分野別セミナー

Laser-matter interactions give rise to thermal gradients in the molten surface layer: (i) parallel
to the surface, ΔT‖, which is driving force for the Marangoni convection or the surface tension
instability, and (ii) vertical to the surface, ΔT⊥, which is driving force for the Rayleigh-
Benard or the thermal convection instability.
Surface tension instability: To generate ΔT‖, a Gaussian power profile on the chromium
nanolayer of 500 nm thickness has been used. The external control parameter was the laser
energy E (at N = constant number of pulses), in the confined configuration of experiment. The
surface tension gradient, ∂σ/∂T‖, caused the formation of traveling waves (TW) inclined
under an angle φ with respect to ΔT‖. The angle φ and the wavevector k change showing
alternation with increasing E, thus establishing the cascade of the left-right inclinations. The
left-inclined and the right-inclined TW can be simulated on the basis of CGLE, taking the
critical wavevector kc, and the critical frequency ωc from the experiment. However, at E ~
100 mJ, the wave inclination vanishes (φ = 0), and the wavevector k decreases to some
constant value (soft mode like behavior). This pretransitional effect is connected with
initiation of the radial fluid flow and the new type of instability which is more efficient energy
dissipation channel, like the RT and the RM instability.
Thermal convection instability: To generate ΔT⊥, a homogenized laser beam and flat power
distribution on the SiON/Si interface layer of 350 – 400 nm thick, have been used. The
external control parameter was the number of pulses N (at E = const.), in the open
configuration of experiment. The Boussinesq conditions cause the formation of domains with
the parallel roll organization, the inclined wavy-like, and the chaotic ones. The Fourier
analysis of these domains gives the structure factor S(k), and the correlation length of the
wavevectors. Numerical simulation based on 2D Swift-Hohenberg equation reproduces the
roll organization together with the secondary instabilities that evolve in time, starting with the
long-wavelength “ZigZag” instability, than both the “ZigZag” and the Eckhaus instability,
and finally with the Eckhaus instability only. However, the micrographic analysis reveals that
at N ≳ 20, the roll structures grow faster at some locations. This indicates the pretransitional
behavior connected with initiation of the fluid flow from the center to the periphery of the
spot and transition into the multiple absolute instability. Such cascade-like absolute instability
under series of pulses, causes the wavy agglomeration of rolls into well separated bands. The
roll structures become interconnected into the pattern of a high topological complexity that
resembles the topology of neural networks.