A. Miotello

Critical assessment of thermal models for laser sputtering at high fluences

A critical assessment of thermal models for laser sputtering at high fluences is presented. It is argued that the model explaining such sputtering by involving a subsurface superheating effect misinterprets the meaning of ‘‘vaporization’’ and ‘‘boiling’’. As a result inappropriate boundary conditions are used, including those for both the surface temperature and for the surface temperature gradient. On the contrary, it is shown that explosive boiling (also termed phase explosion) in the sense pioneered by Martynyuk and by Fucke and Seydel remains the only thermal mechanism able to explain laser sputtering at high fluences.

Comments on explosive mechanisms of laser sputtering

Laser sputtering differs from ion sputtering mainly in the important role played by thermal effects. These include not only vaporization from the extreme outer surface and boiling from an extended near-surface region, but also two additional effects which are important at high fluences and short pulse lengths. The first is phase explosion (also termed “explosive boiling”) in the sense analyzed by Martynyuk and by Fucke and Seydel. For high fluences and short pulses the target is unable to boil because the time scale does not permit the necessary heterogeneous nuclei to form. It therefore approaches more or less closely to Ttc (the thermodynamic critical temperature), homogeneous nuclei form at a high rate, and the near-surface region relaxes explosively into a mixture of vapor and equilibrium liquid droplets. An alternative explosive mechanism, the subsurface heating model, has, however, also been postulated for high fluences and short pulses. The idea began in 1972 with work by Dabby and Paek, and continued in a large number of articles to the present day. In this model vaporization from the surface causes the target to lose the ideal exponential temperature profile (T α exp(−μx), μ being the absorption coefficient) and to develop a modified profile such that the target is much hotter (up to 3000 K) just beneath the surface. We point out that the surface conditions chosen by Dabby and Paek are wrong. We then present a numerical solution to the problem and find that the subsurface region is indeed hotter but only by a few degrees. Explosive release of material therefore cannot occur by this mechanism but only by phase explosion.

On the mechanisms of target modification by ion beams and laser pulses

Abstract A comparison is made between the importance of the following categories of processes in ion-surface and laser-surface interactions: ballistic, thermal-spike, residual-defect-induced, and electronic. It is shown that ballistic processes are important for ion sputtering, ion mixing, and ion composition change. Also, they exist with laser-pulse sputtering but are not important. The picture with thermal-spike processes is the inverse, such processes being unimportant in most ion-surface interactions, but very important indeed with laser-pulse sputtering. In the latter case one must distinguish between normal vaporization and phase explosion, both of which are fundamental. On the other hand, normal boiling may possibly be unimportant owing to the negligible density of heterogeneous nuclei formed in the bulk, ∼ 106 kg−1. However, it is essential to note that, if nuclei can be formed at the surface, the density will be much higher. Sputtering due to subsurface heating probably does not exist. Residual defects are basic to all ion-surface interactions, but most especially to ion mixing. Here most (∼90%) of the observed mixing is due to residual defects and only a lesser amount is due to ballistic events. But the role of thermal spikes in ion mixing appears to have been exaggerated. Finally, electronic processes are somewhat important to ion sputtering but have not yet been quantified for ion mixing and composition change. On the other hand, although they are very important indeed with laser-surface interactions, a basic qualification must be made: the primary laser-surface interactions are dominantly electronic, but most of the deposited energy is rapidly converted from excitation to heat, and it is for this reason that thermal-spike sputtering is so prominent. Finally, we consider the fact that laser sputtering has two aspects. These are the primary processes, which lead to particle expulsion, and the secondary processes. The latter arise due to collisions amongst the emitted particles, with the result that the particles, effectively, “lose memory” of the primary process.

Spectroscopic characterization of thermally treated carbon-rich Si1−xCx films

Amorphous carbon-rich silicon carbide films Si0.45C0.55, deposited on silicon, were obtained by r.f. magnetron sputtering of sintered SiC targets in argon plasma and characterized by means of X-ray photoelectron spectroscopy, Auger electron spectroscopy, Rutherford backscattering spectroscopy and elastic recoil detection analysis. The structural evolution of these films upon thermal annealing at various temperatures in different atmospheres were investigated by means of Raman analysis and IR absorption. The formation of regions of crystallized SiC and diamond-like carbon as well as the hydrogen chemical state evolution are discussed in terms of chemical bondings. The processes of carbon segregation and crystallization of silicon carbide in the films are influenced by (i) high temperature treatments, (ii) annealing atmospheres and (iii) hydrogen dynamical behaviour.

Primary and secondary mechanisms in laser-pulse sputtering☆

Abstract Primary sputtering mechanisms are conventionally grouped in terms of the categories collisional, thermal , and electronic . With pulsed photons one must. in addition, consider the emission of droplets and fragments in thermomechanical processes. Pulsed photons also lead to the situation that the density of emitted particles is sufficiently high for gas-dynamic effects to enter and for the system therefore to lose memory of the primary mechanism. One then distinguishes secondary mechanisms which include outflow, as when a finite reservoir expands, effusion , effusive release from the outer surface without recondensation, and recondensation . effusive release with recondensation. If the photon pulse interacts with the emitted particles then still further secondary mechanisms are relevant due to energy deposition in the plume of emitted particles as well as due to ionization. Finally. the laser-pulse sputtering of the polymer PMMA (polymethylmethacrylate) and the superconductor YBCO ( YBa 2 Cu 3 O 7− x ) is discussed on the basis of explicit photographs of the sputtered particles. In the case of PMMA there are two groups of particles. the first group having primary and secondary mechanisms which are presently unestablishable but the second group being reasonably attributed to thermal primary release and to secondary behavior of the effusion (or recondensation) type. In the case of YBCO there is only one group of particles having a primary mechanism which is almost certainly electronic and a secondary mechanism which is tentatively identified with outflow .

Catalytic effect on hydrogen desorption in Nb-doped microcrystalline MgH2

Mg and Nb-doped Mg films were deposited by rf magnetron sputtering. Morphological and structural analysis were performed by scanning electron microscopy and x-ray diffraction. The desorption kinetics has been investigated by using a Sievert-type apparatus. The overall activation energy and the reaction order controlling desorption are (141±5)kJmol−1H and n≈4 for Mg and (51±5)kJmol−1H and n≈1 for Nb-doped (5at.%) Mg. It is suggested that Nb atoms dispersed in the MgH2 grains catalyzes the dissociation of the hydride phase and that the rate limiting step in the H2 desorption is given by the H atomic migration through interconnected transformed domains of h‐Mg.

Two stages in the kinetics of gold cluster growth in ion-implanted silica during isothermal annealing in oxidizing atmosphere

The growth kinetics of gold clusters, formed by ion implantation in silica, is experimentally investigated. Isothermal sample annealing at 900 °C is performed in air atmosphere for increasing time intervals in the range between 0.5 and 12 h. Two different scaling laws of the cluster average radius with time, t1/2 and t1/3, are evidenced, proving that coarsening, i.e., Ostwald ripening, follows the stage of diffusion limited cluster growth, as the annealing time interval increases. By a comparative analysis of the two regimes of cluster growth, in the framework of linearized models for clustering processes, the value of the surface tension of gold nanoparticles in silica matrix has been evaluated.

Revisiting the thermal-spike concept in ion-surface interactions

In recent years many groups have advocated a thermal-spike model to explain a variety of experimental results in ion-irradiation of solids, as for example sputtering, mixing, compositional change, structural change, and track formation. The latter include crystal-to-amorphous transitions as well as track formation due to MeV/u particles. In this paper we reconsider the phenomena occurring during ion impact of solids looking at the time scale generally indicated as relevant for thermal-spike effects, namely a picosecond scale as shown by molecular dynamics. Sputtering, mixing, and track formation, however, will be analyzed in more detail. We consider first ion-beam sputtering and reiterate (as is already well-known) that yields which increase with the bulk temperature most often indicate merely the onset of normal vaporization. Indeed, only simulations appear to be capable of giving insight even if the information is sometimes tentative. In mixing, ballistic transport is important but not dominant. It is often argued that the additional transport is provided by thermal spikes but it is noted that such an assumption is normally not required by the experimental results. What is more relevant is a role for residual defects such that the total diffusion flux includes (if the defects are chemically guided) a modified Darken factor, or (if the defects are not chemically guided) simply an increased diffusivity. The time scale (min), distances (well beyond the collision cascade), temperature sensitivity (changes of as little as 75 K are relevant), and correlation with vacancy properties (thence with the solid rather than liquid state) which are relevant to these residual defects are not understandable in terms of thermal spikes. We finally consider track formation. Recent work claiming that track formation in solids, irradiated with heavy ions, may be understood in terms of thermal spikes is reconsidered to show that the thermal-spike model is utilized without considering all the relevant phenomena included in irradiation-induced heating and phase transitions. For example, a comparison of fs-laser pulse irradiation of Si with swift heavy-ion irradiation, shows that melting is possible in the first case since the excited electrons have a low and more or less restricted energy while in the case of swift ion-irradiation, the motion of the excited electrons includes a ballistic component which does not favour the localization of the thermal energy necessary to induce lattice melting. It is concluded that track formation is better understandable in a more general framework of defect-induced processes in solids.

Formation of silver nanoclusters by excimer–laser interaction in silver-exchanged soda-lime glass

High-power excimer-laser irradiation (248 nm wavelength) at different energy densities (1.3, 2.3, and 4.6 J/cm2) has been performed on silver-exchanged soda-lime glass. Silver nanoclusters have been obtained with average size depending on the energy density of the laser pulses. The excimer laser pulses induce either the reduction of the silver ions or the heating of the irradiated glass matrix. The high mobility of silver atoms in the liquid phase and the segregation effects at the liquid–solid interface can explain the observed silver atoms clustering.

Ion-beam mixing with chemical guidance. IV: Thermodynamic effects without invoking thermal spikes

Abstract Various studies on ion-beam mixing suggest that the extent of mixing is sensitive to the sign and magnitude of the heat of mixing Δ H m . This suggests a role, not only for random motion, but also for chemical driving forces such that the total diffusion flux includes a modified Darken factor: 1 −α i (1−α i )[ 2h m p RT(1 +p) ]  1− α i (1 − α i )q . Here α i , is the atomic fraction of component i , α i (1 − α i ) h m is the heat of mixing of a regular solution , and p is the ratio of the diffusivities for chemically guided defect motion to those for random motion of all types. We wish to evaluate the parameter p from a variety of recent experiments. We first consider the work of Marton, Fine, and Chambers on profiling multilayers of Ni-Ag and conclude that the narrowing of the Ag profiles, as temperature increases, is connected to a strongly reduced solubility. By fitting the profiles to solutions of the continuity equation corresponding to the above total diffusion flux, values of q are obtained. Provided T can be taken as the ambient temperature, values of p then follow: 0.04 ± 0.01 at 300 K, 0.15 ± 0.02 at 400 K, and 0.24 ± 0.04 at 600 K. We then consider ion-beam mixing experiments based in some cases on solubility changes (11 examples), in some cases on interface broadening versus Δ H m (40 examples), and in still other cases on interface broadening in the presence and in the supposed absence of defect-induced transport (5 examples). Considering all results, reasonably consistent values of p are obtained for temperatures ranging from 77 to 603 K. It is thus possible to understand a variety of experimental results relating to profiling and to ion-beam mixing in terms of chemical driving forces and, moreover, to do so without invoking thermal spikes . In our opinion this approach constitutes a more satisfactory framework for describing mixing (and related bombardment-induced solid-state processes) than that in which thermal spikes are assumed.