We demonstrate the improvement and formation of UV-induced damage on LBO crystal output surface during long-term (130 h) high-power (20 W) high-repetition-rate (80 kHz) third-harmonic generation. The output surface was super-polished (RMS surface roughness <0.6 nm) to sub-nanometer scale super smooth roughness. The surface lifetime has been improved more than 20-fold compared with the as-polished ones (RMS surface roughness 4.0~8.0 nm). The damage could be attributed to the consequence of thermal effects resulted from impurity absorptions. Simultaneously, it was verified that the impurities originated in part from the UV-induced deposition.
© 2013 OSA
Q-switched 355 nm all-solid-state lasers are rapidly becoming the preferred light sources for widespread industrial applications due to their high efficiency and excellent beam quality . A straightforward and efficient way of obtaining high-power 355 nm UV laser is by nonlinear optical (NLO) conversion of the all-solid-state laser emitting at the infrared (IR) 1.06 μm wavelength. The lithium tri-borate (LiB3O5, LBO) is the most outstanding and widely used NLO crystal for its relatively large acceptance and negligible walk-off angle, as well as effective nonlinear coefficient compared with other materials such as CBO, CLBO, and BBO [2,3]. Based on the LBO crystal, a number of high-power high-beam-quality UV lasers have been reported [1,3,4].
Although the laser induced damage threshold (LIDT) value of the LBO crystal for nanosecond UV laser reaches up to 21.9 GW/cm2 (355 nm, 0.9 ns)  or 45 J/cm2 (355 nm, 5 ns)  (much higher for IR and green laser), the UV-induced damage on LBO output surface during long-term third-harmonic generation (THG) is still a bottleneck for the development and application of high-power UV lasers. This damage process, appearing even when the UV average- and peak-power densities are far smaller than the LIDT and thus distinguishing from the direct light-induced opto-mechanical breakdown, i.e., damages appearing immediately after exposure [5,6], implicates a new mechanism. Möller et al. adopted a Q-switched Nd:YVO4 laser and its second harmonic wave (average-power P1064 = 1.6 W, P532 = 1.0 W, repetition rate frep = 20 kHz, pulse duration τ1064 = 10 ns, τ532 = 6 ns) to generate a 15 mW 355 nm UV laser through frequency summing using LBO crystal with surface roughness of 10 nm as the THG module and observed this kind of damage through 192-hour long-term operating process . The damage mechanism under this special low UV power was assumed as the UV-induced deposition and UV-assisted ablation. However, the long-term (>100 h) UV-induced damage feature in the high-power (average-power ≥ 20 W) high-repetition-rate (≥ 80 kHz) UV laser which is the actual industrial demand, has rarely been reported.
In this paper, we utilized a 97.0 W high-power high-beam-quality Q-switched 1064 nm fundamental laser with pulse width of 29 ns at 80 kHz to generate high-power 355 nm third-harmonic wave. The obtained green and UV output powers at 80 kHz were 49.7 W and 20.2 W with the corresponding pulse durations of 27 ns and 26 ns respectively. In order to investigate constraints of lifetime, LBO crystals with different output surface parameters were applied as the THG crystal. Here, the lifetime was defined as the operating duration from the initiation of illumination to the appearance of visible damage spot. Through super-polishing techniques, the uncoated surface lifetime was improved more than 20-fold without damage compared with the as-polished as well as the super-polished but antireflection (AR) film-coated ones. The surface damage decreased the output power and degraded the beam quality of the UV laser. Consequently, the damage formation mechanism is of wide and cardinal significance for the study and application of high-power UV lasers. After 130-hour long-term exposure, a visible damage spot on the uncoated super-polished output surface was observed. Furthermore, comparisons of chemical composition of the UV-affected and unaffected areas were conducted. Then conclusions on the possible damage mechanism were given.
2. Experimental setup
As shown in Fig. 1 , the fundamental wave, passing through a half-wavelength plate (HWP), was shaped by a convex lens with a focal length of 150 mm into the second-harmonic generation (SHG) module and THG unit. To avoid the chromatic aberration of the IR and green laser when passing through a focus lens simultaneously, we didn’t use another lens to focalize the IR and green lasers. In our experiment, the THG crystal was positioned at around the focal point of the 150 mm lens. The output power of the system was measured using an OPHIR NOVAII Thermopile 30(150)A-SV power meter. A Spiricon M2-200s laser beam analyzer was used to survey the beam quality. The pulse signal was probed with a high-speed silicon photoelectric detector as well as a YOKOGAWA DL7200 oscilloscope.
A type I noncritical phase-matching (NCPM) LBO crystal, cutting at θ = 90° and ϕ = 0° with dimensions of 5 × 5 × 25 mm3, was utilized as the SHG module. The crystal was placed in a high-precision temperature-controlled stove and kept at 422.95 ± 0.02 K.
A type II phase-matching LBO crystal, cutting at θ = 44.6° and ϕ = 90° with dimensions of 5 × 5 × 20 mm3, was employed as the THG module. The temperature of the crystal was kept at 333.15 ± 0.05 K. The incident surface of the THG LBO crystal was polished and coated with AR film at 1064 nm, 532 nm, and 355 nm. Samples with different output surface parameters were utilized for comparison.
The as-polished samples (provided by Castech Crystals Inc.) processed with the conventional-polishing method were with surface roughness of 4.0~8.0 nm. Super-polishing which can effectively remove cracks, reduce the embedded polishing remnants, and thus ameliorate the surface roughness, has been reported to enhance the LIDT of fused silica . However, the effectiveness of its influence on the long-term resistance of intense UV exposure is barely discussed. In this paper, the chemical mechanical polishing (CMP) method with the ultra-dispersed nano-scale diamond slurry was employed to precisely polish the output surfaces of LBO samples. Then the surfaces were well cleaned. The roughness of the super-polished surfaces was detected by an atom force microscopy (AFM, NTEGRA, NT-MDT Corp.) with a silicon tip and reached to 0.4~0.6 nm (RMS).
In our experiment, the fundamental frequency laser was provided by a high-power 1064 nm laser generated by a master-oscillator power-amplifier (MOPA) configuration. The average-power at 80 kHz was 97.0 W, with the pulse width of 29 ns and peak power of 41.8 kW. The beam quality factors M2 were smaller than 1.15 in both directions, as shown in Fig. 2 . The maximum 49.7 W green power was obtained at 80 kHz with pulse width of 27 ns, corresponding to a conversion efficiency of 51.2%. The average UV output power was 20.2 W at 80 kHz, with pulse width of 26 ns. The UV diameter size on the output surface of the THG crystal was about 220 μm. The UV average power density on the output surface was 53.1 kW/cm2 and the peak power density was 0.026 GW/cm2. The corresponding values of IR and green laser were shown in Table 1 . These values were well below the LIDT of LBO crystal.
3. Results and discussion
Firstly, the LBO crystals, both of the as-polished and super-polished ones, with film-coated output surface were used as the THG crystal, respectively. The films were all AR at 1064 nm, 532 nm, and 355 nm (coated by Castech Crystals Inc., China and Kogakugiken Corp., Japan). Lifetime of either of the two kinds of crystals was only a few minutes (<10 min) when the UV power reached up to 20 W. Visible damage spots appeared and the UV output power dived. The UV-induced film damage has been widely studied [9,10] and the plunge of LIDT of coated LBO substrates has also been reported . However, to the best of our knowledge, there has not been any efficient way to resist the UV-induced long-term film damage except avoiding film-coating via Breswter-cut . Nevertheless, it needs to be emphasized that demands of the uncoated surface were very rigorous. In our experiment, the two uncoated as-polished LBO crystals with a few nanometer magnitude surface roughness (4.0~8.0 nm) were also used as the THG crystal. However, damage spot became visible on the output surface only after a couple of hours (<5 h). This damage was supposed to be mainly caused by the absorption effect of polishing impurities embedded in the rough surface. Therefore, in order to improve the long-term lifetime, it is very important to fabricate surface with excellent physical and chemical purities, as well as avoiding surface defects and reducing subsurface damage, which were illustrated and corroborated by the performances of the super-polished samples as follows.
In actual applications, the output power and beam quality are the most critical parameters. So dynamic characteristics of these factors during the long-term THG process were investigated for one of the super-polished crystals.
The long-term UV output power was shown in Fig. 3 . Every point in the figure represented a 2-hour average average-power at 80 kHz. The 2-hour average-power stability in standard deviation was smaller than 1.6%, as shown in the inset in Fig. 3. The variation in the short run was mainly due to the temperature fluctuations of the SHG and THG ovens. According to linear fitting, the UV power was slowly decreasing due to the damage of the output surface. The fitting line in the figure was PUV = 20.8-0.0084t, where PUV represented the UV output power and t meant the exposure time. The UV power decreasing velocity was k = 0.0084 W/h. And thus the THG LBO crystal lifetime for 15% decline of the initial power was 371 hours.
During the THG, the beam quality measurements of the UV laser were performed repeatedly in intervals of 10 hours. As shown in Fig. 3, the beam quality appeared dramatic changes after the THG LBO crystal suffering 120 hours exposure. It could be seen clearly that the modulation of power density distribution shapes became increasingly serious along with the exposure time, and thus the beam quality got worse. The beam quality factors in the orthogonal directions were 1.213 and 1.231 initially, changed to 1.258 and 1.323 after 120-hour exposure time, and degraded to 1.330 and 1.453 after 130 hours. When the THG crystal was moved to a new output point, the UV output power and beam quality returned to the initial values. It displayed clearly that the alternations of power and beam quality correlated closely with the surface deformation caused by the simultaneous action of IR, green, and UV laser. In order to investigate the damage mechanism, surface morphology and electron spectroscopy for chemical analysis (ESCA) of the exposed area were conducted.
The super-polished THG LBO crystal output surface detected by the AFM before exposure reached to 0.53 nm in RMS roughness, as shown in Fig. 4(a) . The surface after 130-hour exposure was visualized by a white light interferometer (WLI, MicroXAM-3D, ADE Phase Shift Inc.) and a common optical microscopy (Olympus BX60 Microscope). A damage spot expanded within an area of about 90 μm diameter could be observed, as shown in Figs. 4(b)-4(d). Figures 4(c) and 4(d) were detected by WLI employing 10 × and 50 × object lenses respectively. The detected damage spots were in good agreement under different measurements. Figure 4(c) illustrated that the damage spot was smaller than the UV beam diameter on the output surface. The UV beam could be supposed as an ideal Gaussian beam. Therefore, the UV intensity distribution I(r) at the output surface has the relationship I(r) = I(0)exp(−2r2/w2), where r is the polar coordinate in the radial direction and w is the radius size. Thus, the marginal damage intensity was Idamage = 0.72I(0) = 76.5 kW/cm2 for the 130-hour damage spot in the experiment. Figure 4(d) demonstrated the topography of the damage spot, which explained the degradation of the UV beam quality.
The chemical composition of three areas on the LBO surface after 130-hour THG exposure including the exposed and damaged area (point 1), exposed and undamaged area (point 2), and unaffected area (point 3), as shown in Fig. 4(c), were analyzed by X-ray Photoelectron Spectroscopy (XPS, PHI Quantera system) using AlKα radiation with photon-energy of 1486.6 eV. Every detected point was within an area of 20 μm diameter. The striking depth was about 3 nm. The energetic positions of the determined peaks were referenced to the carbon contaminants C (1s) line, assuming a value of 284.8 eV. The results were displayed in Table 2 and Figs. 5(a) and 5(b).
Li, B and O are the constituent elements of LBO crystal. The crystalline structure is consisted of (B3O7)5- chains arranged as helicoids that are oriented along the axis with the biggest refractive index . The Li+ ions are located between the chains. In the XPS experiment, concentrations of B and O were discovered in all of the three points. However, the Li element could not be probed in point 2 or point 3. The main cause was that the Li+ ions filled in the (B3O7)5- clusters on the surface were substituted by the Hydrogen ions from the polishing slurry during the CMP process according to the Cook model . Besides, the XPS striking depth was only 3 nm and the signal representing Li element was very weak. A mass of carbon element was detected on the surface as a result of adopting the nano-scale diamond polishing slurry. Actually, the contamination of carbon is hardly avoidable even without respect to the polishing slurry, because carbon is the most common element to form ambient matters . While for the exposed and damaged area (point 1) where the finished surface was ablated, the Li element was discovered. The Zn element on the surface could be explained as contaminants during polishing or other process.
In the exposed area, either point 1 or point 2, a significant peak at 103 eV emerged, as shown in Fig. 5(b). The peak position could be assigned to the existence of Si (2p) in the solution of oxy-carbide compounds formed as the SiC/SiO2 or SiOxCy according to the ref [15,16]. Due to the consideration of results published in ref , the emergence of SiO2-configuration could be attributed to the UV-assisted light deposition. The high photon energy of the 355 nm UV laser (3.5 eV) as well as its non-ignorable second harmonic 177 nm photons (7.0 eV) induced the photo-excitation of the crystal surface and photon-dissociation of oxygen (O2) as well as bonds of ambient gaseous molecules including Si to form SiO2 [17–19]. Then, the Si-compounds were chemically bound to the activated surface. The roles of both the IR and green could be neglected during the deposition process due to their relatively low photon energy when regarding of the 7.75 eV large band-gap of LBO . To some extent, it was supported by the fact that no damage was observed on the output surface of the SHG LBO crystal even after much longer time (>>100 h) of exposure.
The deposited SiOxCy impurity gradually altered the reflective rate and absorption rate of the LBO surface. This action plus the absorption effect of the pre-existed polishing remnants brought about considerable heating results and thus lead to the UV-assisted ablation and crack of the exposed area. In our experiment, when the surface was machined to super smooth roughness, the main cause of the damage process was assumed as the UV-induced deposition. On the contrary, when the surface was very rough and lifetime was very short, the embedded polishing remnants was supposed to dominate the process. The irreversible deformation of the output surface introduced the wave-front aberration, intensity distribution modulation, and thus beam quality degradation into the UV laser.
In our experiment, the 120-hour long lifetime without damage benefited from the super smooth output surface, which effectively reduced burrs and cracks and thus reduced polishing remnants embedded in the surface, compared with the as-polished ones with a few nanometer magnitude surface roughness. The surface lifetime of the LBO crystal was improved more than 20-fold without damage. As a matter of fact, the surface lifetime could be expected to be further enhanced by more precise purifying techniques, such as the ion etching technique, to remove the embedded polishing compound in the surface more comprehensively [8,20]. Furthermore, cleaner ambient environment for the THG crystal is also promising for the improvement of its lifetime.
In this paper, we demonstrated the UV-induced damage on the LBO crystal surface during long-term high-power THG. The lifetime of the output-surface with sub-nanometer roughness reached to about 130 hours for the appearance of visible damage spot and 371 hours for 15% UV power decline, which has been a drastic improvement compared with the lifetime of LBO crystal with coated output surface and those with a few nanometer magnitude surface roughness. The damage mechanism was assumed as the thermal effect brought about by the UV-induced deposition and embedded polishing remnants.
This work was partially supported by the National Natural Science Foundation of China (NSFC) (No. 60978032) and the National High Technology Research and Development Program (863 Program) of China (No. 2011AA030208). The authors express their great thanks to Prof. Xuezhan Li for the assistance of XPS experiment and Dr. Mingqian Zhang for the AFM detection.
References and links
1. C. X. Wang, G. Y. Wang, A. V. Hicks, D. R. Dudley, H. Y. Pang, and N. Hodgson, “High power Q-switched TEM00 mode diode-pumped solid state lasers with > 30 W output power at 355 nm,” Proc. SPIE 6100, 610019, 610019-14 (2006). [CrossRef]
2. C. Chen, Y. Wu, A. Jiang, B. Wu, G. You, R. Li, and S. Lin, “New nonlinear-optical crystal: LiB3O5,” J. Opt. Soc. Am. B 6(4), 616–621 (1989). [CrossRef]
3. X. Yan, Q. Liu, H. Chen, X. Fu, M. Gong, and D. Wang, “35.1 W all-solid-state 355 nm ultraviolet laser,” Laser Phys. Lett. 7(8), 563–568 (2010). [CrossRef]
4. H. Hong, Q. Liu, L. Huang, and M. Gong, “High-beam-quality all-solid-state 355 nm ultraviolet pulsed laser based on a master-oscillator power-amplifier system pumped at 888 nm,” Appl. Phys. Express 5(9), 092705 (2012). [CrossRef]
5. Y. Furukawa, S. A. Markgraf, M. Sato, H. Yoshida, T. Sasaki, H. Fujita, T. Yamanaka, and S. Nakai, “Investigation of the bulk laser damage of lithium triborate LBO single crystal,” Appl. Phys. Lett. 65(12), 1480–1482 (1994). [CrossRef]
6. F. R. Wagner, A. Hildenbrand, J. Y. Natoli, and M. Commandré, “Multiple pulse nanosecond laser induced damage study in LiB3O5 crystals,” Opt. Express 18(26), 26791–26798 (2010). [CrossRef]
7. S. Möller, A. Andresen, C. Merschjann, B. Zimmermann, M. Prinz, and M. Imlau, “Insight to UV-induced formation of laser damage on LiB3O5 optical surfaces during long-term sum-frequency generation,” Opt. Express 15(12), 7351–7356 (2007). [CrossRef]
9. C. Gouldieff, F. Wagner, L. Jensen, M. Mende, J. Natoli, and D. Ristau, “Oxide mixtures for UV coatings,” Proc. SPIE 8530, 85300T, 85300T-9 (2012). [CrossRef]
10. S. Papernov, A. Tait, W. Bittle, A. W. Schmid, J. B. Oliver, and P. Kupinski, “Near-ultraviolet absorption and nanosecond-pulse-laser damage in HfO2 monolayers studied by submicrometer-resolution photothermal heterodyne imaging and atomic force microscopy,” J. Appl. Phys. 109(11), 113106 (2011). [CrossRef]
11. M. Mende, L. O. Jensen, H. Ehlers, S. Bruns, M. Vergöhl, P. Burdack, and D. Ristau, “Applying hafnia mixtures to enhance the laser-induced damage threshold of coatings for third harmonic generation optics,” Proc. SPIE 8530, 85300W, 85300W-8 (2012). [CrossRef]
12. W. M. Grossman, J. D. Henrie, and S. Clara, “Third-harmonic generation with uncoated Brewster-cut dispersive output facet,” US Patent, US5850407 (1998).
13. L. M. Cook, “Chemical processes in glass polishing,” J. of Non-Crystalline Solids 120(1-3), 152–171 (1990). [CrossRef]
14. T. Schmid, A. Messmer, B. S. Yeo, W. Zhang, and R. Zenobi, “Towards chemical analysis of nanostructures in biofilms II: tip-enhanced Raman spectroscopy of alginates,” Anal. Bioanal. Chem. 391(5), 1907–1916 (2008). [CrossRef]
15. W. E. Morgan and J. R. Wazer, “Binding energy shifts in the X-Ray photoelectron spectra of a series of related group IV-a compounds,” J. Phys. Chem. 77(7), 964–969 (1973). [CrossRef]
16. M. D. Irwin, C. G. Pantano, P. Gluche, and E. Kohn, “Bias-enhanced nucleation of diamond on silicon dioxide,” Appl. Phys. Lett. 71(5), 716–718 (1997). [CrossRef]
17. P. Boyer, G. Roche, W. Ritchie, and G. Collins, “Laser-induced chemical vapor deposition of SiO2,” Appl. Phys. Lett. 40(8), 716–719 (1982). [CrossRef]
18. K. Awazu and H. Onuki, “Photoinduced synthesis of amorphous SiO2 with tetramethoxysilane,” Appl. Phys. Lett. 69(4), 482–484 (1996). [CrossRef]
19. H. Takao, M. Okoshi, and N. Inoue, “SiO2 films fabricated by F2 laser-induced chemical deposition using silicone rubber,” Appl. Phys., A Mater. Sci. Process. 79(4-6), 1567–1570 (2004). [CrossRef]
20. T. Kamimura, S. Fukumoto, R. Ono, Y. K. Yap, M. Yoshimura, Y. Mori, T. Sasaki, and K. Yoshida, “Enhancement of CsLiB6O10 surface-damage resistance by improved crystallinity and ion-beam etching,” Opt. Lett. 27(8), 616–618 (2002). [CrossRef]