Cascade-gain-switching for generating 3.5-um nanosecond pulses from monolithic low-cost fiber lasers

Abstract

We propose a novel laser configuration that can output 3.5-$\mu$m nanosecond laser pulses based on a simple and monolithic fiber structure. Cascade-gain-switching, which converts the wavelength of nanosecond pulses from 1.55 $\mu$m to 3.5 $\mu$m by two successive gain-switching processes. Instead of using expensive pump sources at special wavelengths and bulky active or passive modulation elements for Q-switching or mode-locking, the cascade gain-switching only requires the pumping of an electric-modulated 1.55-$\mu$m pulsed laser and two continuous-wave (CW) 975-nm laser diodes. They are all standard products for fiber optic communication applications, which can greatly lower the cost of mid-infrared laser pulse generation. To investigate the feasibility of this configuration, we numerically simulated the cascade-gain-switching processes by comprehensive rate-equation models. In single-shot regime, for stable 3.5-$\mu$m pulsed lasing, the CW 975-nm pump should be turned on at least $\sim500$ $\mu$s ahead of the 1.55-$\mu$m pulsed pump. It shows that the pulse width of the 1.55-$\mu$m pump has major impact on the temporal shape of the intermediate 1.97-$\mu$m pulse while has neglected influence on the generated 3.5-$\mu$m pulse. On the other hand, increasing the CW pump power can significantly improve the output peak power and shorten the pulse when the pump power is less than $\sim4$ W. In the repetitive-pulse regime, we found the 3.5-$\mu$m pulse train can be stably outputted when the repetition rate is $<=100$ kHz. As the repetition rate increases, the duration of reaching the stable operation increases. When the repetition rate is large than 100 kHz, the stable operation cannot be established because the rate of consuming the population on the $^4I_{11/2}$ level of $Er^{3+}$ ions is faster than the rate of building the population.