High-efficiency and compact blue source: intracavity frequency tripling by using LBO and BBO without the influence of birefringence
Introduction
In recent years, blue and green lasers have attracted considerable attention. The short wavelengths of these lasers have found applications in laser ablation, laser cleaning, and precision material processing, i.e., semiconductor wafer drilling and marking, diamond film cutting, removal of gold, and the scribbling of photovoltaic panels. Second-harmonic generation (SHG) and third-harmonic generation (THG) techniques are the most common methods of producing green and blue laser outputs from powerful Nd:YAG lasers. Intracavity SHG is an efficient method of extracting the second-harmonic output from a continuous wave (cw) or a low-peak-power pulsed laser. 1-3 The output of these SHG or frequency-tripled lasers is suitable for many of the above-mentioned industrial applications. There has been a tremendous amount of research devoted to obtaining high THG conversion efficiencies.4-6 Unfortunately, in most situations, the higher conversion efficiency was obtained only from high-peak-power laser systems. Typically the power density of these systems is in the range of (or greater than) 1 GW/cm2. Most industrial laser systems do not approach this power density.
|
The author is with Kigre Inc., 100 Marshland Road, Hilton Head Island, South Carolina 29926. Received 1 July 1992. 0003-6935/93/060971-05$05.00/0. © 1993 Optical Society of America. |
Similar to the intracavity SHG, intracavity frequency tripling is an efficient method of producing the third-harmonic output in solid-state lasers, particularly in cw or low-peak-power applications in which the peak power of the laser beam is too low for an efficient frequency conversion external to the resonator. The conversion efficiency of nonlinear processes may be increased substantially if the nonlinear event takes place internal to the laser resonator, where the fundamental laser intensity is several orders of magnitude higher than external to the resonator.7 In intracavity frequency tripling, the THG crystal mixes the fundamental laser intensity with the second harmonic, which is produced by a SHG crystal as an extraordinary (E) ray, regardless of what type of SHG phase-matching condition is utilized. In order for efficient THG to occur, the sum of the fundamental and second-harmonic laser intensities must be high. In intracavity frequency tripling, the intensity of the second harmonic is relatively low; however, the high intensity of the fundamental allows for efficient THG.
In this paper a high-efficiency intracavity frequencytripled Nd:YAG laser that uses lithium triborate (LBO) and barium borate (BBC)) nonlinear crystals is described. The laser output is completely free of the influence of birefringence from the two nonlinear crystals, and no interference intensity variations are observed with independent tuning of the SHG or THG crystals. Two hundred mW of a single-mode third-harmonic output at 354.7 nm is demonstrated
/
by using a Nd:YAG laser with a 3 mm x 3 mm x 6 mm type-I LBO crystal for SHG and a 6 mm x 6 mm x 6 mm type-I BBO crystal for THG. The average input power to the flash lamp is 313 W, which corresponds to 1.25 J/pulse at 250 Hz.
2. Basic Theory and Experiments
The basic theory of intracavity THG is quite different from normal external THG. Generally, efficient frequency tripling depends on the fundamental and second-harmonic photons in a ratio of 1:1 over a broad intensity range.8 But this is not the case for intracavity THG since the fundamental oscillating beam is much stronger than the second-harmonic beam. Normally the ratio is in the range 1:3—1:10.
When birefringent nonlinear crystals are inserted into a laser resonator, particularly in the case of intracavity frequency tripling with two nonlinear crystals in the same laser resonator, there are some problems with the intracavity polarization of the fundamental laser beam. Generally, when these crystals are introduced into the laser resonator, they will depolarize the polarized laser resonator or cause a change in the ratio between the ordinary (O) ray and the E ray in a nonpolarized laser cavity. Type-I and type-Il SHG crystals have different depolarization effects and require different resonator polarizations. Subsequently they have different effects on the THG performance. According to the type of SHG used, the performances of intracavity THG lasers are as follows:
A. Type-Il SHG
A typical example is the use of a type-Il potassium titanyl phosphate (KTP) crystal as the SHG crystal. Type-Il KTP crystals have high nonlinear conversion efficiencies and are frequently used for the SHG of a Nd:YAG laser operating at 1064 nm 9' 10 Unfortunately the phase-matching condition cannot be obtained for the THG of 1064 nm. For intracavity type-Il SHG, an O ray and an E ray of the fundamental wavelength in the crystal must interact with each other. The existence of these orthogonal components at the fundamental wavelength is the main source of the birefringent interference phenomena. Although there are some techniques for solving this problem, ll they do not work well in frequencytripling applications.
In a previous work, 536 mW of UV output was demonstrated from an acousto-optic Q-switched intracavity THG Nd:YAG laser with 2 kW of average input power. 7 This laser utilized type-Il KTP for SHG and BBO for THG. It also demonstrated that if a type-Il SHG crystal such as KTP or BBO is utilized, any tuning action of any intracavity optic components will cause periodic variations in the intracavity laser flux density at 1064, 532, and 355 nm. The nature of such phenomena is the above-mentioned birefringence interference. This periodic variation makes it extremely difficult to align the resonator. Even in the vicinity of the phase-matching angle of the SHG
972 APPLIED OPTICS / Vol. 32, No. 6 /
and THG crystals, several interference peaks appear, which causes the final tuning to be even more complicated and difficult. The system is also unstable, as any slight thermal or mechanical change in the resonator will cause the output to fluctuate.
Another consideration is the polarization match between SHG and THG. After SHG there are two nearly equal O-ray and E-ray fundamental intensity components and an E-ray SHG component. It would be difficult to use both components of the fundamental intensity fully for further THG because the phases of those two components are spatially random and slightly separated. Neither type-I nor type-Il THG's could fully use the from both fundamental components.
B. Type-I SHG
Type-I SHG requires only an O ray of fundamental intensity. When a type-I SHG crystal is inserted in a polarized laser cavity, the oscillated laser beam will travel in the SHG crystal only as an O ray. Hence all the depolarization problems mentioned above in Subsection 2.A will be avoided because there is no fundamental E ray. Furthermore many THG crystal choices are available. Both type-I and type-Il THG phase-matching methods could be used without any depolarization problem since the 532-nm beam is only a secondary beam that is not oscillating. It is also easier to optimize the polarization matching between the SHG and the THG for this case.
In order to optimize an intracavity THG laser factor such as polarization, the walk-off angle, the acceptance angle, the nonlinear coefficient, the cavity mode size, the beam divergency, and so on, must be considered. As an example, a type-Il THG crystal that just matches the output polarization conditions of the type-I SHG may be utilized. However, when the influence of the walk-off angle, which leads to the separation of the O and the E beams, is considered for most situations, the separation of the SHG and the THG are in orthogonal directions. In intracavity frequency tripling, it is usually desirable to have the smallest beam size possible in the nonlinear crystal in order to maximize the power density of the laser beam so that high conversion efficiencies can be obtained. Considering that the laser beam waist is rather small (in this case, for example, it is less than 1 mm) it is important to pay particular attention to the beam separation. A type-I THG crystal may also be used for cases in which there is no beam separation during the THG process. However, the orientation of the proper polarizations of the laser beams still needs to be solved.
The experimental setup is illustrated in Fig. 1. The YAG laser resonator consists of mirrors 1 and 8, YAG rod 5, polarizing calcite wedge 6, and LiNb03 Brewster-angle Q switch 7. The laser beam's polarization is determined by the polarizer's orientation.
A 3 mm x 3 mm x 6 mm type-I LBO crystal is used as the SHG crystal. Without the THG crystal, the intracavity SHG laser produces 750 mW of multi-
2 3 4 6
8
1
Fig. 1. Resonator block diagram: 1, 8, resonator mirrors; 2, THG crystal; 3, wave plate; 4, SHG crystal; 5, Nd:YAG rod; 6, calcite wedge; 7, Q switch.
mode average output, which corresponds to 3 mJ/pulse at 250 Hz, with pumping at 1.25 J/pulse,
The polarized 1.064-11m laser beam travels in the LBO crystal as an O ray that produces the second harmonic, polarized as an E ray. In our configuration, the beam separation of the O ray and the E ray is 36 11m. Although a type-Il BBO crystal with 9 = 38.60 could match the output of the SHG by properly selecting the orientation of the BBO THG crystal, in which the 532-nm beam would travel as an O ray and the 1.064-11m beam would travel as an E ray, it is preferred to use a type-I BBO crystal with 6 = 31.30. In order to match the output polarization of the SHG, a wave plate (item 3 in Fig. 1) is used. This wave plate rotates the polarization of the fundamental beam 900 while maintaining the polarization direction of the SHG beam. Finally, both the fundamental and the SHG beams pass through the THG crystal as O rays. Experimental results confirm the above considerations. Following perturbation of the SHG or THG crystals, no intensity interference variations appear when a type-I SHG method is used.
3. Analysis
LBO is a moderately efficient nonlinear harmonic generation crystal with a low angular sensitivity for harmonic generation. 12-15 Table 1 lists the SHG characteristics of LBO and BBO. 16 With an LBO crystal that is only 6 mm long, the intracavitydoubled SHG output reached 75% of the power obtained with a 7-mm-long BBO crystal. The larger acceptance angle and the smaller walk-off angle, both of which are approximately 1 order of magnitude greater than that of BBO, are important for further THG. This was the primary reason for selecting LBO as the SHG crystal for the intracavity-tripling experiments.
In Table 2 a comparison of type-I and type-Il BBO THG crystals is presented. From this table, it can be seen that the advantages of type-I THG include the greater effective nonlinear coefficient. Although the walk-off angles are similar, there are no additional beam separations in the 1.064-pm and 532-nm beams because of the type-I phase-matching method. There is only a slightly larger acceptance angle for type-Il (E + O E) frequency tripling, which may yield an increased output. However, there is still a potential problem with the beam separation because the 4.50 walk-off angle corresponds to a 0.54-mm beam separation for a piece of a 7-mm-long BBO type-Il crystal. The final frequency-tripling results should be better with a type-I frequency-tripling crystal if there are no special measures to prevent beam separation in a single-mode operation. The other type-Il (O + E —+ E) frequency-tripling method has a slightly larger acceptance angle and a smaller walk-off angle. The beam separation is also in the same direction as the SHG beam when the type-I BBO crystal is utilized for SHG. If the two crystals are properly oriented, the separations of THG and SHG can compensate each other. Unfortunately the nonlinear coefficient for this type of phase-matching method is considerably smaller.
The beam quality of the UV output has been examined with a beam imaging system. The use of this system illustrates the influence of the phasematching factor [sinc(dkl/2)] clearly. The beam divergence of the second-harmonic beam was measured first. In Table 3, the results of the beam divergency measurements are summarized. The SHG acceptance angle of the type-I LBO crystal (13.7 mrad for a 6-mm-long LBO crystal) is much larger than the fundamental beam divergence, even in a multimode operation. The data in Table 3 indicates that the SHG beam divergence has a circular distribution that is determined by only the resonator mode structure and alignment. The green beam from the BBO SHG has an elliptical distribution.
|
Tal. BO and LBO SHG Performance Comparison
Accept Angle Temperature Walk-Off Acceptance Angle
aDeff, effective nonlinear coefficiency. / |
The UV beam divergence was also measured as nonsymmetrical (1.8 x 0.9 mrad). The 1.8-mrad diTable 2. BBO Type-I and Type-Il THG Performance Comparisona
Phase-Matching Walk-Off Accept
Angle Angle Angle Def
Type Polarization (90) (90) (mrad cm) (10-9 esu)
aThe 1.064-, 0.53-, and 0.355-11m wavelengths (Ll, 142, and L3, respectively) are present in all cases.
vergence is in the non-phase-matching direction and depends on only the divergence of the 1.064-11m and 532-nm beams. The 0.9-mrad divergence is in the same plane as the optical axes of the THG crystal; thus the beam divergence in this axis is limited by the acceptance angle of the THG crystal (0.75 mrad FWHM for a 6-mm BBO crystal).
The nonsymmetrical output distribution of the UV indicates that there is a rather large portion of energy that does not convert into LTV because the resonator mode does not match the acceptance angle of the nonlinear crystal. It has been suggested that the UV output efficiency could be increased further by using another THG crystal such as LBO, which has an acceptance angle that is 10 times larger than the BBO crystal. Increasing the acceptance angle also allows for a more symmetrical and higher-quality UV beam. Preliminary experimental results of using LBO for the THG crystal confirm this theory. 1.3 W of multimode UV and an 86% conversion efficiency from 532 to 355 nm were obtained. The output peak power reached 0.59 MW. The author is currently preparing details of this experiment for future presentation.
In order to measure the conversion efficiency of the THG, a special output coupler was fabricated. This output coupler had transmissions at 1.064 pm, 532 nm, and 355 nm of 0%, 90%, and 36%, respectively. The 532-nm output power was measured with the THG crystal far from its phase-matching angle. The output powers at 532 and 355 nm were measured separately, with the 355-nm output maximized by using a filter with a transmission at 532 and 355 nm of 0% and 90%, respectively. 136 and 256 mW of power were measured at 355 and 532 nm, respec-
Table 3. Results of the Beam Divergence Measurementsa
SHG Beam THG
(mrad) Beam (mrad)
|
LBO (6 mm) |
Single |
2.3 |
2.2 |
|
|
|
|
Multi |
5.0 |
4.6 |
|
|
|
BBO (7 mm) |
Single |
2.1 |
1.6 |
|
|
|
|
Multi |
4.5 |
2.0 |
|
|
|
LBO as SHG, BBO as THG |
Single |
|
|
1.8 |
0.9 |
tively, yielding a 35% 532-nm photon conversion and a 56% energy conversion.
4. Conclusion
In summary, a polarized Q-switched resonator in which frequency tripling was realized by using two nonlinear crystals inserted in the laser resonator without the instability of birefringent interference was demonstrated. Currently it is more advantageous to use a type-I SHG method for intracavity tripling since the problems that deal with the usage of type-Il SHG can be avoided. Although the KTP crystal, which has a higher SHG conversion efficiency, was not used, the results are still quite impressive. When the laser is aligned in the TEMoo mode, 200 mW of UV output were obtained. Approximately 35% of the green photons were converted into UV. The pulse width was measured to be 17, 12, and 9 ns for 1.06 pm, 532 nm, and 355 nm, respectively. The peak power of the UV reached 90 kW at a repetition rate of 250 Hz.
Recent experiments in which a 355-nm 1.3-W average power output beam with a 10-ns pulse width and 0.59-MW peak power were obtained at a 220-Hz repetition rate further confirm that this method of intracavity frequency tripling is a compact and efficient method of producing a strong UV output with high peak power and high average power. These experiments should prove useful for many applications including applications in the laser chemistry and semiconductor industries.
