Molecular dynamics and Monte-Carlo simulation of sputtering and mixing by ion irradiation

Molecular dynamics (MD) and Monte-Carlo (MC) simulations of low-energy (<500 eV) Ar ion irradiation on Si substrates were performed in order to investigate the mixing and sputtering effects. Both MD and MC simulation show similar results in sputtering yield, depth profile of projectile and mixing of substrate. For these incident energies, the depth of the mixed region is determined by the implant range of incident ions. For example, when the incident energy is 500 eV, the Ar ions reach a depth of 40 Å so that the Si atoms that reside shallower than 40 Å are fully mixed at an ion dose of about 5.0×10¹⁶ atoms/cm². The resolution of secondary ion mass spectrometry (SIMS) was also studied. It was found that the resolution of SIMS depends on the depth of mixing, which depends in turn on the implant range of the probe ions. This is because the mixing of substrate atoms occurs more frequently than sputtering, so that the information about the depth profile in the mixing region is disturbed.     

Medium-energy ion scattering spectroscopy with a position-sensitive and time-resolving detector

A Si (1 0 0) sample covered with a thin Ho layer was measured with a three-dimensional medium-energy ion scattering spectrometer. The spectrometer is an extended version of a time-of-flight spectrometer for medium-energy ion scattering, equipped with a large position-sensitive detector. The device is used for composition depth profiling and crystal structure determination. The intensity distribution of detected particles was visualized to present medium-energy scattering phenomena. Circular shapes were observed in images created with part of the data containing particles scattered from the surfaces of the sample layers. Images show leading edges of “clouds” of arriving scattered particles were detected using the flat surface of the detector, integrated over 2 ns intervals. The center positions of shapes produced by particles scattered on Ho and Si atoms are different. This is explained by the different kinematic-factor dependences on scattering angle of particles scattered on Si and Ho atoms. The depth resolution of the time-of-flight spectra acquired with the full solid acceptance angle of the detector is limited by the kinematic spread. Using position information of detected particles from the detector, corrections for the kinematic spread and variations of flight path lengths were applied to spectra, increasing the depth resolution.     

Depth resolution of TOF-ERDA using a He beam

Depth resolution of time-of-flight ERDA using a ⁴He beam (He TOF-ERDA) has been studied. The measurement system consists of a time detector of the ion transmission type and a silicon surface-barrier detector. Depth resolution was measured using samples of carbon layers on silicon wafers and ⁴He beams with energies between 3.5 and 10.1 MeV. The depth resolution of 6.0 ± 1.6 nm (FWHM) was obtained with a 3.5 MeV ⁴He incident beam. The measured depth resolution agreed with that evaluated by a calculation. Comparison with other methods such as heavy ion (HI) TOF-ERDA, resonant elastic scattering and nuclear reaction analysis (NRA) was performed. Depth resolution obtained by He TOF-ERDA is superior to that by NRA or resonant elastic scattering, and comparable to that by HI TOF-ERDA.     

On the extraction of neutralisation information from low energy ion scattering spectra

There are two different ways of measuring low energy ion scattering (LEIS), i.e. detection of ions and neutrals by means of time of flight (TOF-LEIS), or detection of ions only by means of an electrostatic analyser (ESA-LEIS). We discuss, how information on charge exchange can be extracted from ESA-LEIS and TOF-LEIS spectra, respectively and which is the level of accuracy that can be expected from these procedures.     

Quantitative depth profiling of deuterium up to very large depths

Quantitative depth profiles of deuterium up to very large depths are achieved from the energy spectra of protons created by the D(³He,p)α nuclear reaction at incident energies up to 6 MeV. The advantages of this method compared to the more often applied resonance method are discussed. For light target materials the achievable depth resolution is mainly limited by geometrical spread due to the finite size of the detector aperture, while for heavy materials the resolution is mainly limited by multiple small-angle scattering. A reasonable depth resolution throughout the whole analyzed depth can be obtained by using several different incident energies. Depth profiling up to 38 μm is demonstrated for a-C:D layers deposited on the limiter of Tore Supra, and up to 7.5 μm in tungsten coatings from the divertor of ASDEX Upgrade.     

Study of ion beam induced mixing during sputter depth profiling of thin films by LEIS

Ion beam induced mixing during sputter depth profiling was studied for tantalum and lead marker layers in silicon with 5 keV neon by low energy ion scattering spectroscopy (LEIS). The diffusion approximation was used to calculate mixing efficiency values (D/JF_d) from decay length measurements. The mixing efficiency values are shown to be sensitive to the preferential sputtering which takes place during ion bombardment. TRIM simulations for Ta/Si are shown to agree with the experimentally determined value for preferential sputtering. Depth profiling at high temperature is shown to separate some of the interrelated mixing mechanisms of radiation enhanced diffusion (RED) and radiation induced segregation (RIS). For the Ta/Si system the mixing efficiency value is observed to remain constant regardless of the 5 keV inert gas ion beam used for depth profiling.     

Development of enhanced depth-resolution technique for shallow dopant profiles

A rapid shrinkage in the minimum feature size of integrated circuits requires analysis of dopants in their shallow source–drain and their extensions with an enhanced depth resolution. Rutherford backscattering spectroscopy (RBS) combining a medium-energy He ion beam with a detector of improved energy resolution should meet the requirement of a depth resolution better than 5 nm at a depth of 10–20 nm in the next 10 years. A toroidal electrostatic analyzer of 4×10⁻³ energy resolution has been used to detect the scattered ions of a medium-energy He ion beam. Five keV As⁺ implanted Si or SiO₂ samples were measured. Depth profiling results using the above technique are compared with those of glancing-angle RBS by MeV energy He ions. Limitations in the energy resolution due to various energy-spread contributions have been clarified.     

Rutherford backscattering (RBS) with lithium ions

⁷Li ions at 4.5 MeV and ⁴He ions at 1.5 MeV were scattered at 170° from the same target under the same conditions. The distribution of elements as a function of distance from the surface is more clearly revealed with ⁷Li than ⁴He. We have rewritten our RBS analysis program to handle all ion beams and to take into account the normal isotopic abundance of the target elements. With 4.5 MeV ⁷Li ions the minority isotopes can not be ignored, even with silicon. At present, the accuracy for determining elemental depth profiles with ⁷Li is limited by the accuracy of Li ion stopping powers.     

On the origin of the LEIS signal in TOF- and in ESA-LEIS

The ion fraction analysis of ⁴He⁺ ions backscattered from various faces of copper single crystals is performed by using time-of-flight (TOF) and electrostatic analyzer (ESA) low-energy ion scattering (LEIS) techniques. When an experiment that integrates over 2π azimuth (typical ESA-LEIS setup) is used, the yield of ions backscattered from the Cu(1 1 0) surface may be given by projectiles penetrated much deeper than just one or two monolayers. The threshold energy for reionization processes for ⁴He⁺ and Cu found earlier by TOF-LEIS is experimentally confirmed by ESA-LEIS.     

Measurement of deuterium and helium by glow-discharge optical emission spectroscopy for plasma–surface interaction studies

To examine the resolution of isotope analysis of hydrogen with glow-discharge optical emission spectroscopy (GDOES), depth profiles of hydrogen and deuterium in a H-containing Ta/D-containing Ti/Ni layered structure were measured. The depth profiles of deuterium could be measured with sufficient resolution in the presence of relatively large amounts of hydrogen and vice versa. In addition, measurements of depth profiles of He implanted in W at room temperature were also performed with Ne plasma. The intensity of the He emissions was sufficiently high at a fluence of 10²⁰ He m^?2 or higher. The depth profiles of He measured in this manner were in good agreement with the results of cross-sectional observations using a transmission electron microscope. Therefore, it was concluded that GDOES with Ne plasma is a promising technique for the depth profile analysis of plasma-facing materials and deposited layers formed on them.     

Depth resolution and recoil cross section for analyzing hydrogen in solids using elastic recoil detection with 4He beam

Two aspects of the elastic recoil detection technique for analyzing H and D are described; i) experimental factors which effectively limit the depth resolution in Al film, and ii) determination of the recoil cross section for H(⁴He, ⁴He)H and D(⁴He, ⁴He)D reactions in the range of 1.5–3.0 MeV energy of ⁴He. Both experimental and theoretical estimates of the depth resolution are presented and are in good agreement each other. The theoretical estimate therefore provides a reliable guide to find optimum resolution conditions. The recoil cross section for H is more than double the theoretical Rutherford scattering value and that for D becomes greater than 30 times Rutherford near the resonance energy of 2.1 MeV ⁴He.     

Neutralization effect for Ne+ scattering from a Cu(110) surface

Intensity and structure of the energy spectra of Na⁺ and Ne⁺ ions scattered from a Cu(110) surface are governed by multiple scattering and neutralization effects. These were studied for ion energies between 600 eV and 1 keV and in the temperature range from 100 to 600 K in the experiment and by computer simulation. Na⁺ scattering directly reflects the crystallographic structure of the (110) surface. The temperature effects can be used to analyze thermal motions of surface atoms in terms of a surface Debye temperature for specific vibrational directions. The contributions of single and multiple scattering events to the energy spectra are analyzed and for Ne⁺ a strong trajectory-dependent neutralization is found. The comparison of the neutralization of Ne⁺ and Na⁺ leads to a Ne⁺ ion survival probability of a few percent for single scattering, less than 1% for double scattering, and a value of less than 10⁻³ for scattering from atoms below the top atomic layer. A simple neutralization model is developed to explain the observed survival probabilities.     

Neutralization of He ions scattered from Ca surface

The neutralization of He ions scattered from a Ca surface, as well as Al and Cu surfaces as references, is discussed on the basis of the incident energy dependence of low energy ion scattering (LEIS) spectra. The neutralization probability obtained for Ca is much higher than that expected on the basis of the Auger mechanism, indicating that the collision-induced (CI) process is dominant for the neutralization. The markedly large background component in the LEIS spectrum for Ca, corresponding to the re-ionization at the surface of He ions scattered and neutralized in the bulk, is consistent with the high probability of a CI process.     

ION BEAM AND SIMS ANALYSIS ON DAMAGE OF GaAs DOPED WITH N^＋

ＩＯＮ　ＢＥＡＭ　ＡＮＤ　ＳＩＭＳ　ＡＮＡＬＹＳＩＳ　ＯＮ　ＤＡＭＡＧＥ　ＯＦ　ＧａＡｓ　ＤＯＰＥＤ　ＷＩＴＨ　Ｎ~＋ＷｕＸｉａｏｓｈａｎ；ＬｉｎＺｈｅｎｊｉｎ；ＪｉＣｈｅｎｇｚｈｏｕ；ａｎｄＹａｎｇＸｉｚｈｅｎ（ＤｅｐａｒｔｍｅｎｔｏｆＰｈｙｓｉｃ?..     

Time-of-flight spectrometry for materials analysis

A time-of-flight spectrometer for investigation of solids has been built. This type of spectrometer can be used for forward- and backward-scattering experiments. The scattered or recoiled ions are discriminated by a pair of C-foils. The emitted secondary electrons are detected by channel plates and are used as start and stop signals. Depth profiles up to 500 nm were measured with a depth resolution of 3 nm. The best mass resolution achieved under optimal conditions is 1 u. For example, the two Ga isotopes from GaAs samples have been separated. He, N, Ne and Ar ions with energies up to 10 MeV have been used as projectiles.     

Considerations about multiple and plural scattering in heavy-ion low-energy ERDA

Low-energy heavy-ion Elastic Recoil Detection Analysis (ERDA) is becoming a mature technique for high-resolution characterization of thin films, i.e. below 50 nm thickness. In combination with a small tandem accelerator (∼2 MV terminal voltage) and beam energies below 20 MeV, it is suitable for routine analysis of key materials in semiconductor technology.At low-energies, however, small angle multiple scattering and large angle plural scattering of ions play a significant role, starting from the first nanometers. Multiple and plural scattering dominate the depth resolution deterioration with increasing depth and, when glancing angles are used, introduce long tails in the elemental energy profiles. Moreover, multiple and plural scattering may affect the elemental relative and absolute quantification. A complete characterization of ultra-thin films thus requires a detailed analysis with accurate simulation of the energy spectra.In this paper we investigate the mechanism of multiple and plural scattering for different combinations of beam/recoil atoms, energies and geometries. Simulations run with the Monte Carlo code MCERD support and generalize the experimental data. The calculations show the relative contributions of beam and recoil ions and highlight the role of ion angular distribution to the formation of tails in the energy profiles.     

A broad beam ion source used for planar sputter-etching of shallow layers from flat samples and determination of their mass density

A special UHV device for sputter-etching of large planar samples is presented, which is based on a broad beam ion source of the Kaufman type. It is designed for a beam of Ar⁺ ions with an energy of only 600 eV and a heat load per area of about 0.17 W cm⁻². The equipment is suitable for a planar removal of nanometer thin layers from the surface of flat samples especially of wafers. In a circular area of 2 cm diameter a homogeneous sputtering was reached with ±5% in depth. By means of an atomic force microscope the surface roughness of highly polished wafers was shown to keep on the sub-nanometer level. The sputtered depth was determined by a Tolansky microscope, the sputtered mass was measured by differential weighing on a calibrated microbalance. Depending on the sputtered target (Si, Ge, GaAs) the sputtering yield was in the range of 0.9–2.0 atoms per ion and was very similar for modified materials like oxidised or Co- and As-implanted samples of Si (±10%). For validation of the method, the mass density of thin layers consisting of pure crystalline Si, GaAs, Ge and of As-implanted Si was determined. Precision and accuracy were shown to be <3% and <6%, respectively. The method was also used successfully for depth profiling of a Si and a GaAs wafer sample by repeated sputter-etching and differential weighing. Density/depth profiles were recorded with 35 and 50 nm steps but the actual depth resolution can be reduced to 5–10 nm and possibly to only 2–3 nm.     

Optimal geometry for GeSi/Si super-lattice structure RBS investigation

He Rutherford Backscattering Spectrometry (RBS) depth resolution is limited by detector energy resolution, He ion energy loss in the sample material, energy straggling and geometry of the experiment. Examples of experimental results are shown for GeSi/Si super-lattice investigation for random and channeling RBS analysis in different geometry of detection. Different detection geometries are discussed and compared. For the case of axial channeling experiments, the incident beam direction is restricted to typically low index axes only. It is shown how to improve the depth resolution and sensitivity in this type of experiments. Effects of sample material as well as contributions from different physical processes including multiple scattering effects are discussed in some detail.     

Depth profiling of hydrogen by detection of recoiled protons

The elastic recoil detection technique (ERD) using a 2.5 MeV ⁴He beam for depth profiling of hydrogen in the near-surface regions of solids is described. The optimization of the experimental conditions such as scattering geometry and analyzing beam energy is discussed. The factors limiting the depth resolution of the method have been evaluated showing that a depth resolution of the order of 20 nm can be obtained. Also presented are typical applications for hydrogen profiling in a silicon matrix.     

On the understanding of positive and negative ionization processes during ToF-SIMS depth profiling by co-sputtering with cesium and xenon

In this paper, ionization processes of secondary ions during ToF-SIMS dual beam depth profiling were studied by co-sputtering with 500 eV cesium and xenon ions and analyzing with 25 keV Ga⁺ ions. The Cs/Xe technique consists in diluting the cesium sputtering/etching beam with xenon ions to control the cesium surface concentration during ToF-SIMS dual beam depth profiling. Several depth profiles of a H-terminated silicon wafer were performed with varying Cs beam concentration and the steady state Si, Xe and Cs surface concentrations were measured in situ by Auger electron spectroscopy. It was found that the implanted Cs surface concentration increases with the Cs fraction in the beam from 0% for the pure Xe beam to a maximum Cs surface concentration for the pure Cs beam. Secondly, the variation of the silicon work function, due to the Cs implantation, was measured in situ and during depth profiling as the shift of the secondary ion kinetic energy distributions. Finally, the positive and negative elemental ion yields generated by the Ga analysis beam were recorded and modeled with respect to varying Cs/Xe mixture. We found that the Si⁻ and the Cs⁻ yields increase exponentially with the decrease of the silicon’s work function while that of Cs⁺ and Si⁺ decrease exponentially, as expected by the electron tunneling model.