[en] The authors describe the principles of doping of semiconductors by ion implantation and discuss some of the experiments that have laid the basis for many practical applications. They describe the implantation equipment, the measures adopted to ensure a uniform and reproducible dose, the collision processes that play a part in implantation, the theoretical principles that make it possible to calculate the penetration depth, etc., and two of the methods that can be used to measure the doping profiles. The latter are high-energy ion scattering and secondary-ion mass spectrometry and their particular merits are discussed. (Auth./C.F.)

[en] Since the high-voltage section of the ion-implantation machine developed at Philips Research Laboratories is not insulated by gas under pressure in a tank, changing the ion source is relatively easy. The machine is therefore particularly suitable for research. The accelerating voltage is high, 800 kV, so that heavy elements can be implanted in heavy substrates. There are two target chambers, one designed for treating silicon slices for IC manufacture, the other designed for other samples, including metals. Implantations in metals can improve properties such as hardness or resistance to corrosion. The half-width of the beam cross-section and the location of the pattern of the beam scan on the sample are evaluated with the aid of a wire frame around the sample; the wire frame is earthed through a milliammeter. The implantation dose is measured with a Faraday cup. Five ion sources have been tested: the hollow-cathode source, the Penning source, the radial-extraction Penning source, the sputter source and the microwave source. Some examples are given to demonstrate the usefulness of the machine: implantation of buried oxygen layers in a silicon slice, implantation of aluminium to improve the corrosion resistance of copper, and the determination of a number of concentration profiles for phosphorus in silicon at different energies. 12 refs.; 15 figs

[en] Boron, aluminium and gallium ions were implanted at 600 keV in a crystalline Si-(001) target for tilt angles 11.6⁰, 33.7⁰, 63⁰, 73⁰ and 83⁰. The depth profiles of the implanted impurities were measured with secondary ion mass spectrometry. These experimental data are compared with simulated impurity distributions from the Monte Carlo code MARLOWE (version 12). Projected range, lateral and longitudinal spread are discussed. (orig.)

[en] In the semiconductor industry complementary metal oxide semiconductor technology is the main stream. The continuing trend towards reduction of the transistor gate length allows for more complex integrated circuits. This puts stringent demands on other transistor properties such as source and drain junction depth. Source and drain are formed using ion implantation. For transistors where source and drain are boron-doped very low implantation energies are needed to obtain shallow implantation profiles. We have characterized boron implants, concentrating on the 100 eV to 1 keV energy range. As-implanted and annealed implant profiles are presented together with an overview of electrical activation and sheet resistance showing that ion implantation is a viable technique for shallow source/drain formation. In this paper some of the mechanisms underlying the formation of implantation profiles are discussed. Using a deactivation technique, we have measured the room temperature silicon self-interstitial diffusivity, D_I. It was found to be at least 10^-7 cm² s^-1. This appears to be a new record experimental value, approaching theoretical values for the silicon di-interstitial. Room temperature migration and clustering behaviour of implanted boron has been investigated by performing ion implantation of the boron isotope ¹¹B into molecular beam epitaxy-grown in situ doped layers. We, for the first time, show that a fraction of the implanted boron migrates deep into the bulk of the Si with substitutional ¹⁰B acting as trap centers for migrating ¹¹B. (orig.)

[en] AISI 52100 steel was implanted at 100 keV with doses in the range 4x10¹⁷-3x10¹⁸ C⁺/cm². Tribological tests were carried out on an oscillating ball-on-disk tester. The friction coefficient of the steel implanted with a dose of 3x10¹⁸/cm² was reduced by a factor of 3 for far more than 6 h. The wear behaviour of the steel shows normal wear up to a depth of 0.15 μm, being the range of the implanted carbon ions. Thereafter, a very low wear rate is obtained for more than 6 h (1.5x10⁵ cycles). The ball shows also no wear during this period. The steel was also studied by Moessbauer spectroscopy for various doses. At the lowest dose, ε-carbide is formed. For the higher doses the amount of ε-carbide increases, and the composition approaches that of ε-Fe_2.2C. The ε-carbide was also found by glancing-angle X-ray diffraction. Fe and C profiles were obtained from the Rutherford backscattering spectra. It was found that besides the ε-carbide a large amount of pure carbon is present in the implanted layer. (orig.)

[en] The amount of stress-induced martensite and its distribution in depth in xenon implanted austenitic stainless steel poly- and single crystals have been measured by Rutherford backscattering and channeling analysis, depth selective conversion electron Moessbauer spectroscopy, cross-sectional transmission electron microscopy and x-ray diffraction analysis. In low nickel 17/7, 304 and 316 commercial stainless steels and in 17:13 single crystals the martensitic transformation starts at the surface and develops towards greater depth with increasing xenon fluence. The implanted layer is nearly completely transformed, and the interface between martensite and austenite is rather sharp and well defined. In high nickel 310 commercial stainless steel and 15:19 and 20:19 single crystals, on the other hand, only insignificant amounts of martensite are observed. (orig.)

[en] The electronics industry demands stamped parts with high performance. Therefore, punching tools like cutting punches with very high precision have to be used. In the case reported, the punches are mounted in a modular system and have to be resharpened or replaced after a certain number of strokes. To increase the lifetime of the punches made of Vasco Wear steel, implantations with carbon, nitrogen, boron and titanium, and co-implantation with titanium and carbon were performed at energies from 50 keV to 200 keV and 600 keV and 700 keV with different doses in the region of several times 10¹⁸ cm^-2, measured perpendicular to the ion beam. A maximum increase in lifetime of a factor of 3.6 was reached. The surface roughness had a large influence on the increase lifetime and the improvement caused by specific ion species. The maximum improvement was obtained for the lowest surface roughness (R_a=0.04 μm). Therefore, when performing the implants, punches with low surface roughness should be used. The most successful ion species were boron and nitrogen for the lowest surface roughness used (R_a=0.04 μm), and after changing the polishing procedure (R_a=0.14 μm) titanium and nitrogen at medium energies (100-200 keV). High energy implantation (700 keV) resulted in an increase of a factor of 2.1 at lower doses (5.6x10¹⁷ cm^-2), but is uneconomical owing to the low current density. In laboratory wear tests (ball on disk) no improvement by ion implantation could be found. These results prove that it is difficult to compare field tests and laboratory tests because of different testing conditions. (orig.)

[en] Reactive ion etching of Si wafers with a CF₄ plasma results in the formation of a very thin layer of SiF_x. Quantification of the fluorine areal density as a function of plasma parameters is important for models describing the etching process. The Eindhoven AVF cyclotron has been used to quantify the amount of deposited fluorine using the ¹⁹F(p,p'γ)¹⁹F reaction detecting the 110 and 197 keV γ-lines with a hyperpure Ge detector. The energy of the proton beam was chosen to be 2.76 MeV thereby limiting the Compton background by suppressing excitation of the fourth excited state (5/2⁺) of ²⁹Si. Quantification was obtained by the use of an implantation standard. A detection limit of 1.10¹⁴ F/cm² was obtained and two series of pilot experiments were carried out in which gas discharge pressure and plasma power density have been varied. (orig.)