Spectroscopy and modeling of solid state lanthanide lasers: Application to trivalent Tm3+ and Ho3+ in YLiF4 and LuLiF4

TLDR: In this paper, a theoretical laser model was developed that is easily adapted to any lanthanide ion in a crystal host, and the model was used to predict diode side-pumped laser performance.

Abstract: Lanthanide series ions are considered in the context of acquiring spectroscopic parameters and their application to modelling of quasifour-level lasers. Tm:Ho codoped crystals of YLiF4 (YLF) and the isomorphs LuLiF4 (LuLF) and GdLiF4 (GdLF) as 2.0 μm lasers are used for illustration of the experimental and theoretical techniques presented here. While these materials have similar physical properties, they differ in the strength of the crystal field at the site of optically active lanthanide dopant ions such as Tm3+ and Ho3+. This is due in part to the size of the Lu3+, Y3+, and Gd3+ ions, which comprise part of the host lattice, but ionicity plays a role as well. This selection of lanthanide: host materials provides a useful basis on which to assess laser materials with regards to changes in the strength of the crystal field at the dopant ion site. It is demonstrated that Tm:Ho:LuLF has a larger crystal field splitting than Tm:Ho:YLF and Tm:Ho:GdLF, leading to smaller thermal populations in the Ho lower laser level. To assess this effect quantitatively, the energy levels of the first ten manifolds in Ho:LuLF have been determined. Measurement of Ho:XLiF4 (X=Y,Lu,Gd) emission cross sections at 2.0 μm, Tm:XLiF4 pump absorption cross sections around 0.78 μm, manifold to manifold decay times and energy transfer parameters in Tm:Ho:XLiF4 systems are also determined to provide a consistent set of parameters to use in laser modeling. The techniques presented here are applicable to any lanthanide series ion in a crystalline host. A theoretical laser model has been developed that is easily adapted to any lanthanide ion in a crystal host. The model is used to predict diode side-pumped laser performance of Tm:Ho:LuLF and Tm:Ho:YLF using input parameters determined from the spectroscopy presented here. An explanation is presented for the improved performance of Tm:Ho:LuLF over Tm:Ho:YLF by modeling the laser. A demonstration that small changes in lower laser thermal population can substantially alter laser performance is noted, an effect that has not been fully appreciated previously.