Optical Interactions with Tissue and Cells XX

BO201

Researchers who studied the potential of nanosecond (ns) and femtosecond (fs) optical breakdown for micromachining of transparent dielectrics have expressed the view that “when ultra-short pulses are used, the behavior of the breakdown threshold changes from a statistical nature to a deterministic one [1].” The larger statistical variations of ns breakdown were explained either by “extrinsic” factors such as impurities or defects or by attributing the onset of avalanche ionization to probabilistic “lucky drifts” of free background electrons in which they gain sufficient energy to cause impact ionization. Initiation of the ionization avalanche by multiphoton ionization across the electronic band gap of the pure material was assumed to be relevant only for ultra-short pulse durations.

In the present study, we show that the threshold sharpness for ns breakdown resembles that of fs breakdown if single longitudinal mode laser pulses of 1064 nm wavelength are focused into pure water at large numerical aperture. By contrast, the statistical variations are nine times larger for regular, non-seeded ns laser pulses exhibiting picosecond intensity spikes originating from longitudinal mode beating. This observation suggests that fluctuations are mainly attributable to irregularities of the laser emission, while ns breakdown is inherently governed by an interplay between multiphoton and avalanche ionization. To verify this hypothesis, we investigated the wavelength dependence of the breakdown threshold, I_th(l), in a range between 725 nm and 1025 nm using single longitudinal mode laser pulses that allow for precise threshold measurements.

We observed a stepwise increase of the breakdown threshold at wavelengths above which one more photon is required to overcome the electronic band gap between valence and conduction band. This is indicative for an intrinsic initiation of the ns breakdown process by multiphoton ionization. The measured distance between the steps corresponds to a band gap of 6.55 eV. The experimental I_th(l) data were compared with model calculations considering multiphoton-, avalanche- and thermal ionization as well as recombination and diffusion losses. The best fit between experimental and calculated I_th(l) curves was achieved when an electron-phonon collision time of 7 fs and a value of 10^15 cm^-3 for the multiphoton-produced seed electron density were assumed.

Presently, we are in the process of investigating the wavelength dependence of femtosecond optical breakdown in a wavelength range between 400 nm and 1000 nm. A monotonous I_th(l) curve will indicate that tunnel ionization is the dominant photoionization mechanism in fs breakdown while the observation of steps will indicate an important role of multiphoton ionization that has been denied by some researchers [2]. Comparison of the measured I_th(l) data with model calculation will also allow to assess the relative importance of photoionization and avalanche ionization that is passionately debated among different research groups.

[1] X. Liu, D. Du, and G. Mourou, IEEE J. Quantum Elect. 33, 1706 (1997)

[2] D. Du, X. Liu, and G. Mourou, Appl Phys B 63, 617-621 (1996)

6. KEYWORDS

Laser, optical breakdown, wavelength dependence, nanosurgery, plasma, cavitation, nanosecond pulses, femtosecond pulses

Short abstract

For optical breakdown in pure water produced by ns laser pulses (1064 nm) focused at large numerical aperture, statistical variations are 9 times less for single longitudinal mode (slm) than for multimode pulses. The high threshold sharpness similar to femtosecond breakdown is indicative for multiphoton initiation. The wavelength dependence of the breakdown threshold between 725 and 1025 nm measured using slm pulses exhibits steps that are consistent with an intrinsic interplay of multiphoton and avalanche ionization and a band gap of 6.55 eV. Model calculations yield a multiphoton-produced seed electron density of 10^15 cm^-3, and an average time of 6 fs between electron phonon collisions.