A diode-pumped picosecond 8-pass amplifier with a liquid-nitrogen-cooled Yb:YAG crystal has been developed. An average output power of 23.7 W with a near-diffraction-limited beam quality (M 2 < 1.2) was obtained at a pulse repetition rate of 80 kHz and a pulse duration of 11.7 ps. Average powers above 20 W were also obtained in the 30–80 kHz repetition rate range. The pulse energy reached almost 1 mJ at the 20 kHz repetition rate.
©2007 Optical Society of America
Compact picosecond lasers with high average power are in high demand for precision micromachining. Using a diode-pumped Nd:YVO4 or Nd:GdVO4 crystal, efficient picosecond regenerative amplifiers have been demonstrated with average power beyond the 10-W level and pulse energies of hundreds of microjoules [1,2]. By adding a booster amplifier behind a regenerative amplifier, an average output power of 22 W has also been demonstrated . On the other hand, using a diode-pumped Yb:YAG crystal with thin-disk configuration, a picosecond regenerative amplifier has been demonstrated with an average power of 10.2 W or a pulse energy of 4.5 mJ . These systems which generate the 4–20 ps range pulses are very attractive for achieving compactness and simplicity because they do not require any pulse stretching and compression.
Cryogenically-cooled Yb:YAG crystal is one of the promising laser materials for the high-energy and high-average-power picosecond lasers. At liquid nitrogen temperature, the Yb:YAG has a low saturation fluence of ~1.5 J/cm2 and a gain band width of ~1.5 nm due to the narrowing of the emission spectrum [5,6]. This saturation fluence is comparable to a typical surface damage threshold of optics for a 10-ps laser pulse. Hence efficient energy extraction from the laser crystal will be achieved without optical damage, when one uses multipass amplification of picosecond pulses. In addition, the thermal properties of the crystal such as thermal conductivity, thermo-optic coefficient (dn/dT), and thermal expansion coefficient are significantly improved at low temperatures [6–9], which is preferred especially for high-average-power operation. In fact, cryogenically cooled Yb:YAG lasers have been demonstrated with high output power up to 300 W and near-diffraction-limited beam quality in continuous wave (cw) and Q-switched nanosecond operation mode [6,10].
Recently, we have demonstrated a diode-pumped picosecond regenerative amplifier using a liquid-nitrogen-cooled Yb:YAG crystal . This amplifier generated 12.5-ps pulses with 3.4-mJ pulse energy at 500-Hz repetition rate. However, the average power was limited by thermally induced effects in the Pockels cell, even though the laser crystal had a further potential for high average power operation. In this work, we adopted a multipass amplification scheme to avoid the average power limitation by the Pockels cell. We have developed a diode-pumped picosecond 8-pass amplifier using a liquid-nitrogen-cooled Yb:YAG crystal. Over 20-W diffraction limited output were obtained in the 30–80 kHz repetition rate range by amplifying 10-nJ-level seed pulses. The pulse energy reached almost 1 mJ at the 20 kHz repetition rate.
2. Experimental setup
A schematic diagram of the 8-pass amplifier system is shown in Fig. 1. The seed source was a passively mode-locked Yb:YAG oscillator using a semiconductor saturable absorber mirror (SESAM). The gain medium was a 25 at.% Yb:YAG crystal with a 2 mm length, which was placed in a vacuum chamber and was cooled to 77 K with liquid nitrogen to match the emission wavelength of the oscillator to the amplifier. The crystal was end-pumped by a fiber-coupled cw laser diode with a wavelength of 940 nm and a power of 1.2 W. The oscillator had a cavity length of 7.1 m, and produced a 21.1 MHz pulse train of 10.7 ps pulses with a wavelength of about 1029 nm. The pulse picker that was consisted of an RTP Pockels cell and thin-film polarizers ejected seed pulses with an energy of about 10 nJ, a repetition rate up to 80 kHz and a pulse contrast ratio better than 2000:1. The beam passed through the mode-matching telescope and the Faraday isolator, and was then directed to the amplifier system.
The gain medium of the amplifier was a diode-end-pumped 5 at.% Yb:YAG rod with dimensions of 12 mm in diameter and 6.6 mm in length, with AR-coated faces. The laser rod was held by a liquid-nitrogen-cooled holder made of copper in a vacuum chamber. Indium sheets were used to produce a good thermal contact between the laser rod and the holder. More details of the holder are described in Ref. . The temperature of the holder, measured with a thermocouple, increased up to 86 K from 77 K in these experiments. The crystal and two AR-coated windows of the cryostat were set at a normal angle for both the pump beam and the amplified beam. A fiber-coupled laser diode was used as a cw pump source. The laser diode was cooled by water with a constant temperature of 25 °C, and was operated with a central wavelength of 938 nm at the maximum output power. The output fiber of the diode had a 600-μm core diameter and a 0.22 numerical aperture (NA). The pump beam was focused into the laser rod through the dichroic mirror (M5) with a 1.2 mm waist diameter. The maximum pump power was 90.6 W. 95% of the pump light was absorbed by the Yb:YAG rod.
The seed beam passed through the thin-film polarizer (TFP1) and was directed into the laser rod by the concave mirror with a focal length of 2 m (M2) and the flat mirrors (M1, M3, and M4). The 1/e 2 diameter of the beam at the laser rod was 1.5 mm. The beam passed through the laser rod and was then immediately reflected back by the convex mirror with a focal length of -5 m (M5), which placed at a position of about 30 mm from the laser rod, with a small angle of 0.5 degree, and passed again through the laser rod. The difference of the beam position between the direct beam and the reflected beam at the laser rod was estimated to be about 0.26 mm, which was small enough compared with the pump area. This ensured a good spatial overlap between the pump area and the amplified beams.
After the two passes through the laser rod, the beam returned to the M2. Referring to the inset of Fig. 1, which shows the beam positions for each pass on the M2, the optical path configuration was following. First, the seed beam was reflected on the lower right side of the M2. Second, the beam was reflected back by the M5 and was returned to the upper left side of the M2 after two passes through the laser rod. Then, the beam was returned back by the flat mirrors (M6 and M7) to the upper right side of the M2, and similarly passed twice through the laser rod. After four passes, the beam passed through the 45 degree Faraday rotator and the second RTP Pockels cell for the optical gating to prevent amplified spontaneous emission (ASE), and was then reflected back by the flat mirror (M9) to the same optical path. Finally, after eight passes, the amplified beam with s-polarization was reflected by TFP1 and ejected.
One key feature of the system geometry was that the concave mirror with the 2-m focal length (M2) constituted a confocal telescope for the amplified beam. Namely, the optical path lengths, between M2 and M5, between M2 and the middle of M6 and M7, between M2 and M9, were the same as the focal length of M2. Therefore, it was ensured that the beam sizes on the laser rod in each pass did not change even if the thermal lensing in the laser rod was changed. The M5 had a focal length of -5 m in order to compensate the weak thermal lensing. By this compensation, the change of the beam divergence due to the thermal lensing was reduced. As a result, large beam sizes to prevent laser-induced damage were ensured in the all optical path of the amplifier system.
3. Results and discussion
Single-pass loss and small-signal gain was estimated by measuring the laser power after the first two passes. The single-pass loss was estimated to be about 2.5%, which was mainly produced by residual transmissions of the reflective mirrors (< 0.5% per bounce) and residual reflections of the AR-coated windows and laser rod (< 0.25% per surface). Figure 2 shows the single-pass small-signal gain coefficient (g 0 l) as a function of incident pump power. The maximum gain coefficient was calculated to be 2.24. The gain coefficient showed a linear behavior at low powers and a tendency to saturate at higher powers. This saturation was attributed to heating of the laser crystal. The temperature distribution within the crystal was calculated by using a two-dimensional steady-state thermal model. The details of this calculation are described in Ref . The peak temperature at the center of the pumped laser crystal was estimated to be about 140 K at the maximum pump power. At 140 K, the emission cross section of Yb:YAG was decreased to about two-thirds of that around liquid nitrogen temperature . Such decrease of the emission cross section in the high temperature region caused the gain decrease.
Figure 3 shows the average output power in 8-pass amplification as a function of incident pump power for five different repetition rates between 20 and 80 kHz. A maximum output of 23.7 W was obtained at 80 kHz repetition rate, corresponding to an optical-optical efficiency of 26%. At a repetition rate of 20 kHz, the output power was decreased by 18% to 19.4 W. The average output powers correspond to pulse energies of 0.3 mJ at 80 kHz and 0.97 mJ at 20 kHz, respectively, as shown in Fig. 4. In order to avoid the damage of the optical components, the pulse energy was controlled not to exceed 1 mJ at the minimum repetition rate of 20 kHz.
At 20 kHz, the gain of the amplifier reached almost 105. The fluence of the amplified pulses at the laser crystal was estimated to be about 0.1 J/cm2, which was more than one order of magnitude lower than the saturation fluence, ~1.5 J/cm2, of the laser crystal at liquid-nitrogen temperature. Therefore, gain saturation was not significant during an 8-pass amplification. According to a calculation using the Frantz-Nodvik model , the gain coefficient (gl) decreased by only about 5% before and after the amplification. Consequently, the saturated gain can be simply estimated to be G ≈ 4 (from G 8 ≈ 105) for all passes. It suggests that the most of the output energy was extracted at the final path.
An output pulse of the amplifier system was detected with a fast photodiode. The amplifier generated a clean output pulse without any noticeable pre- or post-pulses. The measured pre- and post-pulse contrast ratio was better than 105:1. The two RTP Pockels cells used in the system provided sufficient contrast. Figure 5 shows an autocorrelation trace of the output pulses at 80 kHz repetition rate. Assuming a sech2 pulse shape, the pulse width at half-maximum was estimated to be 11.7 ps. The pulse duration slightly lengthened because of gain narrowing in the amplifier.
After focusing the output beam with a convex lens, the beam quality factor M 2 was determined by measuring the beam waist radius and the beam divergence by using a CCD camera, to be less than 1.2 at the maximum output. An image of the output beam measured by a CCD camera is shown in Fig. 6. The output beam profile fits well to a Gaussian. No adjustments of any optical components were performed for compensation of the thermal lensing, when pump power was increased to the maximum. Accordingly, the thermo-optic distortion in the laser crystal was very small in this experiment.
In this system, the average laser power that passed through the Pockels cells was low enough, for example, was of the order of 100 mW at the maximum output. Therefore, the increase in average power was not limited by thermally induced effects in the Pockels cell. The average power limitation was only caused by thermally induced effects in gain media. As described above, the temperature in the crystal was estimated to have increased considerably. It implies that higher average-power cannot be achieved in this system only by increasing the pump intensity. However, by using an additional pump beam at another end of the laser rod and by using a lower-doped crystal with a longer length, it should be possible to achieve a 100-W average-power.
We have demonstrated a high-average-power multipass-amplifier for picosecond pulses with a cryogenically-cooled Yb:YAG crystal. The amplifier produced high pulse energies of 0.3–0.97 mJ at repetition rate in the range of 80–20 kHz. The maximum average power of 23.7 W at 80 kHz repetition rate was obtained with an M 2-factor less than 1.2. This is the highest average power, to the best of our knowledge, obtained by single-stage diode-pumped ultrafast amplifiers with pulse energies above multi-hundreds of microjoules. Further power scaling should be possible by using an additional pump beam at another end of the laser rod, which will also allow improved energy extraction.
References and links
1. M. Siebold, M. Hornung, J. Hein, G. Paunescu, R. Sauerbrey, T. Bergmann, and G. Hollemann, “A high-average-power diode-pumped Nd:YVO4 regenerative laser amplifier for picosecond-pulses,” Appl. Phys. B 78,287–290 (2004). [CrossRef]
2. J. Kleinbauer, R. Knappe, and R. Wallenstein, “13-W picosecond Nd:GdVO4 regenerative amplifier with 200-kHz repetition rate,” Appl. Phys. B 81,163–166 (2005). [CrossRef]
3. M. Hornung, M. Siebold, J. Hein, R. Sauerbrey, and G. Hollemann, “Diode-pumped amplification of 7-picosecond-pulses with repetition-rates up to 100 kHz and 22 W average output power, Conference Digest of CLEO/QELS Europe 2005, CA4-4, Munchen, Germany, June 12–17, 2005.
4. D. Müller, S. Erhard, and A. Giesen, “High pwer thin disk Yb:YAG regenerative amplifier,” OSA Trends in Optics and Photonics Series Vol. 50, Advanced Solid-State Lasers (Optical Society of America, 2001),319–324.
5. J. Dong, M. Bass, Y. Mao, P. Deng, and F. Gan, “Dependence of the Yb3+ emission cross section and lifetime on temperature and concentration in yttrium aluminum garnet,” J. Opt. Soc. Am. B 20,1975–1979 (2003). [CrossRef]
6. D. J. Ripin, J. R. Ochoa, R. L. Aggarwal, and T. Y. Fan, “300-W cryogenically cooled Yb:YAG laser,” IEEE J. Quantum Electron. 41,1274–1277 (2005). [CrossRef]
7. R. L. Aggarwal, D. J. Ripin, J. R. Ochoa, and T. Y. Fan, “Measurement of thermo-optic properties of Y3Al5O12, Lu3Al5O12, YAIO3, LiYF4, LiLuF4, BaY2F8, KGd(WO4)2, and KY(WO4)2 laser crystals in the 80–300 K temperature range,” J. Appl. Phys. 98,103514 (2005). [CrossRef]
8. G. A. Slack and D. W. Oliver, “Thermal conductivity of garnets and phonon scattering by rare-earth ions,” Phys. Rev. B 4,592–609 (1971). [CrossRef]
9. R. Wynne, J. L. Daneu, and T. Y. Fan, “Thermal coefficients of the expansion and refractive index in YAG,” Appl. Opt. 38,3282–3284 (1999). [CrossRef]
10. S. Tokita, J. Kawanaka, M. Fujita, T. Kawashima, and Y. Izawa, “Efficient high-average-power operation of Q-switched cryogenic Yb:YAG laser oscillator,” Jpn. J. Appl. Phys. 44,L1529–L1531 (2005). [CrossRef]
11. S. Tokita, J. Kawanaka, M Fujita, T. Kawashima, and Y. Izawa, “High-energy picosecond regenerative amplifier with cryogenically cooled Yb:YAG,” Technical Digest of CLEO/QELS 2006, CWN4, Long Beach, California, USA, May 21–26, 2006
12. S. Tokita, J. Kawanaka, M. Fujita, T. Kawashima, and Y. Izawa, “Sapphire-conductive end-cooling of high power cryogenic Yb:YAG lasers,” Appl. Phys. B 80,635–638 (2005). [CrossRef]
13. L. M. Frantz and J. S. Nodvik, “Theory of pulse propagation in a laser amplifier,” J. Appl. Phys. 34,2346–2349 (1963). [CrossRef]
14. D. C. Brown, “The Promise of Cryogenic Solid-State Lasers,” IEEE J. Sel. Top. Quantum Electron. 11,587–599 (2005). [CrossRef]