonding configuration by Auger electron spectroscopy (AES) and X-ray photoelectron spectroscopy (XPS), their roughening kinetics by atomic force microscopy (AFM), their microstructure by X-ray diffraction (XRD), and cross-sectional scanning electron microscopy (SEM), and their hardness and elastic modulus by nanoindentation measurements. It was found that the Si bonding status was in the form of Si 3N 4 with low Si contents (≤14 at.%), while it was in the form of both Si and Si 3N 4 at high Si contents (≥18 at.%). The quantification of surface roughening was achieved by calculation of both vertical root mean square (rms) roughness and lateral correlation lengths of the film surface using the height–height correction functions of measured AFM images. For all the Ti–Si–N films, a steady roughness exponent α=0.88±0.04 was determined. It was also found that the improved mechanical properties of Ti–Si–N films with the addition of Si into titanium nitride (TiN) compound were attributed to their densified microstructure with development of fine grain size and reduced surface roughness. The effect of silicon in stabilizing the nanocrystalline TiN/amorphous Si 3N 4 structure is also discussed on the basis of structural and thermodynamical stability.

Target poisoning during reactive magnetron sputtering: Part I: the influence of ion implantation

Gettering plays a minor role during reactive sputtering of silicon in a nitrogen/argon mixture. However, an abrupt increase of the target voltage as a function of the nitrogen mole fraction is noticed which is not expected from the classical models explaining reactive magnetron sputtering. To explain the target voltage behaviour during DC magnetron sputtering of silicon in an argon/nitrogen mixture, a model is proposed which is based on the reactive ion implantation into the subsurface region of the silicon target. The model calculates the concentration of the nitrogen ions implanted into the target and assumes three possible pathways for these implanted ions. A first pathway is the chemical reaction between the implanted nitrogen ions and the target material to form silicon nitride. The implanted nitrogen can also remain in the target as non-reacted nitrogen atoms. Or, the nitrogen atoms can recombine in the target and diffuse from the target. The compound formation results in a decrease of the target surface recession speed or target erosion rate. As the surface concentration of the implanted ions is inversely proportional to the surface recession speed, an avalanche situation becomes possible. This abrupt transition in recession speed is accompanied with a sudden increase of the concentration of non-reacted nitrogen atoms in the target. In this way, the abrupt target voltage change, noticed at a given mole fraction of nitrogen in the plasma, can be understood.

Characterization of amorphous GexSi1−xOy for micromachined uncooled bolometer applications

Thin films of GexSi1−xOy were prepared by reactive magnetron sputtering using simultaneous sputtering of silicon and germanium targets in an environment of oxygen and argon. Silicon and oxygen content were varied from 0 to 30 at. % separately and the effect of the addition of each element on electrical and optical properties of amorphous germanium was studied. The electrical and optical behavior of the compound with varying elemental composition is explained based on the oxidation behavior of the Si and Ge. Increasing the silicon content was found to inhibit the formation of germanium–oxygen bonds. Values of temperature coefficient of resistance as high as −5% K−1 were obtained at moderate resistivity values around 3.8×104 Ω cm. These characteristics could be used to enhance the performance of micromachined uncooled bolometers. The composition control enabled by cosputtering components allows resistivity and activation energy to be tailored to suit different design specifications.

A Study of Environmentally Friendly Titanium Pretreatments for Adhesive Bonding to a Thermoplastic Composite

AbstractThe increased use of fibre reinforced polymer composites for aircraft structures have highlighted the need for reliable methods of bonding these materials to metallic components such as titanium. There is also a need for a simple test to evaluate adhesive bonds with dissimilar substrates, while under load and in a variety of environments. In this study the Boeing wedge test has been adapted for use with metal to composite adhesive bonds. A flexural rigidity matching approach was used with the addition of a backing beam to the composite coupon. The titanium adherends were surface pre‐treated using novel environmental benign methods, namely argon gas plasma and silicon sputtering. Sodium hydroxide anodisation, which is currently used industrially and alumina grit blasting were used as benchmarks. The composite, glass fire reinforced polyphenylene sulphide was alumina grit blasted. An environmental cycle was used to mimic a typical aircraft service cycle; joints were subjected to 24‐hour cycles of ambient, 60 °C water immersion, 60 °C dry and –55 °C dry for 12 days. Crack propagation and locus of failure were monitored after each cycle.

The network structure of amorphous silicon–carbon alloy

The network structure of amorphous silicon–carbon alloy (a-Si 1− x C x ) has been studied over a wide range of x. The a-Si 1− x C x thin films were prepared by sputtering silicon and carbon target with argon in radio-frequency magnetron sputtering equipment. The films were characterized by X-ray photoelectron spectroscopy, optical absorption, infrared absorption, and mechanical measurements. The results showed that the network structure could be classified neither as the random covalent network nor as the chemically ordered covalent network. The structure as a whole was close to the random covalent network, but the Si–Si combination at x>0.5 showed a feature of the chemically ordered covalent network. The film at 0.6< x<0.8 was hard and showed a high energy gap, due to the sp 3 configuration in Si–C combinations.

Molecular dynamics studies of the sputtering of divertor materials

We use molecular dynamics (MD) simulations to study two important erosion processes of divertor materials, namely the chemical sputtering of carbon by hyperthermal deuterium and the self-sputtering of W by redeposited atoms. Although it is generally accepted that MD simulations can provide a good qualitative description of far-from-equilibrium processes, such as irradiation-induced cascades, we show that also good quantitative agreement with experiments is obtained. We further show that the chemical sputtering yield of carbon is reduced by Si doping at impact energies ⩽20 eV. Significant carbon sputtering is still observed for both undoped and doped structures at 5 eV. Silicon sputtering is negligible at all impact energies studied. For W self-sputtering, sputtering yields higher than one are observed at ⩾1 keV for normal incidence and at ⩾500 eV for 20° off-normal incidence.

One and Two-Dimensional Pattern Formation on Ion Sputtered Silicon

ABSTRACTThe evolution of surface morphology during ion beam erosion of Si(111) at glancing ion incidence (60° from normal, 500 eV Ar+, 0.75 mA/cm2 collimated beam current) was studied over a temperature range of 500–730° Celsius. Keeping ion flux, incident angle, and energy fixed, it was found that one-dimensional sputter ripples with wavevector oriented perpendicular to the projected ion beam direction form during sputtering at the lower end of the temperature range. For temperatures above approximately 690° Celsius, growth modes both parallel and perpendicular to the projected ion beam direction contribute to the surface morphological evolution. This effect leads to the formation of bumps (“dots”) with nearly rectangular symmetry.

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.

Effect of the projectile parameters on the charge state formation process in solid sputtering

An effect of atomic and molecular projectile parameters on the charge state formation process in silicon sputtering has been studied. It was found that the electronic subsystem excitation in a subsurface region of silicon depends on projectile parameters and it affects the ionization probability P + of sputtered atoms.

Influence of nitrogen–argon gas mixtures on reactive magnetron sputtering of hard Si–C–N films

Si–C–N films were deposited on p-type Si(100) substrates by dc magnetron co-sputtering of silicon and carbon using a single sputter target with variable Si/C area ratios in nitrogen–argon mixtures. The film characteristics were primarily controlled by the argon concentration (0–75%) in the gas mixture at a fixed 40% silicon fraction in the magnetron erosion track area. The total pressure and the discharge current on the magnetron target were held constant at P=0.5 Pa and I m=1 A, the substrate temperature was adjusted at T s=600 °C by an ohmic heater and the rf-induced negative substrate bias voltage, U b was −500 V. Mass spectroscopy was used to explain differences between sputtering of carbon and silicon in nitrogen–argon discharges. The films, typically 1.0–1.5 μm thick, were found to be amorphous with a very smooth surface ( R a≤0.8 nm). It was shown that the nitrogen–argon gas mixture composition is an important process parameter in a reproducible production of Si–C–N compounds with controlled properties by dc magnetron co-sputtering using a composed C–Si target with variable Si/C area ratios. With a rising argon concentration in the gas mixture, the Si content in the films rapidly increases (from 19 to 34 at.% for a 40% Si fraction in the erosion target area), while the C content decreases (from 34 to 19 at.%) at an almost constant (39–43 at.%) N concentration. An intensified bombardment of growing films by argon leads to its subplantation into the films (up to 5 at.%) and to a decrease in a volume concentration of hydrogen (from 5 to 1 at.%). As a result, the NSi and SiN bonds dominate over the respective NC and SiO bonds, preferred in a pure nitrogen discharge, and the film hardness increases up to 40 GPa. The effect of the negative substrate bias voltage (from a floating potential of −18 to −500 V) on characteristics of the films prepared in a 50% N 2+50% Ar gas mixture is not too marked at T s=600 °C. A decrease in the T s values to 135–210 °C leads to a higher incorporation of hydrogen into the films (up to 6 at. %) and to a stronger influence of the negative substrate bias voltage.