J.P. Dauchot

Influence of pulse duration on the plasma characteristics in high-power pulsed magnetron discharges

High-power pulsed magnetron discharges have drawn an increasing interest as an approach to produce highly ionized metallic vapor. In this paper we propose to study how the plasma composition and the deposition rate are influenced by the pulse duration. The plasma is studied by time-resolved optical emission and absorption spectroscopies and the deposition rate is controlled thanks to a quartz microbalance. The pulse length is varied between 2.5 and 20μs at 2 and 10mTorr in pure argon. The sputtered material is titanium. For a constant discharge power, the deposition rate increases as the pulse length decreases. With 5μs pulse, for an average power of 300W, the deposition rate is ∼70% of the deposition rate obtained in direct current magnetron sputtering at the same power. The increase of deposition rate can be related to the sputtering regime. For long pulses, self-sputtering seems to occur as demonstrated by time-resolved optical emission diagnostic of the discharge. In contrary, the metallic vapor ioniz...

Size and segregation effects on the phase diagrams of nanoparticles of binary systems

The phase diagrams of small particles (with diameters in the nanometre range) are studied theoretically. In the limit where thermodynamical arguments remain valid, it is deduced that the phase diagram of small particles is a function of their size. This is discussed for the cases of eutectics and regular solutions. The effects of surface segregation are also treated and lead to further modifications of the phase diagrams.

Phase diagrams of small particles of binary systems: a theoretical approach

The phase diagrams of small particles (with diameters in the nm range) are studied theoretically. In the limit where thermodynamical arguments remain valid, it is deduced that the phase diagram of small particles is a function of their size. For the case of ideal solutions, it is shown that the lens-shaped solidus-liquidus curves are shifted to lower temperatures when the dimensions of the particle decrease. Additionally, at fixed temperatures between the highest bulk melting point and the lowest melting point of the particle, the relative concentrations of the solid and liquid phases are different in the particle and bulk material.

Study of ZrN layers deposited by reactive magnetron sputtering

Zirconium nitride films are deposited onto borosilicate wafers by reactive magnetron sputtering. The films are analysed in situ by X-ray photoelectron spectroscopy (XPS). We have studied by XPS the effects of the nitrogen partial pressure (1–100%), the subtract temperature (ambient to 450 °C), and biasing (0–80 W) on the stoichiometry of ZrN films. The N1s peak is composed of three components at 397.2, 396.4 and 395.8 eV in binding energy. These components are correlated with the three existing phases of zirconium nitride (ZrN, Zr3N4 and ZrN2). With an increase of the nitrogen partial pressure, a shift of the Zr3d line to the high binding energy and the increase of the N1s component at 395.8 eV are observed. These observations are explained by the charge transfer between Zr to N which increases with P(N2) as previously described for the Ti–N2 system [1]. The bulk stoichiometry is calculated by Rutherford backscattering and nuclear reaction measurements. The resistivity of the films is measured by the four-point probes technique. The reflectivity of the films are recorded by a spectrophotometer in the IR–Vis range. A correlation between the reflectivity and the resistivity is observed. The roughness of the films is measured by atomic force microscopy. The bias voltage has a great influence on the surface roughness and on the reflectivity of the films. The dependence of the ZrNx films structure and morphology with the discharge parameters is established.

Deposition of zinc oxide layers by high-power impulse magnetron sputtering

High-power impulse magnetron sputtering HiPIMS is an attractive technique to grow thin films since it allows one to highly ionize the sputtered metallic vapor. In such circumstances, ion bombardment of the growing film is intense and one can expect to drastically modify the physicochemical properties of the coating. Compared to films deposited by conventional dc bipolar pulsed magnetron sputtering DCBPMS , important alterations of the coating were already observed when using short 7 s high-power pulses for titanium oxide thin film depositions. In Ref. 3, crystallinity was found to be strongly modified. For depositions made at room temperature, a pure rutile phase was observed on grounded steel substrates. Although these results were interesting, some properties such as film surface roughness and density were not investigated. Moreover, the deposition rate was estimated by dividing the physical thickness of the film, as determined by mechanical profilometry, by the deposition duration. Hence, film density was not taken into account and results could be misleading as ion bombardment can increase compactness. In this study, we have chosen to use zinc oxide thin films. Zinc oxide is of great commercial importance since it is used as a seed layer for low-emissivity, silver-based, multilayered stacks employed in low-emissivity windows. For this particular application, it is crucial to get the smoothest surface possible in order to sharpen the interface with the silver layer. By building a sharp interface, one could expect to decrease the emittance of the stack. Fewer silver atoms would be trapped in the interface, allowing more silver atoms to effectively participate in the coating of in-plane electric conduction and infrared reflection.

Growth of ultrathin Ti films deposited on SnO2 by magnetron sputtering

Abstract Low-e multilayers, such as dielectric/Ag/dielectric/glass, are systems extensively used in the field of architectural glass for thermal insulation. However, the physical and chemical phenomena that occur at interfaces are still not fully understood, in particular the function of the sacrificial layer deposited between the dielectric and the silver layers. Most of the time, the sacrificial layer is made of a very thin film of titanium. In order to understand the growth modes of Ti film on SnO2 substrate, as well as the chemical mechanisms taking place at the interface, we have studied by X-ray photoelectron spectroscopy (XPS) the growth of successive amounts of titanium (additions of 0.5 nm) deposited by DC magnetron sputtering on SnO2 substrate. The Ti deposition rate was varied between 0.05 and 0.02 nm/s by varying the current target between 100 and 40 mA in order to determine its influence on the growth mechanism of Ti films. The Ti deposition on SnO2 layers induces the reduction of Sn and partial oxidation of Ti. The Ti films deposited at high deposition rate reach a surface metallic state more rapidly. The XPS results and the fitting of the Sn attenuation signal by a Ti overlayer show that the Ti deposition rate influences the titanium growth mode. We found that the growth mode changes from a Volmer Weber mode for low deposition rate (0.02 nm/s) to a pseudo Frank van der Merwe mode when the deposition rate is enhanced (0.05 nm/s).

Analysis of DC magnetron discharges in Ar- gas mixtures. Comparison of a collisional-radiative model with optical emission spectroscopy

DC magnetron discharges in argon-nitrogen gas mixtures have been characterized by optical emission spectroscopy (OES). Optical lines from the cathode species (aluminium) and the gas mixture (argon, nitrogen) have been measured at constant total pressure, as a function of the gas mixture composition and electrical power. The experimental data are compared with results of a theoretical plasma model which solves self-consistently the Boltzmann equation for electrons and the kinetic equations for aluminium, argon, molecular and atomic nitrogen states. The emission intensity variations of plasma species have been analysed versus the nitrogen relative concentration and electrical power and compared with calculated populations of emitting species with a satisfactory agreement. The variation of aluminium density versus the nitrogen concentration has been deduced by using the line intensity and excitation rates given by the model.

Optical diagnostics of d.c. and r.f. argon magnetron discharges

Abstract The plasmas of d.c. and r.f. magnetron discharges have been analysed by emission spectroscopy for argon and metal atom radiative states and by optical absorption for the Ar(3P2, 3P0) metastable and Ar(3P1, 1P1) resonant atoms. In the 6–200 mTorr pressure range for W = 1–50 W in r.f. and I = 2–50 mA in d.c. the metastable and resonant atoms are three to five times more populous in r.f. than in d.c. The densities are saturated for I > 20 mA in d.c. and W > 5 W in r.f., reaching a maximum value n Ar ( 3 P 2 ) ≈ 5 × 10 10 cm −3 . The emission line intensities I(λ) vary as I(λ) = Wβ in r.f., with β ≈ 0.5 for neutral Ar lines and β ≈ 1 for Ar ion and metal (Al, Mg, Mn) atom lines. In d.c. discharges I(λ) = Iβ, with β ≈ 2 for Ar ion and metal atom lines. These results suggest an electron density variation as ne ∞ W0.5 in r.f. and an excitation mechanism proportional to ne2 for Ar ions and metal atoms in both d.c. and r.f. discharges.

Size effects on the phase diagrams of nanoparticles of various shapes

It is known that the melting point and other phase transitions in nanoparticles vary with their size. Some crystal phases also appear preferentially in nanoparticles. These also depend on the chemical environment and the encapsulation of nanosystems. Although their shape is usually considered to be spherical-like, nanoparticles of various shapes (disks, rods, etc.) may be prepared experimentally. It is shown that, in the limit where thermodynamical arguments remain valid, the phase diagrams of nanoparticles are a function of both their size and shape. This is demonstrated for various geometrical figures. The implications for the modelling of processes involving nanoparticles are discussed.

On the phase diagram of non-spherical nanoparticles

The phase diagram of nanoparticles is known to be a function of their size. In the literature, this is generally demonstrated for cases where their shape is spherical. Here, it is shown theoretically that the phase diagram of non-spherical particles may be calculated from the spherical case, at the same surface area/volume ratio, both with and without surface segregation, provided the surface tension is considered to be isotropic.