Open Access
Add to BibSonomyAbstract
We present a method of third harmonic generation at 355 nm by frequency mixing of fundamental and second harmonic radiation of an ytterbium nanosecond pulsed all-fiber laser in a type-I phase-matched LiB3O5 (LBO) crystal where originally orthogonal polarization planes of the fundamental and second harmonic beams are aligned by an optically active quartz crystal. 8 W of ultraviolet light at 355 nm were achieved with 40% conversion efficiency from 1064 nm radiation. The conversion efficiency obtained in a type-I phase-matched LBO THG crystal was 1.6 times higher than the one achieved in a type-II LBO crystal at similar experimental conditions. In comparison to half-wave plates traditionally used for polarization alignment the optically active quartz crystal has much lower temperature dependence and requires simpler optical alignment.
©2013 Optical Society of America
1. Introduction
Currently, there is a growing demand for high power ultraviolet (UV) pulsed lasers for various industrial applications such as LED scribing, chip dicing, via-hole drilling, plastics marking and others. In comparison with more common IR lasers, UV lasers have an advantage of higher linear and nonlinear absorption of the UV light by some materials and their possibility to achieve smaller focus spots. The majority of commercially available pulsed UV lasers are diode pumped solid state Nd:YVO4, Nd:YAG, or Nd:YLF lasers or Yb-doped fiber master oscillator power amplifier (MOPA) lasers operating near 1 µm wavelength with inter- or intra- cavity frequency tripling or quadrupling (see [1] and references therein). The conventional way of third harmonic generation (THG) employed in most of the UV lasers operating near 0.35 µm consists of a two stage process: second harmonic generation (SHG) in a type-I phase-matched nonlinear optical crystal and sum frequency generation of the fundamental and second harmonics in a type-II phase-matched crystal. Usually LiB3O5 (LBO) crystals are used for both processes due to their high damage threshold, high nonlinearity, low absorption in visible and UV ranges, and high crystal growth yield. The popularity of the described scheme can be explained by its ease of implementation: in the output of the first nonlinear crystal we have the fundamental and double frequency beams polarized in the orthogonal planes which is exactly what is required for type-II phase-matching condition in the second nonlinear crystal. Because of that there is no beam manipulation needed between the nonlinear crystals except focusing of the beams. The alternative way of THG is to use type-I phase-matched crystals for both processes [2]. There is a benefit of doing it that way in getting higher conversion efficiency of sum-frequency generation under type-I phase-matching condition in some of the common nonlinear crystals. Table 1 compares the THG parameters in type-I and type-II LBO crystals. Using the values of effective nonlinear susceptibility deff from Table 1 and the fact that the output intensity of the third harmonic beam is proportional to square of deff, one can calculate that the total conversion efficiency is about 2.2 times higher in the type-I phase-matched LBO crystal than the one in the type-II crystal at 355 nm wavelength at 60° C. In addition, there is no spatial walk-off of the fundamental and second harmonic beams in the type-I phase-matching scheme, which removes the crystal length limitation present in type-II phase-matching [3]. There is, however, non-zero spatial walk-off of the third harmonic beam with respect to the fundamental and the second harmonic beams. This effect leads to some ellipticity of the 355 nm output beam, which is considered to be a minor problem and could be compensated, for example, by an anamorphous prism arrangement. Thus due to higher nonlinearity and the absence of crystal length limitation THG in a type-I LBO crystal is significantly more efficient. This is especially important for devices with low IR pump peak power, such as fiber lasers. Higher efficiency also provides a way to decrease crystal material degradation rate by relaxing the focusing conditions in the THG crystal.
Table 1. THG Parameters for Type-I and II Phase-matched LBO Crystals at 355 nm Wavelength at 60°C
The peculiarity of the type-I phase-matching scheme, however, is the fact that both fundamental and double frequency beams must be polarized in the same plane. That means that a polarization control element is required after the first nonlinear crystal. Usually a wave plate is inserted between the nonlinear crystals for that purpose [6]. This wave plate should simultaneously provide a half wavelength phase shift to the fundamental beam and a whole wavelength phase shift to the second harmonic beam. If the phase axis of such a wave plate is oriented at 45° with respect to the polarization plane of the fundamental beam the wave plate will flip the fundamental beam polarization by 90 degrees, while the polarization of the second harmonic beam will remain unchanged. That way the polarizations of the fundamental and second harmonic beams will become collinear. There are the following drawbacks of the use of the wave plate, however: 1) high temperature dependence of the phase shift (that may lead to the third harmonics power instability), 2) resonant wavelength dependence (which requires very high precision manufacturing of the wave plate, which in turn leads to that element’s high cost), 3) the requirement of precise angular adjustment around the beam propagation axis (which complicates the wave plate installation). It is worth to mention that because of the dispersion of the absolute value of the ordinary and extraordinary indexes of refraction | no – ne |, it is impossible to make a single wavelength phase shift wave plate which works at both fundamental and second harmonic wavelengths. Thus the required wave plate will have a phase shift of considerable number of integer wavelengths (N) and in turn will have N times higher thermal dependence.
In this paper we propose to use optical activity [7] in a nonlinear optical crystal (e.g. quartz, LiIO3, TeO2, etc.) instead of a wave plate to align polarization planes of fundamental and second harmonic beams before a type-I phase-matched crystal used for sum-frequency generation. Specific optical activity (polarization rotation angle per unit length) ρ in these crystals has a strong wavelength dependence (it is monotonically decreases in the near IR- visible range) [8]. Because of that if both fundamental and second harmonic beams are propagating through an optically active material in the optical axis direction their polarization planes will rotate with different speeds and will exactly coincide at a certain position. The minimal length L of the optically active crystal to get the fundamental and second harmonic beams polarizations aligned may be calculated by this formula:
where ρFH and ρSH are specific optical activities at fundamental and second harmonic wavelengths respectively.Let’s consider propagation of the collinear and initially perpendicularly polarized fundamental and second harmonic beams through the optically active crystal of length L given by Eq. (1). At the exit of the optically active crystal the fundamental and second harmonic beam polarizations will rotate by angles ϕFH = ρFH L and ϕSH = ρSH L = 90° + ρFH L around the beam propagation axes, respectively. Thus the fundamental and second harmonic beams will be collinearly polarized at an angle of ϕFH with respect to the original fundamental harmonic polarization.
The benefit of the described method is its much lower temperature dependence. Figure 1 presents the experimentally measured dependence of polarization rotation angle ϕFH of a 6.9 mm long quartz crystal at 1064 nm. As seen in Fig. 1 there is practically no change in ϕFH in the range of 15-55° C. In addition, no precision tuning of the crystal is required.
Fig. 1 Experimental temperature dependence of a polarization rotation angle ϕFH of a 6.9 mm quartz crystal.
Table 2 presents the required minimal optically active crystal length L, polarization rotation angle ϕFH and specific optical activities ρFH, ρSH for SiO2 (quartz) and TeO2 crystals at 1064 and 532 nm, respectively. Whether the resulting polarization rotation angle is negative or positive depends on the type of the optical activity (clockwise or counterclockwise).
Table 2. Properties of Some Optically Active Crystals
2. Experimental results and discussion
The experimental setup is shown in Fig. 2 . An ytterbium nanosecond pulsed fiber laser (IPG Photonics) was employed as a pump source. The output of the pump laser was a linearly polarized single mode beam at 1064 nm wavelength, less than 0.1 nm linewidth (FWHM), 21.5 W average power, 1.5 ns pulse duration (FWHM), 18 kW peak power. The repetitionrate of the optical pulses was adjusted in the range of 20 to 791 kHz. The parameters of the output optical pulse (i.e. pulse energy, duration and peak power) were constant at all repetition rates. All of the experimental data presented in this paper were obtained at 791 kHz repetition rate. The pump laser output beam had a 0.5 mm diameter (FWHM) after a collimator. The beam was focused into a 20 mm long type-I LBO crystal with a 40-mm focal length lens for second harmonic generation. For SHG we used noncritical phase-matching at 150° C. Both input and output surfaces of the LBO crystal were antireflection (AR) coated at 1064 and 532 nm wavelengths. The output radiation of the SHG LBO crystal consisted of orthogonally polarized fundamental and second harmonic co-propagating beams. Those beams were focused into the second LBO crystal with a 35-mm focal length lens for third harmonic generation.
We used type-II and type-I phase-matched 15 mm long LBO crystals for THG. Both crystals had AR coating at 1064 and 532 nm on the entrance side and a SiO2 protective coating on the exit side. The crystal temperature varied from 55 to 155° C. When the type-I crystal was used we placed a 4.39 mm long optically active quartz crystal after the SHG crystal to match the polarizations of the fundamental and second harmonics. The quartz crystal had both sides AR coated at 1064 and 532 nm wavelengths. The output beam from the THG crystal was directed into two dichroic mirrors to filter out the unconverted fundamental and second harmonic radiation. The minimum beam waist diameter of the infrared (IR) beam was located in the middle of the SHG LBO crystal and equaled to about 70 µm. Under these condition an optimum 75% second harmonics conversion efficiency was achieved. We detuned the SHG efficiency from the optimum by adjusting the crystal temperature, however, to optimize a ratio between the second and fundamental harmonics powers for the highest THG efficiency. The position of lens #2 and the position of the THG crystal were adjusted along the beam propagation axis. This way the focus waist diameter in the THG crystal was varied in the range of 50 to 150 µm. We performed an optimization process for each THG crystal studied. The best efficiency for both 15 mm long type-I and II THG crystals was achieved with approximately a 70 µm IR beam waist diameter. Since the polarization plane of the fundamental beam was rotated by the quartz crystal by 27.7 degrees with respect to the polarization plane of the Yb fiber laser output beam (see Table 2), the THG crystal was rotated by the same angle around the beam propagation axis to get the crystal’s Z axis to be aligned with the polarization plane of the incoming fundamental and second harmonic beams.
The experimentally measured dependence of 355 nm radiation power and THG conversion efficiency on 1064 nm pump power is shown in Fig. 3 . The type-II THG crystal was cut for phase-matching at 60 °C temperature, while the type-I one was cut for 100 °C. By angular adjustment of the THG crystal it was possible to get phase-matching at a different temperature. However, the type-II crystal required much large angular tuning than the type-I one for a given temperature change. Because of that we were able to measure the performance of the type-II crystal only around 60 °C, while the type-I crystal was tested in a wider range of 55 to 155° C. Please note that we calculated the total generated 355 nm power from the actually measured one by multiplying by a correction factor of 1.05 responsible for reflection losses in the output crystal surface which did not have an antireflection coating. As seen from Fig. 3, the maximum 355 nm radiation power of 6.4 W was achieved in the type-I THG crystal at 55° C temperature. That corresponds to 30% conversion efficiency from 1064 nm to 355 nm, which is 1.6 times higher than the one obtained in the type-II THG crystal scheme. At 155 °C 7 W of 355 nm power were achieved in the type-I LBO with 33% conversion efficiency. As also seen from Fig. 3 in the type-I crystal we did not reach 355 nm conversion efficiency saturation even at the highest pump power, therefore we expect to get higher efficiencies at higher pump laser peak powers. As mentioned in Introduction, another way to increase efficiency is to use longer type-I THG crystals. Figure 4 shows the IR power dependence of 355 nm output power and THG conversion efficiency obtained in a 27 mm long LBO crystal with both uncoated sides. 8 W of UV were achieved with conversion efficiency of 40% with no saturation reached. Please note that we used a correction factor of 1.05 again to take into account the output surface reflection losses. The input surface losses were not taken into account. The measurement was done at 50 °C crystal temperature. As discussed above at a higher temperature one would get to even higher UV power and higher conversion efficiency. We were not able to verify that, however, because of the operation temperature limitation of the oven used for the 27 mm crystal.
Fig. 3 Third harmonic radiation power (a) and THG conversion efficiency (b) versus fundamental pump power for the type-I LBO crystal at temperatures of 155° C (triangles), 55° C (circles), and type-II LBO crystal at 60° C temperature (squares). Both crystals were 15 mm long.
Fig. 4 Third harmonic radiation power (a) and THG conversion efficiency (b) versus fundamental pump power for the type-I 27 mm long LBO crystal at 50° C temperature.
3. Conclusion
We have demonstrated a method of third harmonic generation based on the use of a type-I phase-matched LBO crystal and an optically active quartz crystal. This method is significantly more efficient than the conventional type-II phase-matching scheme. The output of an all-fiber nanosecond Yb-doped MOPA laser was converted into 8 W of 355 nm radiation with conversion efficiency of 40%. It was shown experimentally that the conversion efficiency obtained in the type-I phase-matched LBO THG crystal was 1.6 times higher than the one achieved in the type-II crystal at similar experimental conditions. In comparison to half-wave plates traditionally used for polarization alignment the optically active quartz crystal has much lower temperature dependence and requires simpler optical alignment.
Acknowledgments
We would like to thank Igor E. Samartsev for fruitful discussions.
References and links
1. A. Starodoumov and N. Hodgson, “Harmonic generation with fiber MOPAs and solid state lasers–technical challenges, state-of-the-art comparison and future developments,” Proc. SPIE 7912, 79120H–1, 79120H-14 (2011). [CrossRef]
2. A. Borsutzky, R. Briinger, Ch. Huang, and R. Wallenstein, “Harmonic and sum-frequency generation of pulsed laser radiation in BBO, LBO, and KD*P,” Appl. Phys. B 52, 55–62 (1991). [CrossRef]
3. C. Jung, W. Shin, B.-A. Yu, Y. L. Lee, and Y.-C. Noh, “Enhanced 355-nm generation using a simple method to compensate for walk-off loss,” Opt. Express 20(2), 941–948 (2012). [CrossRef] [PubMed]
4. K. V. Diesperov, V. G. Dmitriev, and A. A. Kazakov, “Specific features of second-harmonic generation in biaxial nonlinear-optical crystals,” Laser Phys. 6, 1040–1049 (1996).
5. V. G. Dmitriev, G. G. Gurzadyan, and D. N. Nikogosyan, Handbook of Nonlinear Optical Crystals (Springer-Verlag, 1997).
6. D. A. V. Kliner, F. Di Teodoro, J. P. Koplow, S. W. Moore, and A. V. Smith, “Efficient second, third, fourth, and fifth harmonic generation of a Yb-doped fiber amplifier,” Opt. Commun. 210(3-6), 393–398 (2002). [CrossRef]
7. D. Laurence, Barron, Molecular light scattering and optical activity, 2nd ed. (Cambridge University Press, 2009).
8. V. A. Bagan, B. L. Davydov, and I. E. Samartsev, “Characteristics of Cornu depolarisers made from quartz and paratellurite optically active crystals,” Quantum Electron. 39(1), 73–78 (2009). [CrossRef]
