Cryogenic laser operation of a “mixed” Yb:LuYAG garnet crystal

We report on the continuous-wave and passively Q-switched operation of a compositionally “mixed” heavily doped 16.6 at.% Yb:(Y,Lu)3Al5O12 garnet crystal at cryogenic temperatures (100–200 K), pumped by a volume Bragg grating stabilized diode laser emitting at 969 nm. At 140 K, in the continuous-wave regime, a maximum output power of 10.65 W was achieved at ~ 1029 nm with a slope efficiency of 56% (versus the incident pump power), a laser threshold of 1.05 W and excellent beam quality. Using Cr4+:YAG as a saturable absorber, the passively Q-switched laser generated pulses with an energy/duration of 0.15 mJ/201 ns, respectively, at a repetition rate of 39.7 kHz, corresponding to a peak power of 0.39 kW.


Introduction
Operation of any diode-pumped solid-state laser (DPSSL) is associated with a deposition of heat in the laser medium. The heat loading leads to thermal lensing, depolarization loss and even laser ceasing due to a cavity instability or crystal fracture [1]. Ytterbium ions (Yb 3+ ), featuring a simple energy-level scheme, benefit from an in-band pumping scheme leading to a high slope efficiency and weak fractional heat loading (as compared to Nd 3+ ones). This determines the prospects of Yb 3+ -doped materials for the development of high-power DPSSLs emitting around 1 μm. Yb 3+ ions can be efficiently pumped by commercial highpower fiber-coupled InGaAs diode lasers at ~ 976 nm [2]. They also provide a relatively long lifetime of the upper laser level ( 2 F 5/2 ), improving the energy storage capabilities for passively Q-switched operation [3]. Despite their intrinsically weak heat loading, Yb-based DPSSLs suffer from thermal effects, especially at moderate to high pump powers.
One solution to alleviate the thermal effects is to cool the laser crystal down to cryogenic temperatures [4]. There are several effects of such cooling [4][5][6]: (i) the thermal properties of the material improve, i.e., the thermal conductivity increases and thermal expansion and dn/dT coefficients drop, altogether contributing to weaker thermal lensing; (ii) the absorption and stimulated-emission cross sections of the active ions increase leading to higher gain; (iii) the thermal population of the terminal laser level for quasi-three-level gain media notably decreases, thus eliminating the reabsorption loss and reducing the laser threshold; (iv) the lifetime of the upper laser level also increases. In terms of spectroscopic properties, one drawback of the cryogenic cooling is the reduction of the absorption bandwidth. This challenge can be overcome by using volume Bragg grating stabilized diode lasers or employing gain media with inhomogeneous spectral broadening of absorption bands, i.e., crystals with a structure and/or compositional disorder.
Cubic yttrium and lutetium aluminum garnets, Y 3 Al 5 O 12 (YAG) and Lu 3 Al 5 O 12 (LuAG) [7] are well-known laser host 1 3 57 Page 2 of 7 crystals for Yb 3+ doping. Due to the closeness of the ionic radii and weights of the host-forming (Lu 3+ ) and dopant (Yb 3+ ) cations, the LuAG crystal features much weaker dependence of its thermal conductivity on the Yb 3+ doping level, which is essential for high-power DPSSL development [8][9][10]. Yb 3+ ions in LuAG also offer somewhat better spectroscopic properties, namely, slightly broader ZPL in absorption at 969 nm, higher stimulated-emission cross section at ~ 1030 nm and slightly longer luminescence lifetime (as compared to Yb:YAG crystals) [7]. The growth of LuAG crystals is more complicated due to their higher solidification point and higher price of the Lu 2 O 3 reagent [11].
As a result, "mixed" (solid-solution) crystals Yb:(Y 1-x Lu x ) 3 Al 5 O 12 or shortly Yb:LuYAG were suggested to combine the advantages of both parent compounds [12]. In such crystals, the regular garnet structure is distorted by replacing part of the Y 3+ cations by Lu 3+ ones. One can also expect additional inhomogeneous spectral broadening of the spectral bands of Yb 3+ ions in such compounds. This could be particularly relevant for cryogenic lasers. Kuwano et al. reported on the growth of LuYAG crystals with different Lu 3+ contents and their properties [11]. It was shown that such "mixed" crystals exhibit good thermo-physical properties and, because of the use of only partial substitution of Y 3+ by Lu 3+ , their growth is simplified as compared to that of LuAG. Continuous-wave laser operation of Yb:LuYAG crystals and transparent ceramics was studied [13][14][15].
There exist multiple studies on diode-pumped cryogenic laser operation of Yb:YAG [16][17][18][19] and only a few concerning Yb:LuAG [20], and both continuous-wave and passively Q-switched operation regimes have been explored. David et al. reported on a cryogenic (140 K) Yb:LuAG laser pumped by a volume Bragg grating stabilized laser diode delivering 10.5 W at 1030 nm with a slope efficiency of 48.6%, showing a significant improvement over output performance achieved at elevated temperatures.
In the present work, we aimed to study the laser performance of cryogenically cooled diode-pumped "mixed" Yb:LuYAG garnet crystals in the continuous-wave and passively Q-switched operation regimes.

Laser setup
As a laser gain medium, we have used a heavily Yb 3+ -doped compositionally "mixed" (Y,Lu) 3 Al 5 O 12 garnet. A single crystal was grown by the Czochralski method using a [111] oriented seed in an Ar atmosphere and an induction heating system. The crystal was further annealed in air at 1400 °C for 12 h to remove the residual stresses and color centers causing crystal coloration. The annealed crystal was colorless. According to the ICP-AES analysis, the actual crystal composition was (Y 0.601 Lu 0.233 Yb 0.166 ) 3 Al 5 O 12 , corresponding to a doping level of 16.6 at.% Yb 3+ (ion density: N Yb = 22.6 × 10 20 at/cm 3 ). A rectangular laser element was cut from the annealed crystal having an aperture of 4.0 × 4.0 mm 2 and a thickness of 2.0 mm. Both its faces were polished to laser-grade quality with good parallelism and left uncoated. Figure 1 illustrates the schematic diagram of the cryogenic laser setup. An L-shaped cavity was formed by a concave rear mirror M1 coated for high reflection (HR) at 1030 nm (radius of curvature, RoC = − 300 mm), a plane dichroic bending mirror M2 coated for HR at 1020-1200 nm and high transmission (HT) at 900-980 nm, an antireflection (AR, at 1030 nm) coated plano-convex lens (L3) having a focal length of 150 mm and a series of plane output couplers (M3) with a transmission T OC in the range of 10%, 20%, 30%, 40% and 50% in the spectral range of 1020-1070 nm.
The uncoated laser element was pumped by means of a volume Bragg grating (VBG) stabilized fiber-coupled diode laser (numerical aperture, N.A. = 0.22, fiber core diameter: 105 µm) emitting up to 60 W of unpolarized output at 969 nm with a spectral bandwidth (FWHM) of 0.4 nm. The pump thus addressed the absorption zero-phonon (ZPL) line of Yb 3+ ions. The pump beam was collimated and focused into the laser element by two AR-coated achromatic lenses, L1 and L2 (1:2.5 reimaging ratio, focal lengths: f = 100 mm and 250 mm, respectively) resulting in a pump spot diameter of ~ 262 μm.
The laser performance was studied at several temperatures of the laser element. The crystal mounted inside a vacuum chamber was placed between the M2 folding mirror and the L3 intracavity lens at normal incidence. The crystal was wrapped into indium foil and mounted in a copper holder being conduction cooled. To avoid condensation at the sample surfaces upon cooling, the chamber was maintained at a pressure of 10 −5 mbar. The cryogenic temperatures of the active medium were achieved by a closed-cycle helium cryostat (CH-204, JANIS), which provided a maximum cooling Fig. 1 Schematic of the experimental setup for studying the cryogenic laser performance of a 16.6 at.% Yb:LuYAG crystal. LD VBGstabilized laser diode, L1 and L2 a pair of achromatic lenses, reimaging ratio: 1:2.5, M1 rear mirror, M2 dichroic mirror, L3 AR-coated intracavity lens, f = 150 mm, M3 output coupler, Cr:YAG saturable absorber, F long-pass filter power of 13.5 W at 100 K. A Lake Shore temperature controller (DT 670) was used to monitor and maintain the sample temperature, which included a silicon diode sensor and a 50 Ω heater.
The cavity was optimized for maximum output power by adjusting the separations between the optical elements. The measured optimum total geometrical cavity length was 755 mm, which included 301 mm between the M1 rear mirror and the input face of the crystal, 153 mm from its output face to the L3 intracavity lens, and 301 mm from L3 to the M3 output coupler. For passive Q-switching experiments, a saturable absorber (SA) was placed at 200 mm away from the L3 intracavity lens. Using the ABCD-matrix formalism, the mode diameters in the crystal and the SA were calculated to be ~ 220 and ~ 870 µm, respectively. The output power of the laser was measured using a thermal power sensor (S314 C, Thorlabs) and the measurement uncertainty was around ± 3% @ 1064 nm. The laser output was separated from the residual pump using a long-pass filter (FEL1000, Thorlabs) with a cut-off wavelength of 1000 nm and the transmission was around 84% @ 1030 nm.
To characterize passively Q-switched pulses, an ultrafast Si photodiode with a bandwidth of 6 GHz and a rise time of < 500 ps (UPD-500-UP, Alphalas) was used. The pulses were detected and recorded using a digital oscilloscope (Teledyne Lecroy Wavesurfer) with a bandwidth of 600 MHz and a sampling rate of 10 GS/s. The spectra of laser emission from the cryogenic laser were measured using an Ocean Optics compact spectrometer (model USB 2000+) with a spectral resolution of 0.3 nm.

Cryogenic continuous-wave Yb:LuYAG laser
First, the input-output characteristics of the CW cryogenic diode-pumped Yb:LuYAG laser were studied by fixing the output coupler transmission (T OC = 30%) and varying the crystal temperature in the range of 100-200 K with a step of 20 K, as shown in Fig. 2a. The best output characteristics were achieved for an intermediate temperature of 140 K: the laser generated a maximum output power of 8.88 W at 1028.9 nm with a slope efficiency η of 51% (vs. the incident pump power) and a laser threshold of 0.98 W. The optical efficiency at the highest incident pump power of 18.21 W amounted to 51%. For smaller crystal temperatures, the laser performance deteriorated due to the rapidly decreasing pump absorption (see below). When increasing the temperature from the optimum one up to 200 K, the maximum output power decreased (down to 5.93 W), the slope efficiency also decreased (to 38%) and the laser threshold raised (to 2.45 W), as shown in Fig. 2b. For the crystal temperatures in the range of 120-200 K, the output dependences were linear for the studied range of pump powers. Figure 2c shows the temperature dependence of the laser wavelength. With increase in the temperature from 100 to 200 K, it experienced a slight monotonous red-shift from 1028.7 to 1029.3 nm.
After selecting the optimum crystal temperature (140 K), the effect of the output coupler transmission on the laser performance was studied, see Fig. 3a. For 40% OC, the laser generated 10.65 W at 1029 nm with a slope efficiency η of 56% and a laser threshold of 1.05 W. With reducing the output coupling transmission, the laser performance progressively deteriorated: the output power and the slope efficiency decreased (down to 3.94 W and η = 27% for 10% OC). In addition, thermal rollover was observed for T OC = 10% and 20%. It appeared at smaller pump powers when reducing the output coupling probably due to higher intracavity laser intensity. The laser threshold monotonously increased with the output coupling, from 1.05 to 1.43 W. The laser spectra were weakly dependent on T OC and the emission was observed at 1028 nm, Fig. 3b. The laser linewidth was about 1 nm.
The pump absorption in the crystal was measured under non-lasing conditions at the threshold pump power assuming a single pass of the pump radiation. The calculation was done by taking into consideration the Fresnel losses at the uncoated crystal surface (assuming a refractive index n of 1.818). The pump absorption efficiency was about 72% at 140-200 K and below 140 K, it notably dropped (to 47% at 120 K and 18% at 100 K). This decrease in absorption was due to the blue-shift and narrowing of the absorption ZPL of Yb 3+ ions in the garnet crystals. A similar behavior was previously observed for the parent Yb:YAG and Yb:LuAG garnet crystals at the temperatures below 140 K [20,21]. Based on the determined pump absorption efficiency, we also plotted the output dependence of the laser with respect to the absorbed pump power, as shown in Fig. 4a. At 140 K, using 30% OC, the corresponding slope efficiency amounted to 71%. Figure 4b shows a typical far-field beam profile measured at the maximum output power corresponding to the fundamental transverse mode. A perfectly circular beam with a Gaussian intensity distribution was obtained confirming the excellent thermal management on the gain media by cryogenic cooling. Based on the CW laser results, the optimum temperature (140 K) and the output coupler transmission (40%) were chosen for further passively Q-switched laser experiments.

Passively Q-switched cryogenic Yb:LuYAG laser
Passive Q-switching of the cryogenic Yb:LuYAG laser was realized using AR-coated Cr 4+ :YAG SAs with two different initial transmissions at the laser wavelength T SA of 85% and 95%. The input-output dependences for the 1 3 57 Page 4 of 7  Fig. 5a. The average output power increased with the initial transmission of the SA. A maximum average output power of 8.28 W corresponding to a slope efficiency of 49% was achieved for the SA with T SA of 95%. For the SA with a higher modulation depth (T SA = 85%), the laser generated 6.05 W with a lower slope efficiency of 39%. The laser operated on the fundamental transverse mode, see the far-field beam profile depicted at the inset of Fig. 5a. The pulse duration Δτ (FWHM) and the pulse repetition rate (PRR) of the passively Q-switched Yb:LuYAG laser are shown in Fig. 5b. The PRR increased almost linearly with the incident pump power, from 44 to 105 kHz (T SA = 95%) and from 11 to 39.7 kHz (T SA = 85%.) Smaller PRR values for the SA with a higher modulation depth are due to higher pulse energies. The pulse duration decreased with the incident pump power just above the laser threshold, and then showed a weak dependence on the pump level. The shortest pulse durations were 201 ns and 346 ns, measured using the SAs with T SA = 85% and 95%, respectively. The corresponding oscilloscope traces of the shortest single Q-switched pulses are shown in Fig. 6a. The calculated pulse energy, E out = P out /PRR, and the peak power, P peak = E out /Δτ, are shown in Fig. 6b. The pulse energy   Table 1. The laser wavelength was centered at 1028.9 nm, Fig. 6c. Further shortening of Q-switched pulses is expected by increasing the modulation depth of Cr 4+ :YAG SAs or by shortening the cavity length to reduce the cavity roundtrip time [22].

Conclusion
To conclude, the ytterbium-doped "mixed" (yttrium-lutetium) aluminum garnet crystal is attractive for the development of high-power diode-pumped cryogenic lasers emitting at ~ 1030 nm. In the continuous-wave regime, pumping by a VBG-stabilized laser diode at 969 nm, a maximum output power of 10.65 W was achieved at 1029 nm with a slope efficiency of 56% with respect to the incident pump power, at the crystal temperature of 140 K. The optimum  operation temperature and the output performance of the "mixed" Yb:LuYAG crystal were similar to those found for Yb:LuAG. However, the former compound benefits from the simplified and more cost-efficient growth procedure. By employing Cr 4+ :YAG based saturable absorbers, for the passively Q-switched cryogenic laser, the best pulse characteristics (energy/duration) were 0.15 mJ/201 ns at a repetition rate of 39.7 kHz. Future work will focus on active Q-switching using KD*P or BBO based Pockels cells to increase the pulse energy. Other cavity designs will also be studied.