Open Access
Add to BibSonomyAbstract
Multiple pulse nanosecond laser induced damage in the bulk of LiB3O5 (LBO) crystals was investigated at 1064 nm, 532 nm and 355 nm. Scanning electron microscopy of cleaved damage sites confirmed the presence of different zones that have already been reported in the case of KH2PO4 (KDP). Multi pulse measurements reveal a strong decrease of the damage threshold with increasing pulse number at 1064 nm (fatigue effect). A weaker fatigue effect was observed at 532 nm and no fatigue effect was found at 355 nm. This observation is best explained by an inherently statistical light matter interaction generating laser induced damage. Finally, a polarization dependent damage threshold anisotropy was evidenced at all three wavelengths, being strongest at 1064 nm. The results indicate the importance of Li+ vacancy stabilized color centers for the damage mechanism.
©2010 Optical Society of America
1. Introduction
Lithium triborate, LiB3O5, or short LBO, is a high-performing nonlinear optical material that is mostly applied to frequency conversion tasks. The orthorhombic crystals are described by point group mm2 and space group Pna21. The unit cell contains 36 atoms (4 × LiB3O5) [1]. The crystalline structure is made up by (B3O7)5--chains arranged as helicoids that are oriented along the z-axis of the crystal. Here the z-axis is defined as the axis with the highest refractive index. The Li+-ions are located between the chains and a strongly anisotropic ionic conductivity is observed [2]: The ionic conductivity in z-direction is up to 104 times higher compared to the ionic conductivity in x and y directions.
LBO is an optically biaxial crystal with very good UV transparency (down to 160 nm) and is therefore preferentially used in UV-applications like third harmonic generation for Nd:YAG and similar lasers [3]. Recently more than 80% doubling efficiency have been demonstrated using LBO on a large scale nanosecond laser facility [4]. The laser damage threshold of LBO at 1064 nm has been reported to be 1.5 times higher than the one of fused silica [5,6] using a pulse duration of 1.1 ns.
In this paper, a multi pulse nanosecond laser damage study of LBO is presented using three wavelengths (1064 nm, 532 nm and 355 nm). The damage site morphology and damage probability curves as a function of the polarization of the incident light are presented. Further, the development of the laser induced damage threshold with increasing pulse number is studied and conclusions on the possible damage mechanism are given.
2. Experimental details
The used LBO samples were produced by Cristal Laser SA in Messein, France, using a flux method. For this study, five x-cut LBO crystals (i.e. laser propagation direction along the x-axis) of section 10 x 10 mm2 and thickness 5 mm were used. A reproducibility test using two different crystals showed no notable damage probability variations due to crystal quality variations. In particular the results of the multi pulse experiments superpose perfectly [7].
The test procedure applied for our experiments was described in detail in [8]. In summary, an S-on-1 test [9] is carried out using 20 sites per fluence and a maximum of 10’000 laser pulses per site at 10 Hz pulse repetition rate. The results are represented as damage probability curves with error bars corresponding to 68% confidence [10]. The damage probability curves are modeled using a power law distributed damage precursor ensemble. Direct evidence for this type of precursor distribution has been provided in KDP [11].
Bulk damage was investigated at three wavelengths: 1064 nm, 532 nm and 355 nm. Two Quantel Ultra GRM laser sources have been used: one at 1064 nm and the other at 355 nm. The 532 nm light has been obtained by external frequency doubling the IR laser in a 10 mm long KTP crystal and separating the two wavelengths using an infrared blocking filter. Figure 1 shows a schematic of the experimental setup. The pulse durations at 1064 nm, 532 nm and 355 nm were 6 ns, 5.5 ns and 5 ns respectively.
Fig. 1 Schematic of the laser damage setup. One wavelength only is used for a given measurement. The used components are: SHG, KTP type II frequency doubling crystal for 1064 nm; IR BF, infrared blocking filter used together with the SHG crystal if a 532nm measurement is carried out; M, 355 nm mirror on a flip mount; λ/2, half-wave plate for the used wavelength; Pol., Glan-laser polarizer; Detector, pyroelectric energy detector; NG, neutral grey filter; BF, a blocking filter for the used wavelength and f. lamp, fibred halogen lamp. In situ damage detection is performed by imaging scattered light with a zoom lens and a CCD camera.
The laser was focused by a plano-convex fused silica lens as to obtain a spot diameter at the waist of 25 µm (at 1/e2) for all wavelengths. The focus of the test beam is located approximately in the middle of the 5 mm thick crystals and the intensity profile in the focal region is close to a Gaussian beam. As the measurements were carried out in nonlinear crystals special care concerning the laser damage metrology has been taken. In particular, we checked that the aberrations due to the focusing in the birefringent crystal are negligible by imaging the focal plane through two joined LBO crystals (totaling a thickness of 10 mm) and comparing the obtained beam profile with the beam profile in air [10].
For comparison we also measured the 1-on-1 bulk damage threshold of a synthetic fused silica sample with the same setup. At 1064 nm, 532 nm and 355 nm we obtained respectively 100 J/cm2, 100 J/cm2 and 45 J/cm2. The infrared and UV damage threshold values are close to former measurements with a spot diameter of 8 µm using the same laser sources [12].
All damage thresholds given in this article are 0%-damage thresholds for the used conditions, i.e. the highest fluence with 0% damage probability is defined as the experimental damage threshold.
The damage morphology of the irradiated samples was observed ex-situ with a Carl Zeiss Axiotech microscope working with polarized light and a table-top SEM (Hitachi TM-1000).
3. Results and discussion
3.1 Damage location
Many nonlinear crystals like KTP, RTP and KDP preferentially damage in the bulk of the crystal when irradiated with parallel laser beams. During preliminary experiments we observed that LBO behaves similarly for 355 nm irradiation. However, damage appears preferentially on the surfaces when parallel 1064 nm light is used. As this study aims information on the bulk damage threshold, a focused beam with a diameter of 25 µm is used.
3.2 Damage morphology
A comparison of the bulk damage morphology in LBO with respect to KTP is shown in Fig. 2 . Both crystal types, LBO and KTP, are orthorhombic crystals of the same point group and the same space group. The cracks created during the damage process have nevertheless different orientations. Thus, despite the same crystal symmetry, the mechanically weakest directions in LBO are not the same as in KTP.
Fig. 2 Optical microscope images of laser induced damage sites in x-cut LBO (a) and x-cut KTP (b). The used fluences were 240 J/cm2 and 27 J/cm2 respectively. In both cases the observation direction of the images is oriented along the irradiation direction.
The fluences used to create the shown damages were approximately two times the damage threshold in both materials and the used beam diameter was identical. The significant difference in absolute energy deposition during the damage event leads to a much higher number of cracks in the case of LBO. The lateral dimensions of the damage sites are however very similar.
Figure 3 shows a scanning electron micrograph of a bulk damage site that has been cleaved. Similarly to what has been reported for KDP [13] three regions may be distinguished: A void, a shell of modified material that stays in one piece after cleaving, and the zone with the micro cracks that allowed for easy cleaving.
Fig. 3 Scanning electron micrographs of cleaved bulk damage sites generated at high fluence in x-cut LBO. The plane of the break is perpendicular to the laser propagation direction. Different zones are visible: void, modified shell, and crack zone. The inset shows the size of the crack zone.
3.3 Wavelength effect
For comparison with former studies [5,6], Fig. 4 shows laser damage probability curves for a 1-on-1 test carried out in x-cut LBO using y-polarized light at the three wavelengths. The bulk damage threshold is strongly wavelength dependent. Comparing the infrared laser damage threshold of LBO to the one of synthetic fused silica we obtain that LBO exhibits 1.5 times higher laser damage resistivity than fused silica. This value is intermediate to the one reported by Yoshida et al. [6] who reported a factor of 1.3 and the value of Furukawa et al. [5] who measured a factor of 1.8.
Fig. 4 Single pulse damage probability curve of x-cut LBO measured with y-polarized light at 1064 nm, 532 nm and 355 nm. The beam diameter was approximately 25 µm for all wavelengths. The solid lines are generated using a power law distributed damage precursor ensemble.
According to our measurements, the bulk laser damage threshold of LBO in the UV-region at 355 nm, is still at approximately 50% of the bulk threshold of synthetic fused silica. At this wavelength, Yoshida et al. [6] report a similar damage threshold for LBO and synthetic fused silica. Furukawa et al. [5] even measured a 1.7 times higher damage threshold of LBO with respect to fused silica.
The difference in the slopes of the 355 nm damage curve compared to the 1064 nm damage curve is due to the different focal volumes that have been used for the tests. As all measurements have been carried out using the same spot diameter of 25 µm, the infrared beam is the most divergent one and thus the one with the smallest focal volume corresponding to the smallest slope in the damage probability curve.
3.4 Multiple pulse fatigue effect
For functional use, laser damage studies using multiple-pulse irradiation with a high number of pulses are very important. Laser damage measurements with 104 pulses per site have been realized at the three wavelengths. Figure 5 shows the damage threshold as a function of the pulse number.
Fig. 5 Evolution of the bulk damage threshold of LBO versus pulse number measured at 1064 nm 532 nm and 355 nm using x-cut crystals and y-polarized light. The dashed lines are guides to the eye.
For the tests at 1064 nm a strong fatigue effect is observed, i.e. the damage threshold decreases with increasing pulse number. However, a constant damage threshold is measured for more than 1000 pulses. For this wavelength the ratio between the 1-on-1 threshold and the 1000-on-1 threshold is 1.85. At 532 nm the observed fatigue effect is much weaker: the threshold is constant for more than 10 pulses and the ratio between the thresholds is only 1.30. Finally, for 355 nm no fatigue effect is observed.
Besides the practical importance of multi pulse laser damage studies, this data helps to better understand the laser damage mechanism. The fatigue effect at the IR wavelength has been observed before in fused silica as well as in nonlinear crystals [8,12]. At given fluence, spot size and pulse repetition rate, the increasing laser damage probability with increasing pulse number may be understood in two different ways. Either the fatigue effect is caused by a material modification [12,14], or it is a consequence of the statistical nature of nanosecond laser damage [15,16].
The data in Fig. 5 show that the fatigue effect decreases with increasing photon energy. In our opinion this observation is a point against the presence of material modifications, as material modifications are more likely to be generated by short wavelength irradiation.
Concerning the statistical nature of nanosecond laser damage, one might think of statistical fluctuations in the temporal pulse profile of the used multi mode lasers. In fact the temporal pulse profiles of both lasers exhibit multiple sharp peaks in an approximately 5 ns Gaussian envelope. The maximum intensity of our pulses, which is reached during some tens of picoseconds only, is thus strongly fluctuating. The data however does not confirm the hypothesis that the peak intensity fluctuations cause the fatigue effect, because the 532 nm pulses are expected to exhibit stronger intensity fluctuations than the infrared pulses from which they have been produced by frequency doubling but nevertheless the fatigue effect is reduced at 532 nm.
Other types of laser fluctuations should also be considered. If spatial jitter (depointing) or pulse energy variations of the used laser are important, an increasing number of pulses sample an increasing volume and the apparent laser damage threshold decreases. We addressed this question before in KTP using the same laser source and found that the spatial jitter, even combined with the pulse energy fluctuations, is not strong enough to explain the observed fatigue effect [16]. As the fatigue effect in LBO is similar to the one in KTP in the sense that the multi pulse damage threshold stabilizes at approximately at 60% of the single pulse damage threshold, we conclude that in the case of LBO too spatial jitter and pulse energy fluctuations are not strong enough to explain the fatigue effect.
Our data thus supports that the light matter interaction itself is statistical, as expected for quantum mechanical processes like electronic excitation of distinct defect levels. The reduced fatigue effect for short wavelengths also indicates that the probability of interaction between laser and defect approaches unity for short wavelengths. Wavelength dependent single pulse nanosecond laser damage measurements in KDP have shown before that multi photon processes are important for the damage initiation [17]. Multi photon processes may thus play a role in laser damage initiation in LBO too. The defects interacting with the incident irradiation by multi photon processes may be inherent to the material, as recently shown in RTP, where the infrared nanosecond laser damage threshold is independent of important crystal quality variations [18].
3.5 Damage threshold anisotropy
Damage probability curves at the three wavelengths have been measured in 1-on-1 mode for x-cut crystals using y-polarized and z-polarized light (Fig. 6 ). We found that, similarly to RTP and KTP [8,19], LBO exhibits polarization dependent anisotropy of the damage threshold. Our measurements thus differ from earlier data at 1064 nm, where no dependence on polarization direction had been observed [5,6].
Fig. 6 Bulk damage probability curves of LBO for 1064 nm, 532 nm and 355 nm using x-cut crystals (light propagation direction along the x-axis). The solid lines are generated using a power law distributed damage precursor ensemble.
The polarization anisotropy appears clearly at the infrared wavelength (a factor of 3.2 between the 1-on-1 thresholds for y-polarized and z-polarized light) and much weaker at the visible and the UV-wavelengths (factors of 1.15 and 1.25 respectively).
Concerning the polarization dependent damage threshold anisotropy in KTP and RTP [8,19] we proposed an explanation of the increased z-polarization damage threshold based on an increased rate of cation hopping along the z-direction of the crystal, if z-polarized light is used. Ion hopping reduces the lifetime of the color centers that are stabilized by neighboring metal ion vacancies. In consequence, light absorption is less localized in the case of z-polarized light and an increased laser damage threshold is observed.
In LBO too, color centers are stabilized by a nearby Li-ion vacancy [20] and the strongly enhanced ionic conductivity in z-direction [2] indicates a tendency to ion hopping under the influence of a z-polarized electric field. Thus, even if the laser damage threshold in LBO is approximately ten times higher than in KTP, the reason for the polarization dependent anisotropy might be the same in both crystals.
The fact that the polarization dependent damage threshold anisotropy is observed at all three wavelengths indicates that color centers probably play an important role in the damage mechanism over the full spectral range (IR to UV). Further, we might explain qualitatively the differences in the strength of the polarization anisotropy as follows: At 1064 nm, where the defect level excitation is the less probable, a damage event requires a relatively long lifetime of the color center. The probability to observe a damage event is thus strongly affected by the life time reduction caused by z-polarized light. The difference between 532 nm and 355 nm wavelengths may be explained by the wavelength dependence of the laser-ion interaction in the case of z-polarized light. In fact, the electric field of the laser oscillates much too fast in order to allow for a direct interaction. Most probably the photon-ion interaction is mediated by a parametric process for which a weak wavelength dependence is expected. Our results suggest that 355 nm light is more efficient in inducing Li-ion hopping than 532 nm light.
4. Summary and conclusions
We studied nanosecond laser induced damage in x-cut LiB3O5 crystals at 1064 nm, 532 nm and 355 nm. Using synthetic fused silica as a reference material, single pulse measurements showed good agreement with earlier studies at 1064 nm. At 532 nm and 355 nm however our measurements indicate a lower single pulse damage threshold [5,6].
The damage morphology has briefly been described using optical microscopy and scanning electron microscopy of a cleaved sample. The difference in the crack patterns for two crystals of the same symmetry group (KTP and LBO) has been shown using optical microscopy and the detailed damage morphology as observed by the scanning electron micrograph was qualitatively similar to the one reported earlier for KDP [13].
The multi pulse measurements revealed a strong fatigue effect at 1064 nm, a weaker fatigue effect at 532 nm and no fatigue effect for the UV wavelength of 355 nm. The fact that for shorter irradiation wavelength less “fatigue” is observed indicates that in fact the observed effect is rather a statistical effect than a real, material modification driven “fatigue” effect. Further, it seems that the light matter interaction itself is statistical as the statistics of the laser intensity fluctuations between the 1064 nm pulses and externally generated 532 nm pulses is contrary to the observed fatigue effect. This conclusion of an inherently statistic nature of the light matter interaction is compatible with quantum mechanical processes like multi-photon excitation of defect levels in the band gap of the material. The relevant defects may be inherent to the material as recently shown by the independence of the infrared nanosecond laser damage threshold of RTP on crystal quality [19].
Finally, a strong, polarization dependent damage threshold anisotropy has been evidenced at 1064 nm. At 532 nm and 355 nm the damage threshold anisotropy has been observed to a lesser extend. In analogy to a qualitative model existing for KTP [8,21] the results indicate the importance of Li+ vacancy stabilized color centers for the damage mechanism.
Acknowledgments
The authors would like to thank the Délégation Générale pour l’Armement (DGA) for financial support.
References and links
1. Q. Y. Shang, B. S. Hudson, and C. O. Huang, “Infrared and Raman Spectra of Lithium Triborate - Vibrational Assignments and a Correlation with its Nonlinear Optical Activity,” Spectrochim. Acta [A] 47(2), 291–298 (1991). [CrossRef]
2. S. F. Radaev, N. I. Sorokin, and V. I. Simonov, “Atomic Structure and One Dimensional Ionic Conductivity of LiB3O5,” Fiz. Tverd. Tela 33, 3597–3600 (1991).
3. D. N. Nikogosyan, “Lithium Triborate (LBO) - A Review of its Properties and Applications,” Appl. Phys., A Mater. Sci. Process. 58(3), 181–190 (1994). [CrossRef]
4. G. Mennerat, J. Rault, O. Bonville, P. Canal, P. Villeval, B. Rainaud, H. Albrecht, D. Lupinski, A. Kokh, N. Kononova, V. Vlezko, K. Kokh, G. Chériaux, M. Pittman, J.-P. Chambaret, and G. Mourou, “115 J, 85% efficiency second harmonic generation in LBO,” in CLEO/QELS, (OSA Technical Digest (CD), 2008), paper CPDA1. http://www.opticsinfobase.org/abstract.cfm?URI=CLEO-2008-CPDA1
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, LiB3O5, single crystals,” Appl. Phys. Lett. 65(12), 1480–1482 (1994). [CrossRef]
6. H. Yoshida, H. Fujita, M. Nakatsuka, M. Yoshimura, T. Sasaki, T. Kamimura, and K. Yoshida, “Dependences of laser-induced bulk damage threshold and crack patterns in several nonlinear crystals on irradiation direction,” Jpn. J. Appl. Phys. 45(Part 1), 766–769 (2006). [CrossRef]
7. F. R. Wagner, A. Hildenbrand, J. Y. Natoli, and M. Commandré, “Nanosecond-laser induced damage at 1064 nm, 532 nm, and 355 nm in LiB3O5,” Proc. SPIE 7504, 75041M (2009). [CrossRef]
8. A. Hildenbrand, F. R. Wagner, H. Akhouayri, J. Y. Natoli, M. Commandré, F. Théodore, and H. Albrecht, “Laser-induced damage investigation at 1064 nm in KTiOPO4 crystals and its analogy with RbTiOPO4.,” Appl. Opt. 48(21), 4263–4269 (2009). [CrossRef] [PubMed]
9. International Organization for Standardization, “Determination of laser-damage threshold of optical surfaces Part 2: S-on-1 test,” (ISO 11254–2, 2001), p. 29.
10. A. Hildenbrand, F. R. Wagner, H. Akhouayri, J.-Y. Natoli, and M. Commandré, “Accurate metrology for laser damage measurements in nonlinear crystals,” Opt. Eng. 47(8), 083603 (2008). [CrossRef]
11. C. W. Carr, M. D. Feit, M. C. Nostrand, and J. J. Adams, “Techniques for qualitative and quantitative measurement of aspects of laser-induced damage important for laser beam propagation,” Meas. Sci. Technol. 17(7), 1958–1962 (2006). [CrossRef]
12. J. Y. Natoli, B. Bertussi, and M. Commandré, “Effect of multiple laser irradiations on silica at 1064 and 355 nm,” Opt. Lett. 30(11), 1315–1317 (2005). [CrossRef] [PubMed]
13. C. W. Carr, M. D. Feit, M. A. Johnson, and A. M. Rubenchik, “Complex morphology of laser-induced bulk damage in K2H(2-x)DxPO4 crystals,” Appl. Phys. Lett. 89(13), 131901 (2006). [CrossRef]
14. A. E. Chmel, “Fatigue laser-induced damage in transparent materials,” Mater. Sci. Eng. B 49(3), 175–190 (1997). [CrossRef]
15. M. Bass and H. H. Barrett, “Avalanche Breakdown and the Probabilistic Nature of Laser-Induced Damage,” IEEE J. Quantum Electron. 8(3), 338–343 (1972). [CrossRef]
16. F. R. Wagner, A. Hildenbrand, L. Gallais, H. Akhouayri, M. Commandré, and J. Y. Natoli, “Statistical interpretation of S-on-1 data and the damage initiation mechanism,” Proc. SPIE 7132, 71320Y (2008). [CrossRef]
17. C. W. Carr, H. B. Radousky, and S. G. Demos, “Wavelength dependence of laser-induced damage: determining the damage initiation mechanisms,” Phys. Rev. Lett. 91(12), 127402 (2003). [CrossRef] [PubMed]
18. A. Hildenbrand, F. R. Wagner, J.-Y. Natoli, and M. Commandré, “Nanosecond laser induced damage in RbTiOPO4: the missing influence of crystal quality,” Opt. Express 17(20), 18263–18270 (2009). [CrossRef] [PubMed]
19. F. R. Wagner, A. Hildenbrand, J.-Y. Natoli, M. Commandré, F. Theodore, and H. Albrecht, “Laser induced damage threshold of RbTiOPO4: evidence of polarization dependent anisotropy,” Opt. Express 15, 13849–13857 (2007). [CrossRef] [PubMed]
20. W. Hong, M. M. Chirila, N. Y. Garces, L. E. Halliburton, D. Lupinski, and P. Villeval, “Electron paramagnetic resonance and electron nuclear double resonance study of trapped hole centers in LiB3O5 crystals,” Phys. Rev. B 68(9), 094111 (2003). [CrossRef]
21. X. B. Hu, J. Y. Wang, H. J. Zhang, H. D. Jiang, H. Liu, X. D. Mu, and Y. J. Ding, “Dependence of photochromic damage on polarization in KTiOPO4 crystals,” J. Cryst. Growth 247(1-2), 137–140 (2003). [CrossRef]
