Cs D₂ line laser (852.1 nm) pumped by the photoassociation of Cs-Ar, Cs-Kr, and Cs-Xe collision pairs: Impact of rare gas partner on threshold and efficiency

Identifying three and four level laser systems having quantum defects approaching unity has been a priority of laser physics for decades because of the potential of such systems for conversion efficiencies surpassing 50% and, therefore, minimal thermal loading of the gain medium. Foremost among the laser systems identified to date is the optically-pumped Yb:glass fiber laser which is unique because of the absence of Yb³⁺ excited state absorption, and the range in splittings available between Stark sublevels of the ²F_5/2 state.¹ One liability of this laser, however, is that as much as ∼3 kT of thermal energy is dissipated in the fiber for every absorbed pump photon. Although this thermal load is of little consequence when the laser output power is less than 100 W, heat dissipation becomes a challenge when the pump power rises to the kW level. In the vapor phase, a system having potential similar to that of Yb³⁺ is the photoexcited alkali D₂ (np ²P_3/2 − ns ²S_1/2) and D₁ (np ²P_1/2 − ns ²S_1/2) line atomic lasers, where n = 3–6 for Na, K, Rb, and Cs, respectively. When pumped by the dissociation of excited alkali-rare gas atomic pairs,^2–6 the alkali D₂ atomic transition, in particular, offers the benefit of a continuously-tunable quantum defect which is an enormous asset for optimizing laser efficiency while suppressing thermal dissipation by the gain medium.^2–4 Furthermore, the B ²Σ⁺ interatomic potential [correlated with Cs (6p ²P_3/2) + RG(¹S₀) in the separated atom limit, where RG is a rare gas atom] provides for several properties of the atomic upper laser level to be controlled precisely.^5,6 Indeed, Mironov et al.⁵ recently demonstrated the generation of circularly-polarized (σ⁻) coherent radiation on the D₂ line of Cs when the B ²Σ⁺ − X ²Σ⁺ transition of the Cs-Xe complex, also known as the blue satellite of the Cs D₂ line, was photoexcited with a σ⁺-polarized optical field. Of greatest interest is the result that electronic spin polarization of the 6p ²P_3/2 hyperfine states (2 ≤ F ≤ 5) can be manipulated by the pump wavelength which, in turn, specifies the value of internuclear separation (R) at which the excited alkali-rare gas complex is born. In more recent experiments, Mironov and Eden⁶ demonstrated that the sensitivity of CsXe (B ²Σ⁺) dissociation products to the wavelength and polarization of the pump optical field can be exploited to substantially enhance the laser slope efficiency. Controlling the spin polarization of the Cs (6p ²P_3/2) state with a polarized pump field is tantamount to manipulating the degeneracy (g₂) of the upper laser level and, therefore, the small signal gain of the D₂ line laser at 852.1 nm.⁶ The principle underlying these recent spin polarization experiments is that the performance of the alkali D₂ line lasers in alkali-rare gas mixtures is dependent upon detailed aspects of the B ²Σ⁺ state structure. Consequently, the laser efficiency and threshold pump energy, for example, can be expected to be sensitive to the rare gas partner selected for a given alkali atom.

We report here the results of experiments in which the performance of the Cs D₂ line laser is examined as the rare gas partner and temperature are varied. Slope efficiencies above 17% (with respect to absorbed pump pulse energy) are observed for the Cs 852.1 nm laser when Cs vapor-Ar mixtures are photoexcited at 836.7 nm, the peak of the D₂ line blue satellite. For Cs-Kr and Cs-Xe mixtures, the corresponding optical-to-optical conversion efficiencies are 12% and 27%, respectively, for a linearly-polarized pump optical field. Not only are these values the highest reported to date for alkali lasers pumped by the dissociation of the transient alkali-rare gas diatomic molecule and a linearly-polarized optical field, but the dependence of laser output energy and conversion efficiency on pump pulse energy and temperature are found to be closely tied to the structure of the B ²Σ⁺ state. Cs-Ar pairs, for example, are responsible for generating 100 μJ pulses at 852.1 nm (20 kW peak power), and this laser is least sensitive of the three laser mixtures to temperature because of the magnitude of the barrier in the B ²Σ⁺ potential at R ≃ 5.3 Å.⁵ In contrast, the reduced barrier height of the CsXe (B) state results in a slope efficiency of 27% at 453 K (kT = 39 meV) but this quasi-four level system collapses rapidly at higher temperatures. The data reported here, combined with the results of Refs. 5 and 6, demonstrate that the alkali-rare gas B ²Σ⁺ state, the gain medium temperature, and the pump polarization and wavelength can be optimized as a group to realize an atomic laser of maximum peak and average power, as well as robust temperature stability. Although the present experiments entailed single-pass pumping of the laser, long pulse (>10 μs) and CW operation are expected to yield increased efficiency and output power because the time required to build the laser optical field intensity from the noise will no longer be a factor.

Figure 1 is a partial energy level diagram for the CsAr and CsXe diatomic molecules, illustrating the X ²Σ⁺ and B ²Σ⁺ electronic states that are correlated, in the separated atom limit, with Cs (6s ²S_1/2) + Ar, Xe (5p^{6 1}S₀) and Cs (6p ²P_3/2) + Ar, Xe (¹S₀), respectively. Although the X and B interatomic potentials shown are those calculated from simulations of experimental B ← X absorption spectra,⁷ the CsAr potentials are similar to those of Refs. 8 and 9. Photoexcitation of Cs-Ar or Cs-Xe atomic collision pairs in the thermal continuum of the X ²Σ⁺ state populates the B ²Σ⁺ state^2,8,9 of the excited diatomic complex. Absorption of a λ = 842.7 nm photon by a Cs-Xe ground state pair, for example, is known⁵ to place the diatomic onto or near the peak of the barrier in the B state potential at R ≃ 5.3 Å.^5,6 Since the B-X difference potential appears to be a single-valued function for R ≥ 5.2 Å, a one-to-one correspondence exists between the pump optical field wavelength and the value of R at which the excited Cs-rare gas complex is born. Experiments reported recently^5,6 demonstrated that, contrary to expectations, the process by which the excited alkali-rare gas species dissociates is impacted profoundly by the shape of the B ²Σ⁺ interatomic potential itself. Specifically, the barrier in the CsXe (B ²Σ⁺) state of Fig. 1, for example, arises partially from the interaction of the B state with a higher-lying potential correlated with Cs (5d) + Xe (5p^{6 1}S₀) in the separated atom limit. As evidenced by measurements of the spin polarization of the Cs (6p ²P_3/2, 2 ≤ F ≤ 5) dissociation fragment, the CsXe (B ²Σ⁺) state is strongly perturbed in the 5.2 ≤ R ≤ 6 Å region. Because spin polarization of the Cs 852.1 nm laser upper level is known to decrease the effective state degeneracy and, thus, raise the small signal gain, the R-dependent interaction between the B ²Σ⁺ state and the Cs (5d) + Xe (CsXe [²Λ, Λ = 1,2]) potential has the result of lowering the threshold pump power for the laser if the pump wavelength is chosen judiciously. In the present experiments, the pump wavelength is fixed on the blue satellite peak so as to isolate an additional factor determining the barrier heights in Fig. 1 and the shape of the bound potential lying beyond ∼6 Å—namely, the polarizability of the rare gas atom which scales linearly with atomic mass.

A simplified diagram of the experimental arrangement is presented in Fig. 2. A pulsed dye laser, operating at a repetition frequency (PRF) of 10 Hz and pumped by a frequency-doubled Nd:YAG laser (λ = 532 nm), serves as the optical pump for these experiments. Having a pulsewidth of 8 ns (FWHM), the dye laser output was tunable from 833 nm to 850 nm which allows access to the entirety of the D₂ blue satellites for Cs-Ar, Cs-Kr, and Cs-Xe pairs. Pump laser pulses were introduced to, and extracted from, the resonator with two polarizing beamsplitters (PBS). The pump laser pulse energy was adjusted with a combination of a PBS (located upstream of PBS 1, shown in Fig. 2) and a half-wave plate mounted onto a rotation stage. With these optical components, the single pump pulse energy could be varied readily without altering the cross-sectional beam profile. Mixtures of Cs and research grade rare gas were contained within a borosilicate glass cell having a length, diameter, and clear aperture of 10 cm, 2.5 cm, and 2.1 cm, respectively. Separate cells were fabricated for each rare gas, and all experiments were conducted with a rare gas pressure (at 300 K) of 500 Torr which corresponds to a number density of 1.6 × 10¹⁹ cm⁻³. Both windows were tilted by 11 degrees so as to avoid Fabry-Pérot cavity effects.

The optical resonator was completed with a high reflector (HR; R > 99.9% at 852 nm) and an output coupler (OC) which transmitted 10% over the 830–870 nm wavelength interval. Notice that the optical cavity of Fig. 2 allows for only single pass pumping of the Cs 852.1 nm laser when the pump field is linearly-polarized. This design was chosen for convenience because it blocks residual pump radiation from reaching the energy detector. Although only one detector is illustrated in Fig. 2, three calibrated pyroelectric detectors provided pulse-to-pulse measurements of the Cs D₂ line laser output energy, the pump energy absorbed by the alkali-rare gas mixture in a single pass, and a known fraction of the pulse energy of the incoming dye laser beam. Multiple thermocouples, a Cu cold finger, and an oven controller maintained the temperature of the optical cell to within ±1 K during experiments. When measurements at different Cs number densities were required, the cell temperature was stabilized for a period of at least 1 h before experiments resumed.

Before leaving this section, it must be emphasized that the single-pass pumping geometry of Fig. 2 will not allow for the ultimate potential of these lasers to be determined, insofar as the maximum optical-to-optical conversion efficiency and output power are concerned. The reason for this statement is that gain is available for a limited time period, and a substantial fraction of the 8 ns pump pulse width would normally be required for the 852.1 nm optical field intensity to reach laser threshold from the spontaneous emission background. Because the optical path length from the high reflector to the output coupler is 64 cm, the optical field is able to make approximately 4 passes through the optical cell before the pump pulse terminates. Consequently, the Cs (6p ²P_3/2) radiative lifetime of approximately 30 ns¹⁰ dictates that virtually all of the output power in the present experiments is amplified spontaneous emission (ASE). Optical pump and 852.1 nm laser waveforms recorded simultaneously with photodiodes and a 2.5 GHz oscilloscope show that intense ASE is detected within roughly 2 ns of the arrival of the pump pulse and the optical output at 852 nm follows the pump waveform closely.

Figure 3 summarizes measurements of the Cs 852.1 ASE laser¹¹ pulse energies generated from Cs-Ar, Cs-Kr, and Cs-Xe mixtures as the pump pulse energy is varied. All of the data were acquired by exciting collision pairs at the peak of the D₂ line blue satellite (836.7 nm, 841.1 nm, and 842.7 nm for Cs-Ar, Cs-Kr, and Cs-Xe, respectively). In order to provide an accurate assessment of the optical-to-optical conversion efficiencies for these alkali lasers, the influence of the Cs-rare gas pair absorption coefficient (at the pump wavelength) on the data was removed by normalizing the measured Cs laser pulse energy to the pump pulse energy absorbed by the gain medium. Consequently, data are presented for absorbed pump energies up to 4 mJ for CsAr, CsKr, and CsXe in Figs. 3(a), 3(b), and 3(c), respectively. One notices immediately that, on a per-absorbed pump photon basis, the pump energy thresholds are approximately the same for all of the rare gases examined (250–300 μJ/pulse) which suggests that the quantum yield for the production of Cs 6p ²P_3/2 by the photodissociation of Cs-rare gas pairs is, to first order, independent of the rare gas. With regard to the slope efficiency, the highest value (27%) is observed for Cs-Xe pairs at low temperature (453 K). At higher temperatures, however, the performance of the Cs-Xe laser deteriorates rapidly. When T reaches 493 K, for example, the pump energy threshold is more than quadruple the value when T = 453 K and reaching laser threshold is problematic for temperatures above 495 K, regardless of the pump pulse energy. The explanation for this behavior does not lie in the increase in the Cs number density ([Cs]) with rising temperature of the saturated metal vapor because the decline of the laser's performance occurs in an extraordinarily narrow window of kT (<3.2 meV). Rather, this behavior is attributable to: (1) the small activation (energy) barrier in the CsXe(B ²Σ⁺) interatomic potential, and (2) the interaction between B ²Σ⁺ and the ²Λ state [correlated with Cs(5d) + Xe] discussed previously. The first of these is primarily responsible for the deterioration of the Cs atomic laser output at higher temperatures in the present experiments, not only for Cs-Xe mixtures but for Cs-Kr and Cs-Ar as well. Specifically, because the Cs-rare gas collision frequency exceeds 1 THz, the velocity distribution of the Cs 6p ²P_3/2 population thermalizes in 1–2 ns. It is appropriate, therefore, to describe the relative upper laser level (6p ²P_3/2) and Cs-rare gas B ²Σ populations by Boltzmann statistics, despite the short pump pulse duration. For this reason, the degradation of laser performance with increasing temperature is expected to be more pronounced if the pump laser wavelength is increased (thereby photoexciting Cs-rare gas pairs at larger values of R).

Although the Cs-Ar data of Fig. 3 do not represent the highest slope efficiency of the three rare gas partners, the results of Fig. 4 show that this pair does exhibit robust temperature stability. In panel (a) of Fig. 4, the dependence of the slope efficiency of each of the three Cs 852.1 nm lasers on temperature is given and the data corroborate the earlier discussion of the influence of the B ²Σ⁺ potential profile. For these measurements, conversion efficiency was calculated on the basis of the total (incident) pump pulse energy (as opposed to the absorbed pump energy). Because of the B ²Σ⁺ barrier height in the Franck-Condon region associated with the pump wavelength of 836.7 nm (peak of the blue satellite, cf. Fig. 1), CsAr is clearly the most thermally-stable laser of the three Cs-rare gas combinations. Consequently, the thermal population of the B state at the peak of the barrier is negligible for temperatures below 490 K. Above 500 K, however, this Cs laser also deteriorates quickly. One notes that, for T ≤ 493 K, efficiency grows linearly because the Cs number density is rising with temperature, thereby producing more Cs-Ar pairs in the Franck-Condon region (the Cs-Ar pair density, [Cs-Ar], in a specific interval in R is proportional to the product of the Ar and Cs number densities). Furthermore, as was the case for the Cs-Xe laser, the precipitous fall in Cs-Ar output and efficiency occurs in a region of kT that is less than 3 meV in width. In summary, Fig. 4(a) vividly illustrates the descending order of the barrier heights for Cs-Ar, Cs-Kr, and Cs-Xe pairs.^7–9 Below ∼470 K, the slope efficiencies for all three Cs-rare gas pairs rise together but Cs-Xe declines first at higher temperatures because its B ²Σ⁺ barrier is the lowest of the three. The laser output falls precipitously because the thermal population of CsXe(B) transient molecules reduces the population inversion. Stated in other terms, the thermal population of CsXe(B), at the value of R corresponding to the pump terminus, rises sufficiently that the 4 level laser collapses. Similar behavior is observed in Cs-Kr mixtures, and the details of this fundamental mechanism for three- and four-level lasers will be discussed elsewhere. Part (b) of Fig. 4 illustrates several of the conclusions discussed above from a different perspective by presenting the laser slope efficiency, normalized to the absorbed pump pulse energy, as in Fig. 3. The same behavior is evident but the superiority of Cs-Ar mixtures, relative to the Cs-Xe complex, is less obvious because of the magnitude of the Cs-Xe slope efficiency at temperatures below 470 K. The superior performance of Cs-Xe, relative to Cs-Ar and Cs-Kr pairs, at reduced temperatures may stem from the bound portion of the B ²Σ⁺ excited state potential predicted by theory and spectral simulations (cf. Fig. 1 ^7,8) Although shallow, the bound CsXe potential likely serves as a reservoir for Cs(6p) atoms and extends the effective lifetime of the upper laser level.

In conclusion, experimental measurements of the output energy of the Cs D₂ line (852.1 nm) laser, pumped at the peak of the blue satellite in three different Cs-rare gas mixtures, were recorded over a wide range in the Cs number density. The data reported here unambiguously tie the performance of a photodissociation laser to the structure of the intermediate molecular complex. Overall, the highest slope efficiency with respect to absorbed pump energy was determined to be 27% for Cs-Xe at 453 K, a value that is more than an order of magnitude higher than any previously reported efficiencies for Cs D₂ lasers pumped by the photoassociation of alkali-rare gas thermal collision pairs with a linearly polarized optical field.

Interatomic potentials calculated by J. D. Hewitt, and the support of this work by the U.S. Air Force Office of Scientific Research (under Grant No. FA9550-14-1-0002) and CU Aerospace are gratefully acknowledged.