Controlling the Temporal Pulse Shape of the Passively Q-Switched Laser by Nonlinear Feedback Control

Abstract

Q-switched lasers are high intensity laser pulses with extremely short durations of a few nanoseconds. The application of such laser pulses is ubiquitous [1,2]. In many dermatological procedures, intense energy pulses of Q-switched lasers are used to evaporate pigmented cells with high precision while the ultra-short pulse duration prevents damaging the surrounding healthy tissues. These lasers are also used in many industrial applications that require precision cutting, ranging from laser ablation and micromachining to the high-speed drilling of holes. However, the necessary precision of desired pulse characteristics has become increasingly demanding. Ever shrinking dimensions of micromachining call for reliable pulses with predictable properties. Currently, the difficulty of producing identical pulses of prescribed intensity, shape, and duration limits the use of Q-switched lasers. While the authors have previously worked on actively Q-switched lasers [3], this work focuses on passive Q-switching, where saturable absorber materials function as the internal switch in the laser cavity. In this paper theoretical tools from the field of nonlinear dynamical control systems [4] are applied to design a feedback controller circuit, which can produce a Q-switched pulse with customized characteristics by varying the pump rate, which can be influenced by the pump power. The pump power is considered as the input to the system and the photon number is the output of the system. The population inversion, the saturable absorber concentration and the photon number are considered as the states of the system. First it is demonstrated that by applying nonlinear state feedback, the Q-switched laser system can be transformed into a pair of two cascaded integrators. Therefore, the desired output can be customized by applying its second derivative as the input. Cr:YSO Q-switched Cr:LiSAF laser is used as an example to demonstrate our method. Both theoretical analysis of the method and the numerical simulation results are presented.