We report a 1.02 W coherent light at 193 nm and a 2.15 W coherent light at 221 nm generated by sum-frequency generation in CLBO crystals. The conversion efficiency from 221 nm to 193 nm was about 47%. To the best of our knowledge, this is the highest average power of 193 nm and 221 nm coherent light generated by SFG, respectively. These two DUV coherent lights could be used in many applications such as laser machining for their high photon energy features as well as the high power. The 193 nm coherent light also meets the requirements of injection-seeding hybrid ArF excimer laser.
© 2017 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Deep ultraviolet (DUV) coherent light with high spatial coherence and high power is currently attracting significant interest in various applications including spectroscopy, metrology, and laser machining. Particularly, due to its high photon energy feature, it is more effective that to do laser machining of certain materials such as carbon fiber reinforeced plastics (CFRP) by use of a DUV coherent light if it also has a high spatial coherence. Generally, the most prevalent way to obtain the high power and high coherence DUV light is by frequency conversions of the solid-state lasers. Moreover, to obtain higher power such as tens of watt or even hundreds of watt, an excimer laser seeded by a solid-state DUV coherent light will be a suitable solution, maintaining the high power and high spatial coherence simultaneously. Recently, a hybrid ArF excimer laser at 193 nm has been demonstrated with average output power of more than 100 W . This hybrid ArF laser consisted of a solid-state 193 nm coherent light for injection-seeding and a conventional ArF laser amplifier with very high gain. Thus, it is important to develop a high power DUV laser not only for applying itself to the laser machining application but also for seeding the excimer laser which could accomplish a 193 nm coherent light with hundreds of watt output power and high spatial coherence by the aforementioned hybrid ArF laser schematics.
In the past decades, the generation of the solid-state 193 nm coherent light is a challenging but thriving topic, on which many efforts have been made by different research groups as shown in Fig. 1 [2–9]. For the continuous-wave (CW) solid-state 193 nm coherent light, M. Scholz et al. reported a CW 193 nm light with 15 mW average power output directly frequency doubled from 386 nm by use of a KBe2BO3F2 (KBBF) crystal inside an enhancement cavity . J. Sakuma et al. had obtained a 193.4 nm coherent light with 120 mW average power by sum-freqeuncy generation (SFG) in a CsLiB6O10 (CLBO) crystal, also inside an enhancement cavity, which is the highest average power of the CW solid-state 193 nm coherent light so far . On the other hand, for the pulsed solid-state 193 nm coherent light, T. Kanai et al. reported a picoseconds 193.5 nm coherent light with 1.05 W average power output by second-harmonic generation (SHG) in a KBBF crystal , which is the highest solid-state 193 nm power generated directly by SHG until now and is also the highest power of the picoseconds 193 nm coherent light. S. Ito et al. obtained a 0.2 W average power at 193 nm also from a KBBF crystal by SHG with the repetition rate of 6 kHz and nanoseconds pulse duration . The disadvantage of this method was that it started from a Ti:sapphire laser and thus it could not be operated in a high repetition rate (>6 kHz) with a good beam quality simultaneously and easily. In terms of SFG between a pulsed near-infrared (NIR) wavelength laser and a pulsed DUV laser at the wavelength longer than 193 nm, H. Kawai et al. reported a 140 mW average power coherent light at 193 nm by SFG in a CLBO crystal with a pulse duration of 1 ns . N. Umemura et al. obtained a 193 nm coherent light with 200 mW output power by SFG in a K2Al2B2O7 (KABO) crystal . Recently, P. Koch et al. reported their results of a DUV coherent light at 191.7 nm by SFG between 1342 nm and 224 nm in the CLBO crystal with the output power of 240 mW and a high conversion efficiency of 49% from 224 nm to 191.7 nm . 300 mW average power of 193 nm coherent light was obtained by SFG between a DUV laser at 258 nm, which was the fourth harmonics of a 1030 nm Yb:YAG laser, and an NIR Er-doped fiber laser at 1553 nm .
The current hybrid ArF laser seed of the 300 mW solid-state 193 nm coherent light was still not high enough to realize the aforementioned features after the excimer amplifier due to the huge amplified spontaneous emission (ASE) ratio of the hybrid ArF amplifier, which stimulates the demand of increasing the average power of the solid-state 193 nm coherent light to 1 W level according to the conventional ArF laser system. Moreover, a solid-state 193 nm coherent light with higher power will also improve the possibility of applying itself directly to the laser machining even without an excimer amplifier. However, there is still no report on a 1 W solid-state 193 nm coherent light source by SFG until now. Firstly, it is not easy to obtain a high power, high beam quality DUV laser at the wavelength longer than 193 nm by fourth-harmonic generation (FHG) from a 1 µm laser for SFG with another NIR laser. Moreover, parameter changes of the generated high power DUV laser, including the beam size and the beam quality, will affect the conversion efficiency to 193 nm. In the previous result of the 300 mW 193 nm coherent light, the DUV laser at 258 nm was no more than 3 W  and there were no parameter changes during the 258 nm laser generation. However, when the 258 nm laser was increased from 1–3 W to 10 W, its beam quality became worse due to the nonlinear effects , which will induce the wave-front change of the DUV laser, lead to its wave-front mismatching with the NIR laser beam and decrease the conversion efficiency of SFG. At the same time, the DUV laser beam size will also change because of the DUV laser absorption of the nonlinear optical crystal during its generation, which increases the difficulty of achieving an optimizing focused power intensity for efficient SFG processes. In terms of the focused power intensity, the high power DUV coherent lights (such as 258 nm, 221 nm or 193 nm) will theoretically and experimentally have the transmission optics damaged during their generation, which will also make more troubles to achieve the appropriate power intensity as well as the conversion efficiency because the transmission optics such as lenses could not be utilized in the experiments due to the ultraviolet(UV)-induced damage. On the other hand, the power scaling of the NIR laser at 1553 nm is still limited by the lack of the gain medium . From the picture of photon, the conversion efficiency should be affected by the NIR photon depletion if the photon flux of the NIR laser is not high enough comparing to that of the DUV laser. In short, there is a gap between achieving a 193 nm coherent light with average power of hundreds of mW and achieving that with 1 W, which is hard to be crossed.
In this contribution, we report our unique result of a DUV coherent light at 193 nm with a recorded 1.02 W average power achieved by SFG in the CLBO crystals. To the best of our knowledge, this is the highest average power for solid-state 193 nm coherent light by SFG in nonlinear crystals so far. Meanwhile, 2.15 W average power of DUV cohernet light at 221 nm was generated, which is also the highest average power achieved by solid-state laser at this wavelength until now. In order to avoid the damage of the optics induced by the DUV coherent lights, an innovative configuration of the SFG process by concave mirrors was constructed to achieve the final target of the high power 193 nm coherent light.
2. Experimental setup
The schematics diagram of the solid-state 193 nm coherent light is shown in Fig. 2, which started from a DUV laser at the wavelength of 258 nm and a NIR laser at the wavelength of 1553 nm. The repetition rate of the final 193 nm light was 10 kHz in order to match that of a 10 kHz hybrid ArF laser system, which is under developing now. The 258 nm laser was generated from a 35 W Yb:YAG laser at 1030 nm by FHG . The NIR laser at 1553 nm was from an Er-doped fiber master oscillator amplifier (MOPA) . The 221 nm coherent light was generated by SFG between the 1553 nm and the 258 nm lasers. Then, the 193 nm coherent light was finally generated by this 221 nm coherent light and the residual of the 1553 nm laser. The synchronization of the DUV and NIR laser pulses for SFG was achieved by a digtial delay/pulse generator (DG645, Standford Research Systems) [2, 13].
The average power of the 258 nm laser was 10 W  while the power of the 1553 nm laser was 1.7 W. Figure 3 depicts the experimental setup of the two stages of SFG process in detail. The 1st stage SFG occurred between the 258 nm and the 1553 nm laser. The beam size of the 258 nm laser was firstly enlarged by use of two concave mirrors (CM1, CM2) in order to prevent the damage of the optics induced by the DUV laser at 258 nm. The distance between the two concave mirrors led to a 2:3 telescope system. On the other hand, the beam size of the 1553 nm laser was focused by a lens, which made the beam diameter to be almost the same size as that of the 258 nm laser at the 1st SFG stage. The generated coherent light at 221 nm was separated by a dichroic mirror DM2 (HR@221 nm/HT@1553 nm), which was also high transmission for 258 nm. Meanwhile, the residual of the 258 nm and the 1553 nm laser was separated by DM4 (HR@258 nm/HT@1553 nm). The residual of the 1553 nm laser was used for the next stage SFG while the 258 nm was reflected and dumped. The non-linear optical crystal for the 1st stage SFG was a piece of type-I CLBO crystal (5×5×20 mm3) with the phase-matching (PM) angle of 51.9◦ [14, 15]. For the second stage SFG, the frequency conversion crystal was also a type-I CLBO crystal (5×5×20 mm3) but with the PM angle of 61.6◦ [14, 15]. Each CLBO crystal was packaged, respectively, in a cell with the Ar gas purging and heated to more than 150◦C by an oven to avoid the hydration problem .
3. Results and discussion
As shown in Fig. 4(a), the highest output power of 221 nm coherent light was 2.15 W when the power of the 258 nm laser was 9.44 W, corresponding to the conversion efficiency of 22.8% from 258 nm to 221 nm. However, the highest conversion efficiency of 37.7% was obtained at a lower 258 nm laser power of 1.2 W. With the 258 nm laser power increasing from 1.2 W to 9.44 W, the conversion efficiency decreased from 37.7% to 22.8%. One possible reason of this efficiency drop is the beam size change of the 258 nm laser for SFG as shown in Fig. 4(b). These beam sizes were measured after the dichroic mirrors of DM5 and DM6 by a NanoScan (NS2s-Pyro/9/5, Ophir Inc.) at different 258 nm laser powers.
With the power increasing from 1 W to 9 W, the beam diameter of the 258 nm laser changed from about 1 mm to about 0.9 mm, which was firstly inevitable because of the DUV laser absorption of the CLBO crystal of the 258 nm laser generation. Moreover, the UV laser-induced heating issue of the reflection and concave mirrors will also contribute to the DUV beam size change at the input surface of the 1st CLBO crystal for SFG. However, experimentally, the optimization of the DUV laser power intenstiy was done by assuming that there was no change of the beam sizes. Hence, the power intensity at 10 W was higher than the predicted value because of the smaller focused beam size, which would lead to a saturation and thus decrease the conversion efficiency. Moreover, the worse beam quality at a higher DUV laser power due to the nonlinear effects, such as the Kerr lens and the thermal lens effect, inside the CLBO crystal for the 258 nm laser generation , would also decrease the conversion efficiency. This would induce the mismatching of the wave-front with the 1553 nm laser. Furthermore, the beam size of the 258 nm laser changed to be smaller from 1 to 10 W before the 1st stage SFG CLBO crystal while the beam of the 1553 nm laser before the 1st CLBO crystal maintained the same size to that of the 1 W DUV laser. Consequently, when the 258 nm laser power was increased from 1 W to 10 W, the beam overlap between the 258 nm and 1553 nm laser became worse, which also induced the decreasing of the SFG conversion efficiency from 258 nm to 221 nm. In addition, the beam size changing at the 1st CLBO was larger due to the above effects.
On the other hand, from the point of view of photon, the estimated photon flux of the NIR laser at 1553 nm seems to be not high enough comparing to that in , in which the photon flux of the NIR was higher than that of the DUV laser and more than 40% of conversion efficiency was obtained at highest power. Thus, photon depletion of the NIR laser may be another reason for the conversion efficiency decreasing.
By optimistic estimation, at least 40% conversion efficiency from 258 nm to 221 nm due to the results in , corresponding to the 221 nm coherent light power of 4 W, could be obtained at an optimizing DUV power intensity at 10 W by compensating the DUV beam size and quality changing as well as the overlapping of the two beams. Furthermore, increasing the average power of NIR laser at 1553 nm will also be beneficial to increase the conversion efficiency of this stage SFG and thus to obtain a higher power 221 nm coherent light.
After the 1st SFG stage, the 221 nm coherent light and the residual 1553 nm laser with 0.9 W generated the final 193 nm coherent light. As shown in Fig. 3, a CaF2 prism with no coating firstly separated the generated 193 nm coherent light and then the average power was measured by a power meter. The transmission of the CaF2 prism was 92% for the 193 nm coherent light. The highest average power of the 193 nm coherent light measured by the power meter was 0.936 W at the highest 221 nm coherent light power of 2.15 W. Taking the transmission of the CaF2 prism into account, the highest average power at 193 nm was 1.02 W, corresponding to the pulse energy of 0.1 mJ, and the conversion efficiency was 47% from 221 nm to 193 nm. Figure 5(a) demonstrates the output power of the 193 nm coherent light when the 221 nm coherent light power changed as well as the conversion efficiency from 221 nm to 193 nm. There was almost no change of the conversion efficiency when the 221 nm coherent light changed from low to high power according to Fig. 5(a). It would be due to the low power operation where the heating issue did not influence on the beam changes. Figure 5(b) demonstrates the power stability of the generated 193 nm coherent light. It was recorded in half an hour at the average output power of 0.5 W in free running state with no feedback controlling system. It implied that the 193 nm coherent light could be stably operated for a long time in a free running state and the standard deviation of power is about 20.94 mW, corresponding to a 4% fluctuation, which will be beneficial for seeding the ArF excimer amplifier.
During the two SFG processes, no damage occurred on both CLBO crystals in the long operation time. Particularly, for the 1st CLBO crystal, it could even bear the high power 258 nm laser and the generated high power 221 nm coherent light at the same time. Moreover, we found that the damage threshold of the CLBO crystals were even higher than that of the multi-layer coating of the windows and the mirrors.
The 258 nm laser had a beam quality of M2<1.5, which was the best for the power higher than 10 W at this wavelength region so far . The M2 of the 1553 nm laser was and for x-axis and y-axis, respectively. Figure 6(a) and 6(b) shows the temporal pulse shape of the 1553 nm laser and the beam profile of it taken by a BeamGage (Ophir Inc.), respectively. Each of these two lasers had a good beam quality. Therefore, it led the 221 nm and 193 nm coherent lights to have good beam profiles, which is shown in Fig. 7(a) and Fig. 7(b) (measured by the NanoScan), respectively, and indicates the high spatial coherence features for both DUV coherent lights. Consequently, it also shows the possibility of achieving a good laser-machining result.
The pulse duration of the fundamental laser at 1030 nm was about 10 ns [2, 13]. It changed to about 3 ns after FHG to 258 nm . The pulse duration of the 1553 nm laser was about 5 ns as shown in Fig. 6(a). Thus, after the SFG processes, the pulse duration was also about 3 ns for 221 nm and 193 nm coherent light as shown in Fig. 7(c) and Fig. 7(d), respectively. Moreover, the temporal profile difference between the DUV and NIR laser pulses could be another reason of the conversion efficiency change in the SFG process. Furthermore, the estimated linewidth of the 193 nm coherent light was about 4 GHz since the linewidth of the 1030 nm and 1553 nm was 1.5 GHz and 200–300 MHz, respectively [2,13].
In summary, we reported a narrow-linewidth, nanoseconds DUV coherent light generation at 193 nm by SFG in the CLBO crystals with average power of 1.02 W. During the SFG processes, average power of 2.15 W at 221 nm was also obtained, which is the highest power coherent light at this wavelength generated by solid-state laser so far. The conversion efficiency from 221 nm to 193 nm was 47%. To the best of our knowledge, 1.02 W was the highest average power of the solid-state 193 nm coherent light obtained by SFG until now as well as the conversion efficiency of 47% from 221 nm to 193 nm. The configuration of SFG consisted of a telescope system by two concave mirrors to set the appropriate beam size of the DUV laser at 258 nm in order to avoid the optics damage induced by the DUV laser and also consequently to ensure the long term operation of the 193 nm coherent light. The 1.02 W DUV coherent light at 193 nm with high coherence and high stability could be a good seed for the hybrid ArF laser, leading to accomplish a multi-hundreds-of-watt 193 nm coherent light with lower ASE. Above all, the high power and good beam quality DUV coherent lights at 193 nm and 221 nm also implied the potential in the application of laser micro-machining with very high precision by use of their high photon energy characteristics as well as the hybrid laser. In the future, by increasing the laser power at 258 nm and the conversion efficiency from 258 nm to 221 nm via optimizing the telescope ratio, the temporal profile, and the photon flux of the 1553 nm laser, multi-watt level DUV laser at 193 nm is expected as well as the achievement of a hundreds-of-watt average power hybrid ArF laser with a high spatial coherence.
New Energy and Industrial Technology Development Organization (NEDO).
References and links
1. S. Tanaka, M. Arakawa, A. Fuchimukai, Y. Sasaki, T. Onose, Y. Kamba, H. Igarashi, C. Qu, M. Tamiya, H. Oizumi, S. Ito, K. Kakizaki, H. Xuan, Z. Zhao, Y. Kobayashi, and H. Mizoguchi, “Development of high coherence high power 193 nm laser,” Proc. SPIE 9726, 972624 (2016). [CrossRef]
2. H. Xuan, Z. Zhao, H. Igarashi, S. Ito, K. Kakizaki, and Y. Kobayashi, “300-mW narrow-linewidth deep-ultraviolet light generation at 193 nm by frequency mixing between Yb-hybrid and Er-fiber lasers,” Opt. Express 23, 10564–10572 (2015). [CrossRef] [PubMed]
3. M. Scholz, D. Opalevs, P. Leisching, W. Kaenders, G. Wang, X. Wang, R. Li, and C. Chen, “A bright continuous-wave laser source at 193 nm,” Appl. Phys. Lett. , 103, 051114 (2013). [CrossRef]
6. S. Ito, T. Onose, S. Watanabe, T. Kanai, K. Kakizaki, T. Matsunaga, C. Chen, Y. Kobayashi, C. Zhou, J. Fujimoto, H. Mizoguchi, and X. Y. Wang, “A Sub-Watt, Line-narrowing, 193-nm Solid State Laser Operating at 6 kHz with KBBF for Injection-locking ArF excimer laser systems,” in Lasers, Sources, and Related Photonic Devices, OSA Technical Digest (CD) (Optical Society of America, 2012), paper AT4A.7. [CrossRef]
7. H. Kawai, A. Tokuhisa, M. Doi, S. Miwa, H. Matsuura, H. Kitano, and S. Owa, “UV light source using fiber amplifier and nonlinear wavelength conversion,” in Conference on Lasers and Electro-Optics/Quantum Electronics and Laser Science Conference, Technical Digest (Optical Society of America, 2003), paper CTuT4.
8. N. Umemura, M. Ando, K. Suzuki, E. Takaoka, K. Kato, Z. Hu, M. Yoshimura, Y. Mori, and T. Sasaki,“200-mW-average power ultraviolet generation at 0.193 µm in K2Al2B2O7,” App. Opt. 42 (15), 2716–2719 (2003) [CrossRef]
9. H. Zhang, G. Wang, L. Guo, A. Geng, Y. Bo, D. Cui, Z. Xu, R. Li, Y. Zhu, X. Wang, and C. Chen, “175 to 210 nm widely tunable deep-ultraviolet light generation based on KBBF crystal,” Appl. Phys. B 93, 323–326 (2008). [CrossRef]
10. P. Koch, J. Bartschke, and J. A. L’huillier, “All solid-state 191.7 nm deep-UV light source by seventh harmonic generation of an 888 nm pumped, Q-switched 1342 nm Nd: YVO4 laser with excellent beam quality,” Opt. Express 22, 13648–13658 (2014). [CrossRef] [PubMed]
11. H. Xuan, C. Qu, S. Ito, and Y. Kobayashi,“High power, and high conversion efficiency deep ultraviolet (DUV) laser at 258 nm generation in the CsLiB6O10 (CLBO) crystal with a beam quality of M2<1.5,” Opt. Lett. 42, 3133–3136 (2017). [CrossRef] [PubMed]
12. Z. Zhao, H. Xuan, H. Igarashi, S. Ito, K. Kakizaki, and Y. Kobayashi, “Single frequency, 5 ns, 200 µJ, 1553 nm fiber laser using silica based Er-doped fiber,” Opt. Express 23, 29764–29771 (2015). [CrossRef] [PubMed]
13. H. Xuan, Z. Zhao, H. Igarashi, S. Ito, K. Kakizaki, and Y. Kobayashi, “Development of narrow-linewidth Yb- and Er-fiber lasers and frequency mixing for ArF excimer laser seeding,” Proc. SPIE 8961, 89612M (2014). [CrossRef]
14. A. V. Smith, “SNLO nonlinear optics code,” AS-Photonics, http://www.as-photonics.com/products/snlo.
15. N. Umemura, K. Yoshida, T. Kamimura, Y. Mori, T. Sasaki, and K. Kato, “New data on the phase-matching properties of CsLiB6O10,” in Advanced Solid State Lasers, M. Fejer, H. Injeyan, and U. Keller, eds., Vol. 26 of OSA Trends in Optics and Photonics (Optical Society of America, 1999), paper PD15.