Intracavity, common resonator, Nd:YAG pumped KTP OPO

Transcription

1 Intracavity, common resonator, Nd:YAG pumped KTP OPO James Beedell* a, Ian Elder a, David Legge a & Duncan Hand b a SELEX Galileo, Crewe Toll House, 2 Crewe Road North, Edinburgh EH5 2XS, UK b School of Engineering & Physical Sciences, Heriot-Watt University, EH14 4AS, UK ABSTRACT Results are presented for generation of 1064 nm and 1573 nm outputs using a common ring resonator for a laser diodeside-pumped zig-zag geometry Nd:YAG laser slab and three NCPM (non-critically phase-matched) KTP crystals. The performance of the resonator at each wavelength is reported, for various configurations. First a common resonator was tested, then a separate resonator was used to gain an understanding of the performance variation with resonator length and OPO output coupling. A second common resonator was then tested, which had an optimized configuration to improve its efficiency. The conversion efficiency of the final design was 35% with 29 mj output at 1573 nm wavelength for 83 mj at 1064 nm. Keywords: Intracavity OPO, Nd:YAG ring resonator, common ring resonator, shared cavity, KTP OPO 1. INTRODUCTION Pulsed laser sources operating in the eye safe wavelength region around 1.5 µm are used in a number of applications, including rangefinding, active imaging (or burst illumination) and remote sensing. This waveband also offers excellent atmospheric transmission for long-range applications; in general, pulse energies in excess of 10 mj are required for the long range applications. One of the most common routes for generation of large pulse energies at 1.5 µm is to use nonlinear wavelength conversion in an optical parametric oscillator (OPO) to shift the output of a Q-switched Nd:YAG laser at 1064 nm to the eye safe waveband. The ideal nonlinear crystal for this OPO is potassium titanyl phosphate (KTP). It has a high damage threshold suitable for generation of large pulse energies, and shifts 1064 nm from Nd:YAG to 1573 nm in a noncritically-phasematched (NCPM) geometry, which is insensitive to both the angle and temperature of the KTP, hence ideal for use in fieldable laser systems. From previous work at SELEX Galileo, it has been shown that the optical-tooptical conversion efficiency for an extracavity Nd:YAG pumped KTP OPO generating 1573 nm is typically 35%. The efficiency of the OPO is dependent on the intensity of the pump (fundamental) beam. By using an intracavity geometry for the OPO the maximum pump intensity is available, thereby lowering the OPO threshold and enhancing conversion efficiency. The OPO acts as the output coupler for the pump laser. In a low loss (or high gain) design the intracavity OPO will couple out a large fraction of the equivalent fundamental 1064nm laser mode pulse energy. As well as enhancing the conversion efficiency, the intracavity OPO source also has a smaller footprint than the equivalent extracavity OPO arrangement with fewer optical components. There have been a number of published works on intracavity OPOs based on Q-switched Nd:YAG and KTP. 40 mj from a flashlamp-pumped Nd:YAG intracavity KTP OPO, at a conversion efficiency of 0.6% from the input electrical energy to the flashlamp, has been reported [1]. A compact flashlamp-pumped intracavity OPO source which generated 10 mj pulses of duration 1.1 ns has also been demonstrated [2]. Higher efficiencies have been achieved using diodepumped Nd:YAG; 97 mj pulse energies at pulse repetition rates up to 30 Hz, with pulsewidths of 3.6 ns, have been reported from a diode side-pumped Nd:YAG rod laser with a KTP intracavity OPO at 4.2% wall plug efficiency [3]. Also, a conversion efficiency of 10.4% with respect to diode-laser input energy has been reported for a diode side-pumped Nd:YAG rod laser with an intracavity KTP OPO producing an output of 31.5 mj, 1572 nm signal at 10 Hz [4]. *james.beedell@selexgalileo.com; phone ; selexgalileo.com Electro-Optical and Infrared Systems: Technology and Applications IX, edited by David A. Huckridge, Reinhard R. Ebert, Proc. of SPIE Vol. 8541, 85410Q 2012 SPIE CCC code: /12/$18 doi: / Proc. of SPIE Vol Q-1

2 Very few publications could be found where the OPO shared a common resonator with the pump laser. In one report [5] a comparison is made between shared and coupled OPO cavities in a diode-pumped Nd:GdVO 4 laser. The efficiency of the 60 mm long end-pumped shared resonator is reported as 75% of the coupled resonator efficiency. However in this case the laser is passively Q-switched, so is different to an actively Q-switched laser, where externally modulation of the Q-switch can be applied and is independent of the intracavity power density. Residual losses in the Q-switch are significant in a passively Q-switched laser and may affect the efficiency comparison. In addition, this paper compared a 25 mm long coupled resonator with a 60 mm long shared resonator. It is well known that lengthening of an OPO resonator will extend the build-up time and reduce its efficiency. These quoted results all use a short OPO resonator inside the pump laser resonator. One of the primary reasons for this is to reduce the build-up time and duration of the OPO pulse relative to the pump pulse. These results are also for designs where only the OPO wavelength is output. Here we present results for a common, or shared, resonator intracavity OPO design based on laser diode pumped Nd:YAG and KTP. With this design the output can be rapidly switched between 1064 nm and 1573 nm via electrical control of the intracavity polarisation states. These outputs will have a common boresight due to the intracavity mode at each wavelength sharing a common beam path. The common boresight of the two wavelengths during harsh environmental conditions is a key parameter for a real world laser source switching between the two wavelengths. The initial common resonator design fell short of the typical 35% conversion efficiency of an extracavity OPO. To help understand the effect of resonator length and OPO output coupling, an intracavity OPO which used a separate resonator was constructed by introducing 1573 nm fold optics into the resonator. Finally, we present an optimized shared resonator configuration which has lower intracavity loss at 1573 nm and a more suitable waveplate arrangement to allow better control of the two resonating wavelengths. To the authors knowledge, this is the first reported demonstration of a common resonator, intracavity KTP OPO inside an image rotating Nd:YAG ring laser. 2. FIRST COMMON RESONATOR SET-UP Lasers and OPOs for use in fieldable systems where severe environmental conditions will be experienced require resonator designs that are insensitive to alignment perturbations. Resonator designs that are less sensitive to misalignments include a resonator using cross-porro prisms and polarisation output coupling [6]. In this design the selfaligning feature of the Porro prisms is exploited in both axes of the propagating beam by orienting the apex of each Porro prism orthogonally. This design is usually folded to reduce the footprint of the resonator using a corner cube, which also rotates the beam image by 180. A cross-porro prism resonator is essentially a ring resonator where the propagating beam is folded back on top of itself so that the laser slab is passed twice per round trip. Therefore a ring resonator that rotates the propagating beam by 90 will provide the same misalignment tolerance as a cross-porro prism resonator with the same beam rotation. When any element is misaligned by a small amount, the resonator mode can always respond by making small changes in beam position and direction to find a new closed and aligned path [7]. Here a ring resonator was used to produce a 1064 nm Nd:YAG laser with 45 prisms folding the beam around. The 45 prisms were used in this proof of concept design due to their cost and availability as commercial off-the-shelf components. The right hand side of the resonator used two prisms to fold the beam whilst the left hand side of the resonator utilised three prisms to give 90 image rotation. The common resonator intracavity OPO was constructed, as shown schematically in figure 1. The physical round trip length of this resonator was 766 mm. The 90 per round trip rotation of the propagating beam makes the resonator design more resilient to perturbations of the optics. This also produces a more symmetrical laser beam as any thermal lensing in the laser slab is averaged between the x and y axes. Proc. of SPIE Vol Q-2

3 λ/2 plate (achromatic) λ/2 plate nm) λ/4 plate (achromatic) Q-switch (LiNbO 3 ) Polarising beam splitter Output 1064 nm &1573 nm Nd:YAG zig-zag slab 3 KTP crystals (NCPM) Laser diode pumping from side Retroreflector Fig. 1 Layout of the first intracavity, common resonator, KTP OPO and Nd:YAG pump laser. All the optical components within the ring resonator were AR coated at both 1064 nm and 1573 nm wavelengths to allow both wavelengths to resonate. Two exceptions to this were the Nd:YAG laser slab which was uncoated on its end faces and the LiNbO 3 Q-switch which was only AR (anti-reflection) coated at 1064 nm. The end faces of the Nd:YAG slab are cut at 59 from the normal, to allow zig-zagging of the laser mode with a maximized fill factor and therefore pump extraction. The Nd:YAG slab, with a refractive index of 1.81 at 1573 nm has a Brewster s angle of Therefore almost all of the p-pol radiation is transmitted by the surface of the laser slab. A measurement of the uncoated slab 1573 nm transmission was found to be 86%. This loss can be attributed to absorption within the bulk material of the Nd:YAG crystal. The fill factor is not optimized at 1573 nm, so the OPO mode will be reduced in size from 4.75 mm to 4.5 mm. The Nd:YAG slab had a cross-section of 5 mm by 5 mm, and allowed nine bounces of the laser mode. The Nd:YAG slab was pumped from one side using five laser diode stacks, which had a wavelength of 808 nm, and each had dimensions 5 mm by 10 mm. The laser diodes were operated in a pulsed regime of 150 µs pulses at a pulse repetition rate of 10 Hz. While using a laser diode current of 96 A, this produced a pump pulse energy of 675 mj. Three KTP crystals were then inserted into the resonator after the Nd:YAG slab so that the 1064 nm laser output/opo pump and the 1573 nm OPO signal could follow the same path around the ring resonator. The KTP crystals had the dimensions, 9 mm square cross-section and 20 mm long. The KTP crystals were non-critically-phase-matched and therefore no walk-off of the 1064 nm and 1573 nm was introduced. This is a Type II interaction, therefore the 1573 nm radiation produced by the KTP crystals was of the same vertical linear polarization state as the 1064 nm pump incident at the crystals. The achromatic wave plates in the resonator were adjusted to ensure that the 1064 nm laser was in hold-off mode, before the Q-switch operated. The half wave plate at 1573 nm was used in an attempt to separate the polarization states of the 1064 nm and 1573 nm beams, so that the 1573 nm output coupling could be adjusted for optimum performance when the 1064 nm output coupling was maximized. The output coupling for both wavelengths was for the s state polarized light incident on the polarizing beam splitter and the percentage of s state polarized light was controlled using the Q-switch voltage. The retroreflector couples back the counterpropagating beam, which enhances the gain in the forward direction, selecting it preferentially for unidirectional operation. Proc. of SPIE Vol Q-3

4 3. PERFORMANCE OF THE FIRST COMMON RESONATOR The output of the 1064 nm and the 1573 nm was recorded while the Q-switch voltage was increased and is shown in figure 2. The conversion efficiency from 1064 nm to 1573 nm was calculated by measuring the maximum 1064 nm output of 90 mj while the KTP crystals were rotated by 90 to prevent conversion to 1573 nm. Energy (mj) Energy, 1064 nm Energy, 1573 nm Conversion Efficiency Q-switch voltage (kv) 30% 25% 20% 15% 10% 5% 0% Conversion efficiency Fig. 2 Output at 1064 nm and 1573 nm for a range of Q-switch voltages and the conversion efficiency from 1064 nm to 1573 nm. The maximum conversion efficiency was 27% with a 1573 nm output of 24 mj from a 1064 nm pump pulse energy of 90 mj. The stability of the two outputs were recorded over a duration of 5 minutes at the Q-switch voltages 2 kv for 1064 nm output and 5.2 kv for 1573 nm, as shown in figure 3. The average 1064 nm energy was 87 mj with a standard deviation of 2 mj and the average 1573 nm energy was 24 mj with a standard deviation of 0.7 mj Energy (mj) nm wavelength 1573 nm wavelength Time (s) Fig. 3 Output at 1064 nm and 1573 nm over a firing duration of 5 minutes. Proc. of SPIE Vol Q-4

5 The far field beam profile was imaged using a CCD camera at the focal plane of a 1 m focal length spherical mirror as shown in figure 4. Fig. 4 Far field beam profile of the 1573 nm output. The divergence calculated from this beam image was approximately 8.5 mrad with a standard deviation of 0.25 mrad. The standard deviation of the 1573 nm centroid position was 27 µrad in the x axis and 28 µrad in the y axis. In all cases, the far field beam image size is measured using the 90% encircled area calculated in the Spiricon BeamGage beam profiling software. The maximum conversion efficiency of 27% for this configuration is lower than the typical conversion efficiency of 35% for an extracavity KTP OPO configuration. Therefore, a separate resonator was constructed to allow an analysis of the conversion efficiency to be undertaken whilst avoiding high loss optics in the OPO resonator. In addition, this allowed the control of the output coupling at 1573 nm and the OPO resonator length independently of the laser resonator length and output coupling. 4. SEPARATE RESONATOR SET-UP A separate resonator configuration was constructed as shown in figure 5. λ/2 plate (achromatic) λ/2 plate (@1573 nm) λ/4 plate (achromatic) Q-switch (LiNbO 3 ) Polarising beam splitter Output 1064 nm OPO folding mirrors Nd:YAG zig-zag slab 3 KTP crystals (NCPM) Laser diode pumping from side 1573 nm 1573 nm Retroreflector Fig. 5 Layout of diode-pumped Nd:YAG ring laser with intracavity KTP OPO using a separate OPO resonator. The separate OPO resonator utilized two 1573 nm folding mirrors in the laser ring resonator, and a pair of plane mirrors to form the OPO resonator. The output coupler (49%R shown in figure 5) was positioned such that any undepleted 1064 nm pump radiation reflected by the 1573 folding optic would be rejected out through the back mirror of the OPO, Proc. of SPIE Vol Q-5

6 which was highly reflecting at 1573 nm. Due to refraction in the OPO folding optics, the position of the KTP crystals was translated to prevent aperturing of the 1064 nm laser mode. The round trip physical length of the separate resonator was varied in the range 400 mm to 1000 mm. 5. PERFORMANCE OF THE SEPARATE RESONATOR The performance of the separate resonator was tested for four physical round-trip resonator lengths as shown in figure 6, which were 420 mm, 614 mm, 766 mm and 1002 mm. Maximum 1573 nm pulse energy (mj) Physical round trip resonator length (mm) Fig. 6 Dependence of maximum 1573 nm pulse energy on the physical round trip resonator length. A drop in 1573 nm energy from 39 mj to 19 mj is shown for an increase in physical round trip resonator length from 420 mm to 1002 mm. A physical round trip length of 766 mm was tested, which matched the common resonator physical round trip length and approximately matched the OPO build-up time. This separate resonator length produced 1573 nm with an output energy of 25 mj, which was comparable to that produced by the common resonator. This was not expected as although the build-up time was approximately matched, the separate resonator was subject to reduced intracavity loss at 1573 nm, which was mainly due to the Q-switch and the laser slab in the common resonator. This could be explained by cavity length resonances [8], as the optical round trip time would have been more precisely matched in the common resonator. Only the dispersion in the optics would have slightly altered the relative lengths of the two beam paths. For the shortest resonator physical round trip length of 420 mm, three output couplers were tested. As shown in figure 7, the variation in maximum 1573 nm output energy is small for the three output couplers used. The conversion efficiencies for the 36%, 49% and 61% output couplers were 42.7%, 43.4% and 39.2% respectively. The best performance was achieved using the 49% reflectivity output coupler, which produced an output of 39.1 mj. Proc. of SPIE Vol Q-6

7 45 40 Max. pulse energy (mj) % 35% 40% 45% 50% 55% 60% 65% Output coupling Fig. 7 Dependence of maximum 1573 nm pulse energy on the level of OPO output coupling. Fig. 8 Far field beam profile of the 1573 nm output. The far field beam profile for the 766 mm physical round trip length separate resonator was imaged using a CCD camera at the focal plane of a 1 m focal length spherical mirror and can been seen in figure 8. The divergence was calculated to be approximately 6.1 mm with a standard deviation of 0.5 mrad. The standard deviation of the 1573 nm centroid position was 47 µrad in the x axis and 43 µrad in the y axis. 6. OPTIMISED COMMON RESONATOR SET-UP An optimized set-up of a common resonator intracavity OPO was set out as shown schematically in figure 9. The limitation of the first common resonator configuration was that the polarization state of the 1573 nm was not readily known, because the polarization states of the two wavelengths could not be adjusted independently. In the optimised setup, whole wave plates at 1064 nm and 1573 nm were used to provide some control of one wavelength without affecting the other. This time, the 1064 nm was set for hold-off while the Q-switch was unpowered, while the 1573 nm wavelength was set to be unaffected by the powered Q-switch for a more predictable operation. Proc. of SPIE Vol Q-7

8 λ/2 plate 1λ plate (achromatic) (1573 nm) λ/4 plate (achromatic) Q-switch (RTP) 1λ plate (1573 nm) 1λ plate (1064 nm) Polarising beam splitter Output 1064 nm &1573 nm Nd:YAG zig-zag slab 3 KTP crystals (NCPM) Laser diode pumping from side Retroreflector Fig. 9 Layout of the second intracavity, common resonator, KTP OPO and Nd:YAG pump laser. In addition, the LN Q-switch which had a loss of 11% at 1573 nm was swapped for a dual wavelength AR coated RTP Q-switch which had loss of 4%. This reduction in intracavity loss is responsible for a reduced threshold in figure 8 when compared with figure 2. In figure 8, some 1573 nm output is produced before the 1064 nm output reaches its maximum value. In this configuration, the only major intracavity loss at 1573 nm was due to the loss in transmission through the Nd:YAG crystal. 7. PERFORMANCE OF THE OPTIMISED COMMON RESONATOR The conversion efficiency from 1064 nm to 1573 nm was calculated by recording the maximum 1064 nm output of 83 mj while the KTP crystals were rotated by 90 to prevent conversion to 1573 nm. In Figure 10, the maximum output of 29 mj at 1573 nm is shown, which represented a conversion efficiency of 35%. Proc. of SPIE Vol Q-8

9 Pulse Energy (mj) nm Energy 1573 nm Energy Conversion efficiency Conversion Efficiency (%) Q-switch Voltage (kv) 0.00 Fig. 10 Output at 1064 nm and 1573 nm for a range of Q-switch voltages and the conversion efficiency from 1064 nm to 1573 nm. In this set-up, there was less electronic control to select a single output wavelength, than for the first common resonator set-up. A maximum 1573 nm output of 29 mj was available at 3.2 kv and 10 mj of 1064 nm leaked out through the polarising beam splitter. The maximum 1064 nm output was 71 mj at 1.2 kv where 6 mj of 1573 nm was also coupled out. The threshold for 1573 nm conversion was lower in this set-up than for the first common resonator configuration. An energy curve was plotted in figure 11 to show the dependence of 1573 nm pulse energy on the pump diode pulse energy nm pulse energy (mj) Diode pulse energy (mj) Fig. 11 A plot of output average power at 1573 nm wavelength against laser diode pump power. As shown in figure 11, the threshold for 1573 nm output is a diode pulse energy of 335 mj. The energy curve flattens off near the maximum pump energy. With more pump energy available, the output coupling of the 1573 nm radiation could again be optimized for maximum 1573 nm output and this energy curve would have a different slope and threshold. The far field beam profile was imaged using a CCD camera, this time at the focal plane of a 0.5 m focal length spherical lens as shown in figure 12. Proc. of SPIE Vol Q-9

10 Fig. 12 Far field beam profile of the 1573 nm output. The 1573 nm beam and the 1064 nm beam were co-boresighted to within 250 µrad. The divergence of the 1573 nm output, when the Q-switch voltage was set to 3.2 kv for maximum output, was 11 mrad with a standard deviation of 0.7 mrad. The standard deviation of the 1573 nm centroid position was 55 µrad in the x axis and 25 µrad in the y axis. The co-boresight offset was therefore only 2% of the 1573 nm wavelength output divergence. 8. CONCLUSIONS In this paper, we have investigated a common resonator intracavity OPO based on Nd:YAG and KTP. A separate resonator intracavity OPO was also set-up to compare the resonator designs. An optimized common resonator was then investigated. The conversion efficiency of the optimized common resonator configuration was 35%, which was an improvement on the 26% efficiency of the earlier common resonator configuration that had higher intracavity loss. The optimised common resonator efficiency also matched the efficiency of a typical extracavity OPO. As the extracavity would typically have a much shorter physical round-trip length of approximately 200 mm, the build-up time would be approximately four times shorter than for the common resonator, which would improve its efficiency. The advantage of the common resonator design, which effectively off-set this effect was the increased intracavity 1064 nm power density within the intracavity OPO. The overall conversion efficiency from electrical input to the pump diodes, to the 29 mj of 1573 nm output was 2%, for the optimised common resonator set-up. The co-boresight offset was 2% of the 1573 nm wavelength output divergence, which was 11 mrad. The optimised common resonator set-up physical round trip length of 804 mm could ultimately be reduced to an estimated length of 430 mm, to reduce the build-up time of the laser and the OPO. This would further boost the efficiency of the system and reduce the footprint. This is a recommendation for extension to this work, so that the effect of cavity length on the divergence and efficiency in the common resonator can be investigated. Proc. of SPIE Vol Q-10

11 REFERENCES 1. Raevsky, E.V., Pavlovitch, V.L. & Konovalov, V.A., Efficient eye-safe intracavity KTP optical parametric oscillator, Proc. of SPIE 4360, 75 (2002) 2. Mason, P.D. & Perrett, B.J., High-energy, sub-nanosecond pulse duration intracavity pumped KTP OPO at 1572 nm, Advanced Solid-State Photonics, paper MB9 (2007) 3. Liu, X., Lu, C, Wang, X., Sun, B., Cheng, Y. & Chen, J., Efficient intracavity optical parametric oscillator with diode-side-pumped electro-optic Q-switched laser, Chinese Optics Letters 4(11), 664 (2006) 4. Y. Y. Wang, K. Zhong, D. G. Xu, P. Wang, and J. Q. Yao, High-energy pulsed eye-safe intracavity optical parametric oscillator at 1.57 μm Proc of SPIE, 7276, (2009). 5. Y. F. Chen and L. Y. Tsai, Comparison between shared and coupled resonators for passively Q-switched Nd:GdVO4 intracavity optical parametric oscillators, Applied Physics B: Lasers & Optics, vol. 82, pp (2006). 6. W. Koechner, [Solid-State Laser Engineering], Springer New York (2006). 7. A. E. Siegman, [Lasers], OUP (1986). 8. M. Henriksson, L. Sjöqvist, V. Pasiskevicius, and F. Laurell, Cavity length resonances in a nanosecond singly resonant optical parametric oscillator Optics Express 18, (2010). Proc. of SPIE Vol Q-11