## Abstract

There has been recent interest in diode pumped metastable rare gas lasers (DPRGLs) and their scaling to higher powers, due to the advantages of excellent beam quality and high quantum efficiency. In this paper, a cw diode pumped rare gas amplifier (DPRGA) with single-pass longitudinally pumped configuration is studied theoretically based on master oscillator and power amplifier (MOPA). A five-level kinetic model of DPRGAs is first established. Then, the influences of gain medium density, pump and seed laser intensities and gain length on DPRGA performance are simulated and analyzed. The results of numerical simulation agree well with those of Rawlins et al.’s experiment. With the best set of working parameters, the amplification factor reaches 22.18 dB, at pump intensity of 50 kW/cm^{2} and seed laser intensity of 100 W/cm^{2}. Parameter optimization is helpful for design of a relatively high-power DPRGL system.

© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

## 1. Introduction

Recently, significant progress has been made in the development of a hybrid gas phase/solid state laser systems: diode pumped alkali vapor lasers (DPALs) [1–7]. However, there exist two main challenges for DPALs: active chemical properties of the gain medium (usually alkali vapor and hydrocarbons) and precise temperature control (the effective gain medium concentration is vaporized by heating). To solve these issues of DPALs, Han and Heaven et al. proposed an optically pumped metastable rare gas (Rg*) laser (OPRGL) in 2012 [8]. The gain medium of OPRGLs is their usage of rare gases (Rg) only, which makes these lasers inherently chemical inert. OPRGL is an alternative to DPAL because of their similar lasing mechanism, which combines the advantages of solid lasers and gas lasers. As a two-step pumping scheme, OPRGL includes Rg electrical discharge and optical pumping processes. When a diode laser is used as pump source, diode pumped rare gas lasers (DPRGLs) can be regarded as beam conversion systems with a wavelength in the near infrared range. In DPRGLs, a high-power laser with excellent beam quality is produced by a diode laser with poor beam quality at room temperature. In addition, DPRGLs have some other characteristics: high quantum efficiency, the narrow line-width (~MHz), medium recyclable flow heat dissipation, light weight and compactness.

Based on the advantages as mentioned above, DPRGLs have attracted increasing interest of researchers around the world. In recent years, a series of experiments and theoretical models have shown that DPRGLs are of high efficiency, high power and excellent beam quality of DPRGLs in different ways [8–16]. In 2015, Rawlins et al. have firstly realized a cw Ar* laser by microwave frequency microplasma discharge [10]. The cw laser power is 22 mW of a single OPRGL oscillator with narrow linewidth and high beam quality (Gaussian beam profile), which meets the requirements of seed laser in a cw MOPA laser system.

At the same time, researchers have done a lot of theoretical work. Demyanov et al. have established cw OPRGLs model with optical conversion efficiency of 60% based on three-level scheme under consideration of the discharge power density to sustain sufficient Ar* density in 2013 [12]. Subsequently, Yang et al. and Gao et al. developed a longitudinally and transversely diode pumped double-pass model by five-level scheme, respectively [13,14]. And we have established a complete dynamic model of DPRGLs systems with high efficiency and high power output [15]. To increase the power of these hybrid laser systems including DPRGLs and DPALs, till now, higher intensity and narrow-banded (~20 GHz) pump sources have been employed, such as multiple diode laser arrays or stacks of diode laser arrays [16,17]. Besides, similar to in traditional solid-state and fiber laser systems, there are two common ways for DPRGL power scaling: one is to extract power from a single oscillator, the other is to use the master oscillator power amplifier (MOPA) configuration.

Until now, diode pumped alkali amplifiers (DPAAs) have been demonstrated in both experimental and theoretical [18–23]. Zhdanov et al. have firstly presented Cs DPAA with 21.6 dB amplification in 2008 [18], and the first Rb DPAA with 7.9 dB amplification has been demonstrated in 2010 [19]. However, to date, all the high power DPRGL systems are performed using single oscillator, the DPRGA systems may be required for power scaling. Compared with a single oscillator, as an alternative power scaling approach, diode pumped rare gas amplifier (DPRGA) or DPRGA chain configuration is beneficial for the power scaling with excellent beam quality and reduces thermal effect of high power lasers.

In this paper, based on a five-level scheme, a kinetic model of a DPRGA system is established by taking Ar-He gas mixture as an example in order to research the characters of DPRGAs. Then, the validity of our model is verified by comparing with Rawlins et al.’s experimental results. Finally, the influence of optimization factors on single-pass amplification factor and efficiencies are analyzed in the simulation, which are helpful for designing an efficient high power DPRGA system.

In Section 2, the physical model and numerical approach of a single-pass DPRGA based on five-level scheme are introduced. In Section 3, validity of the model is verified by a comparison with Rawlins et al. experimental results. In Section 4, some key factors of laser characteristics are calculated and analyzed.

## 2. Physical model of DPRGA description

For ease of understanding and explanation, the schematic diagram of an Ar* laser MOPA system is shown in Fig. 1. The seed laser, denoted as *I _{sl}*, is produced by the Ar* oscillator, and is injected in amplifier with pump light to obtain the output laser

*I*. Both of oscillator and amplifier contain a homogeneous gain medium, namely helium and metastable Ar* atoms, which is produced by discharge near atmosphere. The difference between oscillator and amplifier is that the later one only considers the single-pass transmission without cavity mirror.

_{laser}The energy level scheme for excited rare gases is more complicated than that of the alkalis. The complete energy levels associated with the DPRGA systems contain four 1s-levels (from 1s_{2} to 1s_{5}, Paschen notation) and ten 2p-levels (from 2p_{1} to 2p_{10}). In our model, however, the kinetic of the DPRGA model can be described as a five-level structure, including the lower pump level 1s_{5}, the upper pump level 2p_{9}, the upper laser level 2p_{10}, and the resonant level 1s_{4} and the 2p_{8}. This simplification was based on the assumptions: one is that the population transfer from 1s_{2} and 1s_{3} levels is ignored because the intermultiplet transfer data of manifold 1s is incomplete, and the other is that we only consider the 2p_{8} level and ignore the 2p_{1}~2p_{7} levels because of the small energy gap between the levels of 2p_{8} and 2p_{9} (about 0.01 eV). A five-level model based on more complete energy levels and population transfer channels is shown in Fig. 2.

The detailed lasing mechanism of DPRGA is as follows: the metastable 1s_{5} state of Rg (excepting He), as the common laser and pump lower level, is generated by discharge. Optical pumping on the 1s_{5}→2p_{9} transition with subsequent collisional relaxation to the laser upper level (2p_{10} state) creates inversion between 1s_{5} and 2p_{10} states undergoing absorption and amplification processes of seed laser.

In Fig. 2, the red dotted and black lines represent the spontaneous emission processes and the collisional relaxation processes, respectively. And the orange and yellow thick lines respectively describe the pumping and lasing processes. In our model, we mainly concern the repopulation of the five related energy level by optical pumping and corresponding laser performance, and the repopulation of these levels by discharge is not included. It is assumed that the pump and seed laser intensities are homogenous along the longitudinal direction (the optical axis). Our model of DPRGA based on rate equations is described as follows:

*n*is defined as Δ

_{ij}*n*=

_{ij}*n*‒

_{i}*g*/

_{i}*g*×

_{j}*n*, in which

_{j}*n*and

_{i}*g*(the subscripts

_{i}*i*,

*j*= 1,2,3,4,5) represent the population densities and degenerate degrees of 1s

_{5}, 1s

_{4}, 2p

_{10}, 2p

_{9}and 2p

_{8}energy levels, respectively.

*A*and Δ

_{ij}*E*, between

_{ij}*i*and

*j*energy level, are the spontaneous emission Einstein coefficients and the energy gaps, respectively, which are provided by [24].

*k*are the collisional relaxation rate constants, which can be obtained by

_{ij}*k*=

_{ij}*k*

_{ij}^{Ar}×

*y*

_{Ar}+

*k*

_{ij}^{He}×

*y*

_{He}.

*y*

_{Ar}and

*y*

_{He}respectively represent the mixture fraction of argon and helium.

*k*

_{ij}^{Ar}and

*k*

_{ij}^{He}are taken from [10,12,25,26]. The reaction processes and the value of

*k*and

_{ij}*A*considered in the rate equations are shown in Table 1 and Table 2, respectively. In order to simulate the DPRLAs systems with different temperatures effectively, the rates are modified by an Arrhenius temperature scaling [10]. In our model, gain medium density, as a constant, is expressed as the initial Ar (1s

_{ij}_{5}) number density which is investigated in detail in our previous work [15].

*l*is length of the gain medium.

_{gain}*k*is Boltzmann constant, and

_{B}*T*is the absolute temperature in Kelvin.

*σ*

_{31}(

*λ*) is described the excitation cross sections in the laser emission processes for laser amplifier, which is expressed as below:

*λ*

_{31}

*=*912.3 nm is the wavelength of the seed laser and the output laser of the amplifier. The spectral linewidth of the laser transition, denoted as Δ

*ν*

_{31}, is described as below [27]:

*k*

_{31}

^{He}and

*k*

_{31}

^{Ar}are equal to 5 × 10

^{−14}cm

^{3}/s and 0.6 × 10

^{−11}cm

^{3}/s, respectively, in Table 1.

*N*represents the density of particles at atmosphere.

In addition, the pump rate, the seed laser transition rate and the laser transition rate in the amplifier are represented by Γ* _{p}*, Γ

*and Γ*

_{sl}*, respectively. And Γ*

_{laser}*is described as below [28]:*

_{p}*η*and

_{del}*η*, assumed to be 100% in calculation, represent the delivery efficiency of the pump beam and the mode-matching efficiency, respectively.

_{mod}*hv*is pump photon energy, in which

_{p}*h*is the Planck constant and

*v*is pump light frequency.

_{p}*dI*(

_{p}*λ*)/

*dλ*is the spectrally resolved pump light intensity, which can be expressed bywhere

*I*denotes the pump intensity, and we assume that the lineshape is shown as

_{p}*g*(

_{p}*λ*) in a Gaussian distribution [14]:

In Eq. (11), *c* is the speed of light, Δ*ν _{p}* and

*ν*are described, respectively, the spectral linewidth in FWHM and the corresponding frequency with the wavelength of the pump source. In order to obtain higher pumping efficiency, Δ

_{p}*ν*needs to match the corresponding Ar* absorption linewidth (FWHM) which is related to absorption cross section

_{p}*σ*

_{14}(

*λ*) in Eq. (9). Since the density of Ar* is produced by discharge under atmosphere, we assume that the homogeneous broadening (Lorentzian profile) dominates in broadening process. Therefore,

*σ*

_{14}(

*λ*) can be given as below:

*λ*

_{41}

*=*811.5 nm is the pump wavelength,

*ν*

_{41}is the corresponding frequency at 811.5 nm. Δ

*ν*

_{41}is not only related to the total pressure in discharge but also related to the partial pressure ratio of helium and argon, similar to Eq. (8), which is described as below:

*k*

_{41}

^{He}and

*k*

_{41}

^{Ar}are equal to 2 × 10

^{−12}cm

^{3}/s and 2.5 × 10

^{−11}cm

^{3}/s, respectively, in Table 1.

In our model, the seed laser is assumed to be in a single frequency and amplified being absorbed. The seed laser transition rate Γ* _{sl}* is given by

*σ*

_{13}(

*λ*) is described as the absorption cross sections for laser amplifier.

The laser transition rate Γ* _{laser}* can be expressed as

*hv*is laser photon energy.

_{l}In the case of cw operation, the solutions of the rate equations are discussed under the steady-state conditions of *d* / *dt* = 0. Combined with Eqs. (9), (14) and (15), the rate equations can be solved to obtain the dependent variables such as number density *n _{i}* and output laser intensity

*I*. Then the energy conversion channels can be expressed as follows:

_{laser}*P*,

_{laser}*P*, and

_{fluorescence}*P*are the output laser, fluorescence power and waste heat in amplifier.

_{heat}*S*is the cross section area of the laser beam.

*V*is the mode volume, which can be expressed by

*V*=

*l*×

_{gain}*S*.

## 3. Comparison with Rawlins et al.’s experimental results

In this section, validity of the model is verified by a comparison between the simulation by our model and the experiment results by Rawlins et al. The model parameters are set to ensure consistency with the experimental conditions [10]: gas pressure *P* = 1 atm, mixture fraction of Ar *y*_{Ar} = 2%, gas temperature *T* = 600 K, pump linewidth Δ*ν*_{41} = 2 GHz. The dependence of gain length on amplification factor at the pump intensity *I _{p}* = 1320 W/cm

^{2}(corresponding to 800 mW) and the seed laser intensity

*I*= 20 W/cm

_{sl}^{2}(corresponding to 12 mW) is shown in Fig. 3. In literature [10], the single-pass amplification factor in probe experiment is about 9, and in our calculation is 9.84 (that is 9.93 dB in Fig. 3). The calculated results agree qualitatively with the experimental results. The main reason for the relatively larger calculated results is that we assumed the delivery efficiency of the pump beam

*η*and the mode-matching efficiency

_{del}*η*to be 100% in simulation.

_{mod}## 4. Simulation results and discussions

In this section, the influence of several key factors on amplification factor are simulated and analyzed, and the calculations results could provide guidance for designing an efficient single-pass DPRGA system in future. We first analyze the influence of the initial Ar(1s_{5}) density (Section 4.1) on single-pass amplification factor and efficiencies of DPRGAs separately. Then, with different pump intensities *I _{p}*, seed laser intensity

*I*and linewidth of pump laser Δ

_{sl}*λ*, the dependences of the single-pass amplification factor on the gain length

*l*(Section 4.2) are respectively discussed. Finally, the influence of pump and seed laser intensities on single-pass amplification factor is determined (Section 4.3).

_{gain}#### 4.1 Influence of the initial Ar(1s_{5}) density

The gain medium density is a very important factor in designing a DPRGA system. In our model, the gain medium density is expressed as the initial Ar(1s_{5}) number density which is obtained by discharge. Yang et al. proposed that the bottleneck effect of slow relaxation rate *k*_{21} can be compensated by increasing the initial Ar(1s_{5}) number density (3 × 10^{13} cm^{−3}) to realize a cw linear lasing [13]. And the number density of metastable Ar* atoms under conditions of a dc microdischarge was measured in [29] to be 10^{14} cm^{−3}. As a result, the initial Ar(1s_{5}) number density is set as the order of 10^{14} cm^{−3} in our calculations.

Then we study the influence of the initial Ar(1s_{5}) density on single-pass amplification factor. In the calculation, the parameters are set as follows: total gas pressure is 1 atm (room temperature) mixture with 2% Ar and 98% He (i.e. *y*_{Ar} = 2% and *y*_{He} = 98%), the pump intensity is *I _{p}* = 50 kW/cm

^{2}with linewidth of 50 GHz (FWHM), and the seed laser intensity is

*I*= 100 W/cm

_{sl}^{2}. The selection of the main parameters is analyzed in detail. According to the literature [10], properly elevated temperature may be beneficial to laser performance because of faster energy transfer kinetics. Given that the lack of collisional rate constants at higher temperature (also the reliable temperature scaling laws), we provide the calculations at room temperature for a conservative estimation [13]. The simulation results are shown in Fig. 4.

In DPRGA system, uniform propagation of pump light in gain medium is beneficial to obtain higher energy laser. The uniform propagation characteristic of pump light in gain medium can be guaranteed by choosing suitable Rayleigh length, that is, by selecting the appropriate gain length. With different gain lengths of 1 cm, 5 cm, 10 cm, 15 cm, 30 cm, the dependences of the single-pass amplification factor on the initial Ar(1s_{5}) number density are shown in Fig. 4(a). It can be seen from Fig. 4(a) that the equivalence between gain length and initial Ar(1s_{5}) density. For a certain column Ar(1s_{5}) density with different gain length, the maximum single-pass amplification factor keeps the same (~22.18 dB) theoretically. But at the same amplification factor, the gain length is inversely proportional to the initial Ar(1s_{5}) density. At the same time, the Rayleigh length can be expressed as *L* = *πω*_{0}^{2} / *λ* (*ω*_{0} represents the beam waist of the pump light), which is proportional to the cross section area of pump light at a definite wavelength. However, according to the relationship between power and intensity (*P* = *I* × *S*), the pump power is proportional to the cross section area of the pump light at a certain pump intensity. To summarize, for identical intensities, the gain length is inversely proportional to the initial Ar(1s_{5}) density and is proportional to pump power. However, the higher initial Ar(1s_{5}) density poses a challenge to discharge conditions. So, it is necessary to keep a balance between the gain length and the initial Ar(1s_{5}) density in designing a DPRGA system. In our simulation model, gain length is set as a moderate value of *l _{gain}* = 15 cm when we study the influence of other factors on amplification factor.

Then the influence of initial Ar(1s_{5}) density on laser performance at different pump intensities is shown in Fig. 4(b). For a given pump intensity, we require an optimal initial Ar(1s_{5}) density to obtain the highest amplification factor, which is the result of a balance between pump absorption and non-lasing losses. To get higher amplification factor, we need a higher pump intensity and higher initial Ar(1s_{5}) density. With different pump intensities of 5 kW/cm^{2}, 10 kW/cm^{2}, 30 kW/cm^{2}, 50 kW/cm^{2}, 80 kW/cm^{2}, 100 kW/cm^{2}, the amplification factor could reach 11.79 dB, 14.95 dB, 19.90 dB, 22.18 dB, 24.25 dB, 25.24 dB at the optimal initial Ar(1s_{5}) density of 3 × 10^{13} cm^{−3}, 5 × 10^{13} cm^{−3}, 1 × 10^{14} cm^{−3}, 1.6 × 10^{14} cm^{−3}, 2.4 × 10^{14} cm^{−3}, 2.9 × 10^{14} cm^{−3}, respectively.

Figure 4(c) shows the relationship between the amplification factor and the initial Ar(1s_{5}) density for six different seed laser intensities under the same pump intensity. For the same column Ar(1s_{5}) density, the amplification factor is decreased with the increase of seed laser intensity. At initial Ar(1s_{5}) density of 1.6 × 10^{14} cm^{−3}, the optimal amplification factors are 32.14 dB, 25.16 dB, 22.18 dB, 15.28 dB, 12.42 dB and 4.23 dB under the seed laser intensities of 10 W/cm^{2}, 50 W/cm^{2}, 100 W/cm^{2}, 500 W/cm^{2}, 1 kW/cm^{2} and 10 kW/cm^{2} respectively.

Finally, the influence of initial Ar(1s_{5}) density on DPRGA performance at different pump spectral linewidths of 2 GHz, 10 GHz, 20 GHz, 50 GHz, 100 GHz, is shown in Fig. 4(d). The results show that the effect of pump spectral linewidth on DPRGA performance decreases with the increasing of initial Ar(1s_{5}) density. The reason is that the probability of collision between pump light and Ar(1s_{5}) particles decreases at lower initial Ar(1s_{5}) density. Therefore, DPRGA performance becomes more sensitive to the pump linewidth. In addition, the amplification factor decreases with the increase of pump linewidth at the same column Ar(1s_{5}) density. According to the results shown in Fig. 4, we also can clearly see that the amplification factor increases dramatically with the initial Ar(1s_{5}) density (in range of 1 ~10 × 10^{13} cm^{−3}) to its maximum value, and then it decreases slowly with the increase of the initial Ar(1s_{5}) density (greater than 1.5 × 10^{14} cm^{−3}) at a certain condition. Under the condition of *l*_{gain} = 15 cm, the aperture of the gain medium need to be 6 cm^{2} to realize 100 kW output for the pump intensity of 50 kW/cm^{2} and the seed laser intensity of 100 W/cm^{2}. The results show that a DPRGA is a compact and high power output laser system.

Figure 5 shows the relationship between the efficiencies of DPRGA and the initial Ar(1s_{5}) density. Parameters are the same as mentioned in Fig. 4. With the increase of the initial Ar(1s_{5}) density, we need to increase the pump absorption efficiency *η _{absorb}*, which, otherwise, will lead to the decrease of the laser extraction efficiency

*η*. For the absorbed pump power, there exist three main exit channels: laser, fluorescence and heat. From Fig. 5, we can see that both the fluorescence efficiency

_{opt-abs}*η*and the heat efficiency

_{fluo-abs}*η*will increase as the initial Ar(1s

_{heat}_{5}) density increases, which induces the decrease of

*η*. The results show that the maximum laser extraction efficiency is about 33%. The phenomenon is caused by the absorption of seed laser and the single-pass pumping process.

_{opt-abs}#### 4.2 Influence of the gain length

The gain length is another important factor in designing a DPRGA system. In Fig. 6, single-pass amplification factor is represented as functions of the gain length for different simulation conditions. The initial Ar(1s_{5}) density is set as 1.6 × 10^{14} cm^{−3} and the other parameters are set as what listed in Section 4.1. Firstly, we study the influence of gain length on DPRGA single-pass amplification factor at different pump intensities [Fig. 6(a)]. The results demonstrate that the amplification factor increases obviously with the gain length to its optimal value, and then it decreases slowly with the increase of gain length at a certain pump intensity.

In addition, we find that the amplification factor increases with the increase of pump intensity at the same column gain length. And the optimal gain length increases with the increase of the pump intensity. The optimal gain lengths are 9.9 cm, 15 cm and 22.1 cm at the pump intensities of 30 kW/cm^{2}, 50 kW/cm^{2} and 80 kW/cm^{2} respectively. Therefore, the gain length should be optimized to achieve the optimal amplification factor in designing a DPRGA. In Fig. 6(b), the amplification factor is represented as functions of gain length at different pump spectral linewidth. When the gain length increases, the probability of collision between pump light and Ar(1s_{5}) particles dramatically increases, and the amplification factor becomes less sensitive to the change of pump spectral linewidth. According to the result of Fig. 4(a), the gain length is inversely proportional to the initial Ar(1s_{5}) density. To obtain a DPRGA system with highest energy output, it is necessary to keep a balance between gain length and initial Ar(1s_{5}) density. From Fig. 6(c), it can be seen that the maximum single-amplification factors are 32.14 dB, 22.18 dB and 12.41 dB at the seed laser intensities of 10 W/cm^{2}, 100 W/cm^{2}, and 1 kW/cm^{2}, respectively, under the optimal gain length *l _{gain}* = 15 cm.

#### 4.3 Influence of pump laser and seed laser intensities

According to the experimental results of Rawlins et al., the output of the DPRGL is approximately dozens of milliwatts (i.e., W/cm^{2}) [10]. As a high power five-level DPRGA system, the required pump intensity is several kW/cm^{2}. In our calculation, the dependence of amplification factor on seed laser intensity at different pump intensities is shown in Fig. 7. It is seen that, for a constant seed laser intensity, the amplification factor increases with the increase of the pump intensity. In addition, when the seed laser intensity is weaker, the amplification factor is larger.

In DPRGAs, when the seed laser is absent, the Rg* works in recycling mode 1s_{5}→2P_{9}→2P_{10}→1s_{5} to convert pump photons into laser photons, and the cycling step 2P_{10}→1s_{5} is dominated by spontaneous emission. And when the seed laser is present, the cycling step 2P_{10}→1s_{5} is dominated by stimulated emission, which is much faster than the spontaneously emission process. Thus, the presence of seed laser accelerates the total atomic recycling rate, including the pump absorption step 1s_{5}→2P_{9}. And for a constant initial Ar(1s_{5}) density, the pump absorption fraction is dramatically enhanced.

## 5. Conclusions

In this paper, we set up a five-level scheme model for a DPRGA by considering the conservation of the number density. The design of DPRGA is very promising in achieving high power output. By analysis of the kinetic of laser amplification processes, we established a physical model to simulate the output characteristics of the DPRGAs systems. The results predicted by our model agree well with the experiment results in Ref [10]. We simulated and analyzed the influences of the initial Ar(1s_{5}) density, gain length and the intensities of pump and seed laser on amplification factor. From the simulation results, we can reach the following three conclusions. Firstly, the maximum single-pass amplification is existed under an optimal initial Ar(1s_{5}) density. Secondly, higher initial Ar(1s_{5}) density can result in the less sensitiveness of pump spectral linewidth and shorter gain length. Thirdly, the total atomic recycling rate can be accelerated by injecting the seed laser into the system compared with a single oscillator, which is beneficial for obtaining higher power output.

According to the simulation results, the DPRGA system is promising in achieving high power output by keeping a balance of some main parameters. The conditions of the maximum amplification factor of 22.18 dB (i.e.100 kW output) are derived as follows: the pump intensity is 50 kW/cm^{2}, the seed laser intensity is 100 W/cm^{2}, the gas pressure is 1 atm and the gain length is 15 cm. However, considering the only single-pass pumping in DPRGAs, the laser extraction efficiency of DPRGAs (~33%) is lower than that of a single oscillator (~57%) [13]. Thus, we can provide a high efficiency DPRGA system by multi-pass pumping, and a higher power output by a chain of two or more amplifiers structure.

## Funding

National Natural Science Foundation of China (NSFC) (61575072, 61775067); National Key Research and Development Project of China (2016YFB1100302).

## References

**1. **W. F. Krupke, “Diode pumped alkali lasers (DPALs) - A review (rev1),” Prog. Quantum Electron. **36**(1), 4–28 (2012). [CrossRef]

**2. **B. V. Zhdanov and R. J. Knize, “Review of alkali laser research and development,” Opt. Eng. **52**(2), 021010 (2012). [CrossRef]

**3. **J. Zweiback, A. Komashko, and W. F. Krupke, “Alkali vapor lasers,” Proc. SPIE **7581**, 75810G (2010). [CrossRef]

**4. **B. D. Barmashenko and S. Rosenwaks, “Detailed analysis of kinetic and fluid dynamic processes in diode pumped alkali lasers,” J. Opt. Soc. Am. B **30**(5), 1118–1126 (2013). [CrossRef]

**5. **A. V. Bogachev, S. G. Garanin, A. M. Dudov, V. A. Eroshenko, S. M. Kulikov, G. T. Mikaelian, V. A. Panarin, V. O. Pautov, A. V. Rus, and S. A. Sukharev, “Diode-pumped caesium vapour laser with closed-cycle laser-active medium circulation,” J. Quantum Electron. **42**(2), 95–98 (2012). [CrossRef]

**6. **B. Zhdanov and R. J. Knize, “Diode-pumped 10 W continuous wave cesium laser,” Opt. Lett. **32**(15), 2167–2169 (2007). [CrossRef] [PubMed]

**7. **Z. Li, R. Tan, W. Huang, and D. Zhang, “Quasicontinuous wave linearly polarized rubidium vapor laser pumped by a 5-bar laser diode stack,” Opt. Eng. **53**(11), 116113 (2014). [CrossRef]

**8. **J. Han and M. C. Heaven, “Gain and lasing of optically pumped metastable rare gas atoms,” Opt. Lett. **37**(11), 2157–2159 (2012). [CrossRef] [PubMed]

**9. **J. Han, L. Glebov, G. Venus, and M. C. Heaven, “Demonstration of a diode-pumped metastable Ar laser,” Opt. Lett. **38**(24), 5458–5461 (2013). [CrossRef] [PubMed]

**10. **W. T. Rawlins, K. L. Galbally-Kinney, S. J. Davis, A. R. Hoskinson, J. A. Hopwood, and M. C. Heaven, “Optically pumped microplasma rare gas laser,” Opt. Express **23**(4), 4804–4813 (2015). [CrossRef] [PubMed]

**11. **P. J. Moran, N. P. Lockwood, M. A. Lange, D. A. Hostulter, E. M. Guild, M. R. Guy, J. E. McCord, and G. A. Pitz, “Plasma and laser kinetics and field emission from carbon nanotube fibers for an advanced noble gas laser (ANGL),” Proc. SPIE **97290**, 97290C (2016).

**12. **A. V. Demyanov, I. V. Kochetov, and P. A. Mikheyev, “Kinetic study of a cw optically pumped laser with metastable rare gas atoms produced in an electric discharge,” J. Phys. D Appl. Phys. **46**(37), 375202 (2013). [CrossRef]

**13. **Z. Yang, G. Yu, H. Wang, Q. Lu, and X. Xu, “Modeling of diode pumped metastable rare gas lasers,” Opt. Express **23**(11), 13823–13832 (2015). [CrossRef] [PubMed]

**14. **J. Gao, Y. He, P. Sun, Z. Zhang, X. Wang, and D. Zuo, “Simulations for transversely diode-pumped metastable rare gas lasers,” J. Opt. Soc. Am. B **34**(4), 814–823 (2017). [CrossRef]

**15. **S. Long, Y. Qin, H. Chen, X. Wu, M. Li, X. Tang, and T. Wen, “Two-stage excitation model of diode pumped rare gas atoms lasers,” Opt. Express **27**(3), 2771–2782 (2019). [CrossRef] [PubMed]

**16. **B. V. Zhdanov, A. Stooke, G. Boyadjian, A. Voci, and R. J. Knize, “Rubidium vapor laser pumped by two laser diode arrays,” Opt. Lett. **33**(5), 414–415 (2008). [CrossRef] [PubMed]

**17. **B. V. Zhdanov, J. Sell, and R. J. Knize, “Multiple laser diode array pumped Cs laser with 48 W output Power,” Electron. Lett. **44**(9), 582–583 (2008). [CrossRef]

**18. **B. V. Zhdanov and R. J. Knize, “Efficiency diode pumped cesium vapor amplifier,” Opt. Commun. **281**(15–16), 4068–4070 (2008). [CrossRef]

**19. **D. A. Hostutler and W. L. Klennert, “Power enhancement of a Rubidium vapor laser with a master oscillator power amplifier,” Opt. Express **16**(11), 8050–8053 (2008). [CrossRef] [PubMed]

**20. **B. V. Zhdanov, M. K. Shaffer, and R. J. Knize, “Scaling of diode pumped Cs laser: transverse pump, unstable cavity, MOPA,” Proc. SPIE **7581**(75810F), 75810F (2010). [CrossRef]

**21. **P. Bai-Liang, W. Ya-Juan, Z. Qi, and Y. Jing, “Modeling of an alkali vapor laser MOPA system,” Opt. Commun. **284**(7), 1963–1966 (2011). [CrossRef]

**22. **Z. Yang, H. Wang, Q. Lu, W. Hua, and X. Xu, “Modeling of an optically side-pumped alkali vapor amplifier with consideration of amplified spontaneous emission,” Opt. Express **19**(23), 23118–23131 (2011). [CrossRef] [PubMed]

**23. **A. I. Parkhomenko and A. M. Shalagin, “An alkali metal vapor laser amplifier,” J. Exp. Theor. Phys. **119**(1), 24–35 (2014). [CrossRef]

**24. **A. Kramida and Y. Ralchenko, J. Reader, and NIST ASD Team (2018), “NIST atomic spectra database,” https://physics.nist.gov/asd.

**25. **J. Han, M. C. Heaven, G. D. Hager, G. B. Venus, and L. B. Glebov, “Kinetics of an optically pumped metastable Ar laser,” Proc. SPIE **8962**, 896202 (2014). [CrossRef]

**26. **W. Wieme and J. Lenaerts, “Excimer formation in argon, krypton, and xenon discharge afterglows between 200 and 400 K,” J. Chem. Phys. **74**(1), 483–493 (1981). [CrossRef]

**27. **P. A. Mikheyev, A. K. Chernyshov, N. I. Ufimtsev, E. A. Vorontsova, and V. N. Azyazov, “Pressure broadening of Ar and Kr (n+1)s[3/2]2→(n+1)p[5/2]3 transition in the parent gases and in He,” J. Quant. Spectrosc. Radiat. Transf. **164**, 1–7 (2015). [CrossRef]

**28. **R. J. Beach, W. F. Krupke, V. K. Kanz, S. A. Payne, M. A. Dubinskii, and L. D. Merkle, “End-pumped continuous-wave alkali vapor lasers: experiment, model, and power scaling,” J. Opt. Soc. Am. B **21**(12), 2151–2163 (2004). [CrossRef]

**29. **C. Penache, M. Miclea, A. Bräuning-Demian, O. Hohn, S. Schössler, T. Jahnke, K. Niemax, and H. Schmidt-Böcking, “Characterization of a high-pressure microdischarge using diode laser atomic absorption spectroscopy,” Plasma Sources Sci. Technol. **11**(4), 476–483 (2002). [CrossRef]