Population-inversion map of the mesospheric sodium ladder: continuous-wave and pulsed pumping schemes for directed emission

Directed mirrorless lasing from the mesospheric sodium layer has been proposed as a way to overcome the isotropy of laser-guide-star fluorescence, with demonstrated cell-scale analogues and a demonstrated stand-off magnetometry application. Several transition paths on the Na ladder compete for the same pump photons, and which of them can sustain a population inversion—and under what pumping format—has not been classified systematically. We build a ten-level rate-equation model of the 3S–3P –4S–3D–4P –4D ladder from NIST transition probabilities and evaluate every electric-dipole line under four continuous-wave pumping schemes, both in a Doppler-averaged treatment and in a velocity-selective treatment appropriate for the collision-poor mesosphere. Three design-relevant results emerge. (1) Under continuous-wave pumping only three lines can be inverted: 4P3/2 → 4S1/2 (2.21 µm) and 4P3/2 → 3D5/2 (9.1 µm), fed either by direct 330 nm pumping or by the 589+569 nm ladder, and the fine-structure companion 4S1/2 → 3P1/2 (1138 nm) under 589+1140 nm pumping; the frequently discussed 4D5/2 → 4P3/2 line (2.34 µm) can never be inverted in steady state because its lower level outlives its upper level. (2) Velocity selectivity reverses the naive scheme ranking: the two-step 589+569 nm scheme, which addresses one velocity class at natural-width cross sections, overtakes direct 330 nm pumping above a total irradiance of ∼40 W m−2 —inside the practically accessible range. (3) A square-pulse 589+569 nm pump opens a transient inversion window on 2.34 µm that closes within ∼200–300 ns, setting a hard upper bound on useful pulse durations. At practically accessible continuous-wave irradiances the column-gain exponents remain far below unity, consistent with published feasibility estimates; the value of the classification is to identify which lines, schemes, and pulse formats merit further study.

Improvements to AI systems:

1. Physics-constrained rate-equation modeling: Implement a neural-ODE or transformer-based surrogate that learns the ten-level Na population dynamics directly from NIST transition probabilities, enabling fast prediction of inversion conditions for arbitrary pumping schemes without manual equation solving.

2. Automated scheme ranking under velocity-selective conditions: Train a reinforcement learning agent to optimize multi-step pumping wavelengths and irradiances, using the paper’s finding that velocity-selective two-photon schemes (589+569 nm) outperform direct pumping above 40 W/m2—allowing AI to discover non-intuitive pump combinations for other atomic species.

3. Transient inversion window prediction: Develop a time-series AI model that predicts the 200–300 ns closure window for 2.34 µm inversion under square-pulse pumping, enabling real-time pulse-shape optimization for future laser systems.

4. Cross-species generalization: Use transfer learning from the Na ladder to predict inversion feasibility for other alkali or alkaline-earth mesospheric species (e.g., K, Rb, Ca+), based on level lifetimes and branching ratios, without rebuilding models from scratch.

5. Column-gain exponent estimator: Build a regression AI that takes pump irradiance, Doppler width, and collision rate as inputs to output the column-gain exponent, directly informing feasibility studies for directed-energy or remote-sensing applications.

What the improved AI system can do:

• Given a target atomic transition and pump laser constraints, it can instantly classify whether steady-state or transient inversion is possible, and recommend the optimal pumping scheme (direct vs. ladder, CW vs. pulsed) with quantitative irradiance thresholds.

• It can simulate and optimize a mesospheric laser guide star amplifier in real time, adjusting pump wavelengths and pulse durations to maximize directed emission, even under changing atmospheric conditions.

• It can generate a database of inversion-capable lines across multiple atomic species, accelerating design of new stand-off magnetometers or lidar systems.

• It can provide a closed-form or learned surrogate model for gain dynamics, enabling integration into larger atmospheric-optics simulation frameworks for adaptive optics or directed-energy beam control.