Optical phase conjugation in the short-time range: Generation of sub-nanosecond laser pulses with Q-switching by stimulated Brillouin scattering (SBS)

Abstract

We studied Q-switch1,2 of a pulsed Nd: YAG laser using the phase-conjugating properties of stimulated Brillouin scattering (SBS) inside the laser cavity. The laser cavity (see Fig. 1) consists of a flashlamp-pumped Nd:YAG rod as the active medium and a conventional two-mirror arrangement (M1, M2) with a SBS-cell containing acetone surrounded by two lenses (L1, L2). Caused by a non-confocal setup of the lenses, one can achieve high geometrical losses (per round trip) of the laser radiation emitted from the Nd:YAG rod. The low feedback between the resonator and the active medium continues as long as the intensity in focal point F inside the SBS-cell is below threshold for SBS. Thus a high inversion population is built up in the Nd:YAG rod during the increase of the flashlamp intensity. At high intensities of the laser wave-phase-conjugated backward scattering starts from F in the liquid switching the resonator quickly from high to low losses (Q-switching). As a result a short laser pulse is generated, which has—caused by the non-confocal setup—a measurable focal point outside the cavity. Depending on the relation between gain of the Nd:YAG rod, SBS reflectivity of the acetone and light intensity we found one, two, or three pulses of 2–3 ns duration with 10 mJ energy each, separated by the round-trip time between outcoupling mirror M2 and SBS scattering point F. These results can be explained qualitatively in the scope of ABCD-matrix theory and Gaussian beam propagation. Numerical calculations of the beam propagation permit to draw conclusions concerning the existence of a focal point and the "prehistory" of the phase-conjugated pulses. Even if starting 0.1 μm apart from the optical axis, a laser beam will exceed the cross section of the Nd:YAG rod (7-mm diameter) in 10 round trips only, which is equivalent to high geometrical losses of the laser wave in the resonator and a short "prehistory." Indeed, the laser pulse appeared 80 μs after the peak of the pump pulse (FWHM 500 μs). By changing the oscillator configuration (resonator length, position, and focal length of lenses), we achieved pulse peaks with 100 ps duration on top of a broader background pulses (signal to background up to 10:1). By using a saturable absorber outside the laser resonator, the background was suppressed and up to 1 mJ pulse energy left. The sub-nanosecond pulses were recorded with a streak camera at the second harmonic (see Fig. 2). These pulses appeared 100 μs before the peak of the flash lamp intensity. As explanations for the highly reduced pulse lengths of the output pulses optical breakdown by generation of a plasma in focal point F, mode-locking by multiple reflection between M2 and F and pulse shortening by oscillation of the reflection point F in the SBS-cell will be discussed.