1.1.

Motivation for DPAL Research

Diode-pumped alkali lasers (DPALs) provide high-power and multi-wavelength output, advantages not found in other laser systems that uniquely benefit many applications.¹ The DPAL is pumped by diode bars or stacks through the alkali’s ${\mathrm{D}}_2$ transition to its ${}^{2}P_{3/2}$ state, then collisionally relaxed to the ${}^{2}P_{1/2}$ state where it lases in the near-infrared (NIR) spectrum along the ${\mathrm{D}}_1$ transition. DPAL performance is optimized when spin-orbit relaxation with a buffer gas (usually a rare gas or small hydrocarbon) is much faster than the optical excitation rate.^2–4 A rubidium DPAL was first demonstrated in 2003,⁵ and in less than 10 years, these laser systems have been scaled to kilowatt power levels,^6,7 analytic lasing models have been developed,^2,3,8 and operating optically pumped alkali vapor laser wavelengths have been verified to span the atmospheric transmission window.^9,10 The alkali gain medium’s ability to produce various beams at a variety of wavelengths from ultraviolet through long-wave infrared is a versatility that requires further study and measurement.^11–14 Several comprehensive DPAL reviews^4,15,16 outline the benefits of this laser’s high quantum efficiency, gaseous gain medium, reduced thermal issues, diode pumping, and scalability to high output powers. This paper examines two key characteristics of a Cs alkali laser: (a) the interplay of amplified spontaneous emission, two-photon absorption, and four-wave mixing as pumping mechanisms and (b) the relation among temperature, vapor density, and laser output power.

1.2.

Experimental Context

The interplay of amplified spontaneous emission, two-photon absorption, and four-wave mixing in alkali vapor lasers is explained in several ways. Gai et al.¹¹ showed that modulating a cesium vapor laser with an infrared laser modulated the UV and blue laser output and concluded that amplified spontaneous emission and four-wave mixing were the key operating mechanisms. Pitz et al.¹⁷ theorized that both cascade lasing and four-wave mixing may both be occurring and may also be in competition with one another for producing the blue beam. Sell et al.¹⁸ examined Rb vapor lasers and arrived at a similar conclusion. In contrast, Sulham et al.¹⁴ concluded that two-photon absorption was the key mechanism for explaining Rb and Cs alkali laser output. Our experiments and analysis provide additional evidence that multiple transition mechanisms are involved.

Along with focusing on the optical transition mechanisms, considerable research examined other mechanisms to improve the performance of alkali lasers,¹⁵ including the uses of buffer gases,¹⁷ the relative advantages of transverse and longitudinal pumping,¹⁵ and the role of heat transfer.¹⁹ Substantial work has also examined the role of temperature and density in Rb and K vapor lasers. For a K vapor cell at 553 K with 5.4 torr of helium (He) buffer gas, Pitz et al.¹⁷ reported that the threshold was $\sim 1\mathrm{mJ}/\mathrm{pulse}$ and the slope efficiency for the blue beam increased by a factor of 4 as the number density increased by a factor of 2.5. Studies of Rb lasers showed that at a cell temperature of $T=140°\mathrm{C}$, 100 mW of ultraviolet light at 398 nm was measured as a harmonic to 1 W of a 794.8-nm beam.²⁰ In another Rb laser, the maximum output power was $3.2\mu \mathrm{J}$ for a pump power of 3.3 mJ, and slope efficiencies increased from 0.03% to 0.5% for Rb as cell temperatures increased from $T=175°\mathrm{C}$ through 250°C. Cs blue emission studies performed in Ref. 14 showed efficiencies on the order of 0.02% to 0.06% for $T=175°\mathrm{C}$ to 200°C for pump transition $(6^2{S}_{1/2}\to 7^2{D}_{3/2})$ and blue emissions for the Cs $(7^2{P}_{3/\mathrm{2,1}/2}\to 6^2{S}_{1/2})$ transition. For a scaled Cs system, $10\mu \mathrm{J}$ per pulse was achieved for 100 mW, 4-ns pump pulse at 10 Hz.¹⁴ Recent measurements in Ref. 11 reported Cs ultraviolet $(8^2{P}_{3/2}\to 6^2{S}_{1/2})$ and blue $(7^2{P}_{3/2}\to 6^2{S}_{1/2})$ emissions after pumping the $6^2{S}_{1/2}\to 7^2{D}_{5/2}$ transition. Maximum powers achieved in this experiment were 700 and 150 nJ for the ultraviolet and blue emissions, respectively, for a pump pulse power of 0.8 mJ.¹¹ These findings highlight the role of temperature and density in Rb and K laser operation. In contrast, the role of temperature and vapor density in Cs alkali lasers has received less attention.

1.3.

Overview of Experimental Results

This work extends the optically pumped Cs vapor cell transition $6^2{S}_{\frac{1}{2}}\to 7^2{D}_{\frac{5}{2}}$ ultraviolet and blue emissions characterization study outlined in Ref. 11 by including the emissions for the Cs vapor cell pump transition $6^2{S}_{\frac{1}{2}}\to 7^2{D}_{\frac{3}{2}}$. A functional form equation detailed in Ref. 21 to quantitatively evaluate Rb near IR laser energy at 2.73 and 2.79 m for Rb vapor cell transitions $6^2{P}_{3/\mathrm{2,1}/2}\to 6^2{S}_{1/2}$, respectively, was applied in this study to assess the Cs ultraviolet and blue emissions. This equation, although not a theoretical solution to the Cs vapor cell emission performance, is sufficient to represent the pump threshold, bleached limit, and slope efficiency dependence on pump energy, providing further validation for the functional fit. Last, this study extends the Cs vapor cell blue emission dependence on Cs vapor density and temperature outlined in Ref. 14 by including visible emissions at 150°C and validating that pump pulse versus blue beam energy dependence detailed in Ref. 14 includes ultraviolet $8^2{P}_{3/\mathrm{2,1}/2}\to 6^2{S}_{1/2}$ as well as blue $7^2{P}_{3/\mathrm{2,1}/2}\to 6^2{S}_{1/2}$ emissions. The energy levels of Cs are shown in Fig. 1.

Fig. 1

Cs energy levels.

3.1.

Cs Emission Spectrum

Figure 4 shows the partial Cs emission spectrum from ultraviolet ($>386\mathrm{nm}$) through NIR ($<1500$) nm for pump transitions $6^2{P}_{\frac{1}{2}}\to 7^2{D}_{\frac{5}{2},\frac{3}{2}}$. This figure shows the emission wavelength diversity at a temperature and associated Cs vapor density of 200°C and $n=1.8\times {10}^{15}{\mathrm{cm}}^3$, respectively. Our emission results, in several cases, were amplified extensions of the full Cs fluorescence from ($>360\mathrm{nm}$) through NIR ($<1600$) nm reported in Ref. 23 with the exception of our recorded emission wavelengths 775.5, 778.0, and 861.7 nm, which have no corresponding fluorescence signatures in Ref. 23. The full spectrum Cs vapor measurements in Ref. 23 were taken at a cell temperature of 247°C (Cs vapor density of $n=8.9\times {10}^{15}{\mathrm{cm}}^3$) after pump excitation of 455.7 nm $6^2{S}_{\frac{1}{2}}\to 7^2{P}_{\frac{3}{2}}$. The spectrum in Ref. 23 showed atomic transitions up to the ionization limit. The different pumping schemes and Cs cell temperatures between our measurements and the fluorescence spectrum in Ref. 23 may account for the differing emissions in our results.²³ The spectrum in Ref. 23 shows evidence of ionization in the plasma broadening of higher F states and blue Cs recombination continuum between 430 and 510 nm not observed in our measurements as well as weak forbidden transitions $n{P}_{\frac{3}{2}}\to 6P_{\frac{3}{2}}$ for $n=9$, 10, and 11 in Ref. 23 not present in our observations.

3.2.

ASE Laser Pulse Evolution

Figure 5 shows 672.3-nm lasing output for Cs vapor cell temperature of 150°C for pump beam energies less than 8 mJ using the equipment setup in Fig. 2. The target state, Cs $7^2{D}_{3/2}$, was produced by two-photon absorption of the ground state, Cs $6^2{S}_{\frac{1}{2}}$ at a wavelength near 767.8 nm. The lasing was detected by a C31034A photomultiplier tube with narrow bandpass filters. A Thorlabs FES0750 short-pass filter with a cutoff wavelength of 750 nm was used to minimize scattered pump light yet also allowed over 88% transmission for 672.3-nm emissions. A second Andover Corp 670.8-nm filter with a center wavelength and 10-nm bandwidth filter was used in conjunction with the short-pass filter.

Fig. 5

Cs 672.3 nm ($7^2{D}_{3/2}\to 6^2{P}_{1/2}$ laser output for cell temperature $T=150°\mathrm{C}$ and pump energies of ○—5.8 mJ, □—6.9 mJ, and ▵—7.6 mJ.

The 670.8-nm filter had a 66% transmission at 672.3 nm. The scattered pump laser intensity was measured and subtracted from the observed emission profiles by tuning off resonance, thereby minimizing the scattered light contribution to the $7^2{D}_{\frac{3}{2}}\to 6^2{P}_{\frac{1}{2}}$ transition emission. Energy measurements could not be performed for 672.3-nm output beams due to the minimal separation between red ASE and pump beam as well as the associated damage risk presented to the RjP-735 pyroelectric energy meter detector. Pump energies were kept below 8 mJ per pulse due to saturation effects on the C31034A photomultiplier for pump pulse energies above this energy per pulse level. ASE at 672.3 nm was also observed for vapor cell temperatures at 175°C and 200°C, and initial indications suggest for a given pump intensity, 672.3-nm ASE intensity increases for higher Cs alkali density. Clearly, 672.3-nm lasing occurred at 150°C, and beam output increased with increasing pump energies. A more detailed analysis is required to determine how the Cs red beam output scales with alkali density and pump intensity and what Cs vapor density is required for the onset of red lasing.

Figure 6 shows the evolution from fluorescence emission (heatpipe temperature $<75°\mathrm{C}$) for laser beam output (heatpipe temperature $>100°\mathrm{C}$) for Cs emission 672.3 nm $7^2{D}_{\frac{3}{2}}\to 6^2{P}_{\frac{1}{2}}$. The target state, Cs $7^2{D}_{\frac{3}{2}}$, was produced by two-photon absorption of the ground state, Cs $6^2{S}_{\frac{1}{2}}$, at wavelength 767.8 nm. The population of the excited state was measured using the C31034A photomultiplier tube to detect the side fluorescence for temperatures at and below 75°C and lasing emissions for Cs heatpipe temperatures at or above 100°C. Figure 6 indicates that 672.3-nm lasing onset occurred between cell temperatures of $T=75°\mathrm{C}$ and $T=100°\mathrm{C}$ corresponding to Cs vapor densities of $n=3.1\times {10}^{12}{\mathrm{cm}}^3$ and $n=1.5\times {10}^{13}{\mathrm{cm}}^3$, respectively. The rapid decrease in intensity at $T=100°\mathrm{C}$ after reaching peak intensity compared with longer measured decay rates at $T=75°\mathrm{C}$ and below, consistent with previous fluorescence decay profiles,^24,25 is indicative of pulsed lasing activity. The time scale of these emissions for cell temperatures $T=100°\mathrm{C}$ approaches pulses of 20 to 30 ns, roughly two to three times the 10 ns pump pulse duration. Figure 6 also confirms previous observations made during the 7D spin-orbit study²⁴ that heatpipe temperatures above 75°C result in longer decay rates due to radiation being absorbed and remitted before escaping the heat pipe. The laser dye pump pulse at the entrance window of the heatpipe was $\sim 16\mathrm{mJ}$ per pulse.

Fig. 6

Cs 672.3 nm $(7^2{D}_{3/2}\to 6^2{P}_{1/2})$ laser development from florescence Cs vapor at $T=25°\mathrm{C}$ through laser emissions at $T=100°\mathrm{C}$ through $T=150°\mathrm{C}$.

3.3.

ASE Laser Output

The dependence of the ultraviolet and blue output energies on pump energy at three temperatures for the pump transitions $6^2{S}_{\frac{1}{2}}\to 7^2{D}_{\frac{5}{2}}$ and $6^2{S}_{\frac{1}{2}}\to 7^2{D}_{\frac{3}{2}}$ are shown in Figs. 7 and 8, respectively. Output energies grow linearly until pump pulse energies of 6 to 8 mJ and gradually level off through the highest pump energy used at 31 mJ. For all emission wavelengths, the energies increased with cesium vapor density. The uncertainty in the output energies represents the 95% confidence bounds for the associated lasing intensities measured by the C31034A photomultiplier tube. A functional form, first used and described in Ref. 21, was used to fit these results

Fig. 7

Ultraviolet (387, 388 nm) and visible (455, 459 nm) for pump transitions $6^2{S}_{\frac{1}{2}}\to 7^2{D}_{5/2}$ for three temperatures and corresponding Cs vapor densities: (□) $T=150°\mathrm{C}$, $n=2.26\times {10}^{14}{\mathrm{cm}}^{-3}$, (○) $T=175°\mathrm{C}$, $n=6.81\times {10}^{14}{\mathrm{cm}}^{-3}$, and (▵) $T=200°\mathrm{C}$, $n=1.82\times {10}^{15}{\mathrm{cm}}^{-3}$.

Fig. 8

Ultraviolet (387, 388 nm) and visible (455, 459 nm) for pump transition $6^2{S}_{1/2}\to 7^2{D}_{3/2}$ for three temperatures and corresponding Cs vapor densities: (□) $T=150°\mathrm{C}$, $n=2.26\times {10}^{14}{\mathrm{cm}}^{-3}$, (○) $T=175°\mathrm{C}$, $n=6.81\times {10}^{14}{\mathrm{cm}}^{-3}$, and (▵) $T=200°\mathrm{C}$, $n=1.82\times {10}^{15}{\mathrm{cm}}^{-3}$.

Threshold conditions were achieved for both pump transitions for pump energies of 0.5 to 3.5 mJ, as shown in Fig. 9. Error bounds in Fig. 9 are reported for the 95% confidence intervals for the $E_{\mathrm{th}}$ fit parameter. Consistent with results reported in Ref. 21, for higher pump intensities, there were not sufficient cesium atoms in the pumped volume to process all the incident pump photons. The population in the ground state was depleted, and the pump transition became transparent. As the alkali density increased, the maximum output energy increased linearly because additional pump photons can be processed. The linear fits in Fig. 10 indicate a small positive intercept, which could be interpreted as a threshold requirement or may indicate minor curvature at higher alkali densities. For both pump transitions, the trend is for the ultraviolet emissions (387, 388) nm to reach higher bleached limits than the blue (455, 459) nm transitions; however, the results from Eq. (1) were insufficient to adequately capture the $E_m$ for 387 nm emission for pump transition $6^2{S}_{\frac{1}{2}}\to 7^2{D}_{\frac{3}{2}}$. Error bounds in Fig. 10 are reported for the 95% confidence intervals for the $E_m$ fit parameter. The slope efficiencies in Fig. 11 also improve as the Cs density increases. The threshold for the ASE beams for both ultraviolet and blue occurred for Cs densities at $T=120°\mathrm{C}$ (Cs density of $n=4.9\times {10}^{13}$).

Fig. 9

Dependence of the initial energy, $E$’th, on the cesium density, when pumping $6^2{S}_{1/2}\to 7^2{D}_{5/2}$ (open symbols) and $6^2{S}_{1/2}\to 7^2{D}_{3/2}$ (filled symbols) for 387 nm (○, •), 388 nm (□, ▪), 455 nm (◊, ♦), and 459 nm (▵,▴).

Fig. 10

Dependence of the maximum achieved output energy, Em, on cesium density when pumping $6^2{S}_{1/2}\to 7^2{D}_{5/2}$ (open symbols) and $6^2{S}_{1/2}\to 7^2{D}_{3/2}$ (filled symbols) for 387 nm (○, •), 388 nm (□, ▪), 455 nm (◊, ♦), and 459 nm (▵,▴).

Fig. 11

Dependence of the initial slope efficiency, $\eta$, on cesium density when pumping $6^2{S}_{1/2}\to 7^2{D}_{5/2}$ (open symbols) and $6^2{S}_{1/2}\to 7^2{D}_{3/2}$ (filled symbols) for 387 nm (○, •), 388 nm (□, ▪), 455 nm (◊, ♦) and 459 (▵,▴).

4.1.

Two-Photon Saturation Intensity

Both threshold and pump efficiency should depend on the absorption cross-section. Two-photon absorption cross-sections are quite low. Although Cs two-photon cross-sections have not been reported in the literature, calculations using a methodology outlined in Ref. 15 predict two-photon cross-sections for pump transitions $6^2{S}_{\frac{1}{2}}\to 7^2{D}_{\frac{5}{2},\frac{3}{2}}$ to be $6.0\times {10}^{21}{\mathrm{cm}}^4/\mathrm{W}$ and $2.5\times {10}^{21}{\mathrm{cm}}^4/\mathrm{W}$, respectively. At the threshold pump energy of 3 mJ (the low end of our pump energies), the corresponding pump intensity, $I$, is $2\times {10}^6\mathrm{W}/{\mathrm{cm}}^2$. To derive the two-photon saturation intensity, we consider the time dependence of the excited state number density²⁶

The saturation intensities, $I_{\text{sat}}$ for pump transitions $6^2{S}_{\frac{1}{2}}\to 7^2{D}_{\frac{5}{2},\frac{3}{2}}$, are 30 and $50\mathrm{kW}/{\mathrm{cm}}^2$, respectively. With saturation parameters of $S=I/I_{\text{sat}}$ of 64 and 41, the sample is strongly bleached for both pump transitions.

4.2.

Optical Excitation Pathways

Although the primary lasing mechanism for many Cesium alkali lasers is often assumed to be two-photon absorption followed by cascade lasing,^{14,19,21,24,27} four-wave mixing can also serve as the central transition mechanism. For example, Gai et al.¹¹ explained the 455.6-nm output from a Cesium vapor cell pumped via 767.2 nm with four-wave mixing, i.e.,

They experimentally demonstrated modulating a $2.42\text{-}\mu \mathrm{m}$ input beam modulated the 455.6-nm output beam, strongly supporting the argument of FWM as the key mechanism. Although Sulham et al.¹⁴ concluded that two-photon absorption was the key mechanism because they observed blue light when the cesium cell was pumped at 743.2, 672.2, and 767.8 nm, they also observed an infrared beam of less than $2.5\mu \mathrm{m}$, which is consistent with the four-wave mixing explanation offered by Gai et al. However, four-wave mixing cannot explain all the laser output we observed because some of the required intermediate transitions, $7^2{D}_{\frac{5}{2}}\to 7^2{P}_{\frac{1}{2}}$ and $7^2{D}_{\frac{5}{2}}\to 8^2{P}_{\frac{1}{2}}$, are forbidden. As photoionization is orders of magnitude less likely than absorption at these wavelengths,²⁷ we reject photoionization as a plausible mechanism. The excitation pathways are most likely a combination of cascade lasing triggered by two-photon absorption and ASE mediated by FWM.