1 Introduction

In line with the trend to higher average output power, ytterbium lasers are becoming increasingly more popular worldwide. Yb:YAG is frequently used for lasers with high average and high peak power thanks to a favorable combination of laser (large emission cross-section and long lifetime) and material (high thermal conductivity) characteristics, as well as low cost [1,2,3,4,5,6,7,8,9,10]. Works on the development of pulse-periodic Yb:YAG lasers with high average and high peak power for scientific and commercial applications, for example, HiLASE [10], Genbu [6], DiPOLE [3] and others, are underway in different countries. However, Yb:YAG possesses some features that should be taken into consideration. One of the serious problems arising in disk laser amplifiers is that the energy stored in the active element (AE) is strongly limited by effects of the amplified spontaneous emission and parasitic generation caused by high transverse gain due to their aspect ratio. Up to date maximum energy extracted from one Yb:YAG thin disk AE with water cooling is ~ 100 mJ, therefore different kinds of cryogenic cooling are used to increase the output energy [1,2,3,4,5,6,7]. Unfortunately significant unwanted Yb:YAG emission cross-section narrowing and displacement of its central wavelength λpeak occur on cryogenic cooling. At a temperature of 80 K, the spectral width is comparable with the displacement of λpeak, hence, for a cryogenic Yb:YAG amplifier it is demanded either to artificially shift the generation wavelength of the Yb:YAG laser oscillator operating at room temperature, to cool the Yb:YAG crystal or to use some other special front end. Furthermore, for many applications, such as generation of radiation in different spectral ranges from X-ray to terahertz [11,12,13,14,15], efficiency of a nonlinear process can be increased at higher intensity due to shorter pulse duration. Different scientific groups published their results on ultrafast cryogenic Yb:YAG laser development [16,17,18,19,20,21], but reported output pulse transform limited duration was about 1–5 ps due to narrow gain bandwidth. One possible solution to these problems is to use media with emission cross-section smaller than that of Yb:YAG, so as to suppress the effect of amplified spontaneous emission, and with a broader emission spectrum at cryogenic temperatures supporting amplification of room temperature Yb:YAG lasers. The candidates meeting the above requirements are the broadband media Yb:KYW, Yb:KGW, and Yb:CaF2 widely used for operation with ultrashort pulses [22,23,24]. However, the emission cross-section of Yb:CaF2 is too small for efficient extraction of the energy stored in the AE; and the heat conductance of Yb:KYW and Yb:KGW, which is lower than that of Yb:YAG, limits average output power. An alternative is laser ceramics of rare-earth metal sesquioxides Yb:Y2O3, Yb:Lu2O3 and Yb:Sc2O3 the heat conductance of which is higher than that of Yb:YAG and the emission cross-section is broader [25,26,27]. The lower emission cross-section for these sesquioxides compared to Yb:YAG allows more efficient energy storage in the AE, and higher than in Yb:CaF2 enables more efficient extraction of the stored energy. The thermal conductivity that is higher than in Yb:KYW and Yb:KGW permits broadband radiation to be amplified at a higher average power. Also, the central wavelength and the amplification bandwidth of Yb:Y2O3 and Yb:Lu2O3 allow using a cryogenic amplifier together with a conventional front end based on water cooled Yb:YAG crystals. The technologies of these ceramics are being actively developed and modified but to date have only been used for constructing high-power disk oscillators or regenerative amplifiers [28,29,30,31,32]. We believe that the research aimed at improving the characteristics of disk multipass amplifiers using these media to create lasers with both high average and peak power is very timely.

The first section of this work concerns the investigation of the characteristics of Yb:Y2O3 ceramic samples, and the second one the development and construction of a high-power multipass cryogenic pulse-periodic amplifier with the AE made of Yb:Y2O3 ceramics.

2 Investigation of the Yb:Y2O3 ceramics characteristics

The AE for our cryogenic disk Yb:Y2O3 amplifier was made of a 1 mm-thick laser ceramic sample 15 mm in diameter, doped with 3.5 at.% ytterbium ions produced by Konoshima Chemical Co., Ltd., Japan. The quality of the material was assessed in a series of experiments on studying the laser and optical characteristics of the sample. The transmission spectrum of the ceramic sample (Fig. 1) was measured by a spectrophotometer (SF-256UVI, 190–1100 nm range, resolution 1 nm, LOMO, Russia). Losses far from the resonant absorption maxima coincided to a good accuracy with the Fresnel reflection at two uncoated faces (0.82) for the refractive index of 1.89 [33], which indicates that there was no evidence of high absorption or scattering in the sample. Higher precision measurements by the method described in the work [34] showed that the nonresonant absorption at a wavelength of ~ 1 µm was only 0.01 cm−1, which is a good result for laser ceramics, sufficient for operation at high average power.

Fig. 1
figure 1

Transmittance versus wavelength in a 1-mm-thick Yb:Y2O3 3.5 at.% doping ceramic sample

Full size image

With the use of the pinhole method, which allows measuring lifetime at the upper operating level in thick samples without error induced by reabsorption [35], we obtained a value of 0.82 ms at room temperature, which agrees well with the data of the works [25, 33]. In the next series of experiments, we placed the sample into a vacuum cryogenic chamber with a system of cooling by liquid nitrogen. At a temperature of 77 K, the reabsorption from the lower operating level may be neglected; therefore, the lifetime at the upper laser level is readily measured by the fluorescence decay with time. The lifetime was 1.12 ms, which is also in a good agreement with the result obtained in Ref. [25].

Further, the fluorescence spectrum was measured by a spectrometer (AvaSpec Dual, resolution 0.3 nm, range 180–1100 nm) at temperatures of 300 K and 77 K. A pump source at 940 nm was used for excitation of ytterbium ions. The obtained data and the Füchtbauer–Ladenburg formula [36] were used to calculate the wavelength dependence of the emission cross-section (Fig. 2). The calculated value of ~ 1.3 × 10−20 cm2 for the emission cross-section at 1030 nm agrees well with the data of the studies [25, 33, 37]. As was to be expected, on cooling to 77 K, the emission band centered at 1030 nm became narrower from ~ 12 nm to ~ 4.5 nm FWHM as in Ref. [25] and the peak value increased up to ~ 2.35 × 10−20 cm2.

Fig. 2
figure 2

Emission cross-section versus wavelength in a Yb:Y2O3 sample with 3.5 at.% doping at 300 K (solid curve) and 77 K (dashed curve)

Full size image

3 Constructing a Yb:Y2O3 disk ceramic amplifier

A multipass disk amplifier was developed on the basis of the active multipass scheme [38] the parameters of which are calculated by a known focal distance of the AE lens [39] specified by different phase distortions. We used a disk AE with a dielectric mirror on one face, and an antireflective coating on the other. When the AE was mounted mirror side towards heat sink and cooled to 77 K with liquid nitrogen, “cold” phase distortions arose due to the difference of the heat expansion coefficients of the AE and the heat sink. During operation at high average power, thermally induced phase distortions occurred in the AE, leading to the appearance of a thermal component of the lens.

The phase distortions were measured to a high accuracy by the method of phase-shifting interferometry [40]. To do so the AE in the vacuum cryogenic chamber was placed in one of the Michelson interferometer arms as a backward mirror (Fig. 3).

Fig. 3
figure 3

Schematic diagram of phase distortions measurement in a disk active element of a cryogenic laser head by the method of phase-shifting interferometry

Full size image

Examples of the obtained distribution of the optical path difference as a function of the transverse coordinate on reflection from the AE are presented in Fig. 4a for a pump diameter of 3 mm. The lens power in the AE per one reflection was calculated by the parabolic approximation in the pump area (Fig. 4a, b). The plots illustrate one of the merits of our mounting method, namely, the “cold” and the hot lenses have different signs and partially compensate each other. This allows choosing any needed operating power up to ~ 200 W.

Fig. 4
figure 4

a Optical path difference as a function of transverse coordinate in a cryogenic laser head based on a disk Yb:Y2O3 AE 15 mm in diameter at pump power 0 W (thin dash-and-dot curve), 150 W (thin solid curve) and 200 W (thin dashed curve), and parabolic approximation in the beam central area at pump power 0 W (thick dash-and-dot curve), 150 W (thick solid curve) and 200 W (thick dashed curve). b Lens power versus absorbed pump power

Full size image

The next step was measuring small signal gain per one reflection from the AE cooled to 77 K with liquid nitrogen using a milliwatt input signal, first at a low thermal load at a pump pulse duration of 3 ms and a repetition rate of 5 Hz, and then at a high heat release with cw pumping (Fig. 5). As expected, in a pulse-periodic mode, the small signal gain G grew linearly as a function of pump power and was restricted (at ~ 150 W) by the parasitic generation appearing in the transverse direction of the AE and observed with the oscilloscope. At cw pumping the emission cross-section decreased due to AE heating and operation at a higher pump peak power became possible without reaching the parasitic lasing threshold (G = 1.25).

Fig. 5
figure 5

Small signal gain in Yb:Y2O3 AE cooled with liquid nitrogen versus peak absorbed pump power at cw pumping (diamonds) and at a pulse repetition rate of 5 Hz (squares)

Full size image

For simplicity and convenience of adjustment, the simplest modification of an active multipass scheme [41] with one large spherical mirror for transferring the image from the AE to the backward mirror was used [39, 42]. Based on the data presented in Figs. 4b and 5 and the formulas presented in Ref. [39], we calculated the optical scheme for operation at a cw pump power of 200 W, when the thermal lens in the AE was not too large and the increase of the small signal gain as a function of pump power almost ceased. The distance from the AE to the large spherical mirror with curvature R = 1600 mm was 1500 mm, and the distance between the large spherical mirror and the backward mirror with curvature radius r = 800 mm was 1715 mm (Fig. 6).

Fig. 6
figure 6

Block diagram of the front end and schematic of the 9 V-pass Yb:Y2O3 cryogenic disk amplifier

Full size image

The multipass disk amplifier was tested on a specially designed front end (FE). The radiation from the home-made seed laser [43] with an average output power of 0.84 W and spectrally limited pulse duration of about 1.2 ps at a repetition rate of 11.5 kHz passed to a chirped volume Bragg grating (CVBG) (produced by “OptiGrate”, USA, coefficient of stretching in time per one reflection 220 ps/nm at spectral width 2.2 nm) stretcher, as a result of which the pulse duration increased to ~ 0.5 ns. After that the power was increased from 0.6 to 7 W by means of a 2-pass thin rod Yb:YAG amplifier [44] and was fed to the input of a disk cryogenic Yb:Y2O3 amplifier (Fig. 6).

For pumping the disk AE a scheme with four reflections from the AE and 3 mm spot diameter on the crystal was constructed, with the signal beam diameter on the crystal being 1.5 mm. The AE was continuously pumped by a module with a fiber output and a maximum power of 600 W at 940 nm, linewidth 2.5 nm (fabricated by LaserLine, Germany). A maximum output power of 15.8 W was attained as a result of nine reflections of the amplified radiation from the AE (Fig. 7a). Note that, because of strong changes in the AE lens (Fig. 4b), the output power was measured starting from a pump power of 50 W at which the diffraction losses associated with the difference in the parameters of the optical system and lens power in the AE significantly decreased. The spectra of the input and output signals were measured with the spectrometer (AvaSpec Dual, resolution 0.3 nm, range 180–1100 nm) (see Fig. 7b). Their full coincidence with a linewidth of ~ 1.2 nm FWHM points to a good agreement between the central wavelength of the Yb:YAG front end and the wavelength with maximum gain cross-section of Yb:Y2O3 cooled to cryogenic temperature, as well as to the possibility of amplification in this medium of subpicosecond pulses. Note that the amplification bandwidth in liquid nitrogen-cooled Yb:YAG active elements was limited to ~ 0.7 nm [45].

Fig. 7
figure 7

a Output power of a 9 V-pass disk cryogenic Yb:Y2O3 amplifier as a function of absorbed pump power, and spatial beam distribution at the exit at maximum output power. b Spectrum of the input (solid curve) and output (dashed curve) radiation of a disk cryogenic Yb:Y2O3 amplifier versus wavelength

Full size image

The amplifier output parameters were restricted by deterioration of the beam quality with an increase in the number of passes. A spatial beam distribution at the exit of the amplifier, right after a large spherical mirror at maximum output power is shown in the insert to Fig. 7a. One can see a ring appearing around the principal beam which we attribute to the difference of the phase distortion profile from the parabolic one, especially at a high pump power. To eliminate this problem, the quality of AE mounting on the heat sink should be improved. For more efficient extraction of the energy stored in the AE, it is intended to increase the output power of the front end. Works on the change-over to a broader-band front end with a pulse repetition rate of 15 kHz and pulse energy of ~ 0.5 mJ at < 300 fs duration are also planned.

4 Conclusion

We present a new concept of using Yb:Y2O3 instead of Yb:YAG in cryogenic booster amplifiers that enables enhancement of the efficiency and increasing one of the key parameters—peak power—due to a broader emission cross-section. Beyond that point systems with a cryogenic booster amplifier get an opportunity to use a low-cost front end based on Yb:YAG crystals operating at room temperature without loss in spectral width and the need to match the wavelength in cascades operating at room and cryogenic temperatures.

The first multipass disk pulse-periodic amplifier on the basis of a cryogenically cooled ceramic Yb:Y2O3 active element was constructed. The amplifier output power is 15.8 W at pulse repetition rate 11.5 kHz, pulse duration 0.5 ns and spectral width 1.2 nm. Given an appropriate source, broader-band radiation may be amplified with a further compression to the subpicosecond range of pulse durations. Currently high-quality Yb:Y2O3 laser ceramic disk samples a few centimeters in diameter are commercially available. Their aperture is comparable to that of Yb:YAG monocrystals and is quite sufficient to be implemented in Joule-level laser setups.