A high-efficiency optically pumped vertical-external-cavity surface-emitting laser emitting 20 W at a wavelength around 588 nm is demonstrated. The semiconductor gain chip emitted at a fundamental wavelength around 1170-1180 nm and the laser employed a V-shaped cavity. The yellow spectral range was achieved by intra-cavity frequency doubling using a LBO crystal. The laser could be tuned over a bandwidth of ~26 nm while exhibiting watt-level output powers. The maximum conversion efficiency from absorbed pump power to yellow output was 28% for continuous wave operation. The VECSEL’s output could be modulated to generate optical pulses with duration down to 570 ns by directly modulating the pump laser. The high-power pulse operation is a key feature for astrophysics and medical applications while at the same time enables higher slope efficiency than continuous wave operation owing to decreased heating.
© 2014 Optical Society of America
Practical and cost-effective yellow lasers are needed in many important medical applications including dermatology, eye surgery and novel imaging methods. In these applications, overall results can be improved, damage to healthy tissue decreased, or resolution increased by implementing a laser that can be tuned to a preferred wavelength and work in pulsed operation instead of continuous wave [1–3]. The yellow spectral range is particularly interesting in medicine due to increased interaction with hemoglobin in blood, which has high absorption peaks in this range. In addition, some of the most common fluorescent markers often used in fluorescence based medical imaging have their depletion wavelength in the yellow range . For medical applications pulsed operation is often preferred due to increased efficiency and decreased damage to healthy tissue; for example in photocoagulation based eye surgery the ideal pulse width is in the order of 1 µs . Furthermore, the availability of reliable yellow lasers able to operate in pulsed mode would impact scientific application areas, such as astronomy, where lasers emitting narrow linewidth at the sodium absorption line (589nm) can be used to create a laser guide star for earth-based telescopes .
The yellow spectral range cannot be reached via direct emission from semiconductor lasers, which is the preferred laser technology when taking into account compactness, cost, efficiency, reliability, and wavelength coverage. The shortest wavelength in the orange region recently demonstrated for direct emission is 599 nm yet the efficiency of the GaInP material system used in this case is rather modest . Alternative yellow laser solutions make use of amplification and frequency conversion in solid-state systems  but the complexity, price, and often their limitations in power and wavelength coverage, render them unsuitable for a wide exploitation. A much more attractive path to generate yellow radiation has emerged with the development of vertical-external-cavity surface-emitting lasers (VECSELs), also known as semiconductor disk lasers (SDLs) . These are compact, power scalable laser sources that are able to maintain good beam quality even when emitting output powers in excess of several watts to several tens of watts [9,10]. Owing to the wavelength versatility of semiconductor gain region they can cover an extremely large emission spectrum by direct emission from 670 nm up to 5000 nm [11–14], yet not without gaps, and can be tuned over tens of nanometers [15,16]. Moreover, unlike the solid state disk lasers, owing to the lower carrier lifetimes the output of VECSELs/SDLs can be modulated on a time scale of a few hundreds of nanosecond by directly modulating the pump laser that at the same time can alleviate the need for advanced cooling . VECSELs emitting in the 1140-1260 nm range can be efficiently converted to yellow region using intra-cavity nonlinear crystals  however their output powers and their conversion efficiency has been rather modest compared to VECSELs operating at 1064/532 nm. This is to large extent due to the fact that longer wavelength gain regions bring more challenges compared to standard InGaAs/GaAs gain mirrors used for 960-1060 nm window.
In this article, we demonstrated a frequency doubled yellow VECSEL emitting 20 W output power with moderate cooling. The measured output power is an improvement of over 12 W over the previously reported yellow VECSEL . Furthermore, the conversion efficiency from absorbed pump power to yellow output power has been improved from 17% to 28%. For GaInAs QW gain material the previous state of the art results were 5 W of 589 nm radiation with ~16% conversion efficiency . Frequency doubled yellow radiation has also been demonstrated with QDs, but with significantly lower output power and even lower efficiency (~5%) compared to QW structures .
Besides the CW operation, we have also implemented a pulse modulation scheme using direct modulation of the pump lasers. The demonstrated pulse duration in the range of 1µs is a good fit to requirements pertinent to medical and guide star applications. Even if the pulse duration is rather long, it could still ease the thermal load and provided further means to increase the power and improve the overall system efficiency.
2. Experimental setup
The semiconductor gain mirror was grown by molecular beam epitaxy (MBE) with an active region that incorporated 10 GaInAs/GaAs/GaAsP quantum wells grown on top of a 25.5-pair AlAs/GaAs distributed Bragg reflector (DBR). The structure is illustrated in a more detailed manner in Fig. 1. The gain structure was designed to be anti-resonant at 1180 nm. For efficient thermal management, the gain mirror was diced into 2.5 mm x 2.5 mm chips that were capillary bonded to a wedged (2°) intra-cavity CVD diamond heat spreader, which was attached to a water-cooled copper mount with indium foil. The outer surface of the heat spreader was antireflection coated for the designed 1180 nm emission. The operation of the gain material at the fundamental wavelength has already been tested and the results were published in . A maximum output power of 23 W at ~1180 nm was obtained with a 97% reflecting output coupler.
The VECSEL cavity was formed by the gain mirror, a folding mirror (RoC = 75 mm) and a flat end mirror in a V-shaped configuration with the first arm having a length of 102 mm and the second 47 mm. A 1.5 mm thick birefringent filter (BRF) and a 100 µm thick etalon were placed along the first arm of the cavity to achieve wavelength tuning and linewidth narrowing. For efficient second harmonic generation, a 10 mm non-critically phase matched lithium triborate crystal (NCPM LBO) (and later a 10 mm critically phase matched (CPM) LBO crystal) was inserted near the flat end mirror where the mode waist was situated. The crystal facets were flat and had an antireflection coating for the fundamental as well as for the yellow radiation. Phase matching was achieved via temperature tuning the crystal with a TEC operated copper oven. The cavity configuration is shown in Fig. 2.
All the cavity mirrors were highly reflective for the fundamental radiation. The flat end mirror was also highly reflective for the yellow radiation, however the folding mirror was highly transmissive (R < 5%) for the yellow radiation. Hence the yellow radiation was extracted through it. For the CW experiments, the gain mirror was pumped with a 200 W 808 nm diode laser with a focused spot diameter of about 510 µm. The mode diameter on the gain mirror was approximately 400 µm and inside the LBO crystal it was 230-164 µm. For the pulse modulation we used a lower power pump that could be driven up to peak powers of 71 W by current pulses with amplitudes of ~50 A and pulse durations in the range of 1 µs.
3. High power CW operation
At first, the laser was tested for high power continuous wave yellow output. Several etalon and BRF configurations with varying thicknesses were tested in order to find the optimal pair for stable high power operation, which lead to the selection of 1.5 mm thick BRF and 100 µm thick etalon. We also tested a critically phase matched (CPM) LBO, however the highest power was obtained with the NCPM LBO. For the high power operation measurement the mount temperature was estimated to be in the range of 5 to 8°C. The temperature of crystal oven was set to 38.3 °C; this was optimized according to the emission wavelength of the VECSEL to satisfy the phase matching condition for efficient frequency doubling. The power conversion graph generated from the measured laser output is shown in Fig. 3 (a). A maximum of 20 W of frequency doubled output power was measured for absorbed pump power of about 75 W, which was obtained by subtracting measured reflected pump power (5.13% of incident power) from the incident pump power. The mount temperature at the maximum output power was 8.3 °C. The mount temperature was measured next to the gain chip. The maximum conversion efficiency (absorbed pump power to yellow output) of ~28% was achieved at 16 W of output power, which is an improvement of 11 percentage points compared to state of the art yellow VECSEL results reported so far .
The emission spectrum of the VECSEL, set by the etalon (FSR ~5 nm, bandwidth ~0.37 nm), was centered at 588.1 nm with a linewidth (FWHM) of <0.2 nm, shown in Fig. 3 (b). The beam profile for the yellow radiation was preserved circular, with some distortions, as shown as inset in Fig. 3 (a), through the measurement range. We did a separate measurement for the output beam quality corresponding to 10 W (power level available at the time of measurement due to NCPM LBO crystal degradation, see next page) of output power. The obtained M2 value was < 1.5 in horizontal and vertical directions. A scanning Fabry-Pérot interferometer was used to measure the mode spectrum for the fundamental wavelength. When the laser was tuned to 589 nm, single-wavelength operation was observed for yellow output power of ~10 W, but the operation was unstable.
Next, the 1.5 mm BRF (placed at Brewster’s angle) was rotated around an axis normal to its surface in order to test the tunability of the VECSEL. For this experiment the absorbed pump power was kept constant at 55 W. The corresponding mount temperatures for the chosen absorbed pump power ranged between 9.2 and 13.5 °C depending on the emission wavelength; the temperature increased towards the edges of the tuning bandwidth due to the reduced gain chip efficiency and consequent increased heat generation. The temperature of the crystal oven was optimized for each wavelength separately. This proved to be challenging because the crystal temperature change was large enough to cause the output power to fluctuate. The 100 µm etalon was taken out each time before the emission wavelength was tuned by rotating the BRF in order to avoid a wavelength jump determined by the FSR of the etalon, and put back in for linewidth narrowing after the laser was tuned to a chosen wavelength.
The tuning spectrum is presented in Fig. 4 with each line having a full width half maximum of ~0.2 nm. The highest output power (15 W) was obtained for emission wavelength of about 588 nm. The tuning bandwidth was ~26 nm ranging from about 576 to 602 nm.
To test the operation at higher temperatures, power curves for different gain mirror mount temperatures were measured. The same cavity configuration was used as described earlier except the nonlinear crystal was changed from 10 mm NCMP LBO to a 10 mm CPM LBO. The change was made, because significant deterioration was observed in the NCPM LBO crystal performance during the course of the experiments and a suitable replacement was not available. We believe the crystal degradation was caused by damaged AR coatings on the facets of the NCPM LBO, which was obvious as observed with an optical microscope. The temperature of the CPM LBO was stabilized with the same copper oven that was used for phase matching with the NCPM LBO. The temperature stabilization was necessary because the stray light of the laser, arising from the parasitic reflections of the gain chip and the BRF, and the pump would have otherwise heated the copper oven of the crystal, and consequently, the crystal above its phase matching temperature (room temperature). The oven temperature was kept constant at 23.1 °C and efficient phase matching was achieved by having the CPM LBO critically cut for type I SHG at 1178 nm and having control of the significant crystal tilts.
The power curves were measured for coolant temperatures of 0, 10, 20 and 30 °C, which corresponded to mount temperatures of 4.5-12.8, 13.9-22.1, 22.9-31.2 and 32.2-38.7 °C, respectively. The dependence of the mount temperature on the absorbed pump power is presented in Fig. 5, which shows that the heat load generated by the increasing pump power led to an increase in the mount temperature of the gain mirror. Whereas, the resulting output power for each temperature setting against absorbed pump power is plotted in Fig. 6. We could not achieve the 20 W of output power that was obtained with the NCPM LBO, because of the CPM LBO crystal properties. The two crystals were ordered from different manufacturers, which can already cause a difference of several Watts in the output as observed while testing two NCPM LBOs from different manufacturers with same specifications. In addition, NCPM LBO is expected to yield higher efficiency because it does not suffer from the phenomena of spatial walk-off and is less sensitive to a slight misalignment of the beams. Furthermore, the acceptance angle of the CPM LBO is smaller than that of the NCPM LBO, which could have an impact on the conversion efficiency. The highest output power of 14.6 W was achieved for lowest mount temperature as expected due to the better luminous efficiency of the semiconductor material as the temperature is decreased. Similarly, the output power is expected to drop for the higher mount temperatures, as is observable from Fig. 6. However, for lower pump powers, the output power was higher for some elevated mount temperatures, which can be seen as overlapping of curves in Fig. 6. This phenomenon was most likely caused by better alignment of the laser wavelength with the gain and with the CPM LBO crystal phase matching condition.
4. Pulsed operation
Heating of the gain chip structure by the incident pump power causes thermal rollover and clearly limits the CW operation, as can be seen from the above power curves. In order to overcome the thermal effects and improve conversion efficiency the yellow laser was tested in pulsed mode. The same cavity configuration was used as for the elevated mount temperature experiments, but a pulsed laser was employed for pumping. The LBO crystal temperature was kept near room temperature and the chip mount temperature was measured to be 21 °C throughout the measurements. The pump laser was driven with a driver capable of producing pulse widths up to ~1.5 µs. Two Si biased detectors were used to monitor the light pulses generated by the pump laser and the VECSEL. The pulse waveforms for the pump laser, as well as the yellow output at the maximum operation of the pump laser, are shown in Fig. 7. The respective FWHM pulse widths are 1.64 µs and 1.08 µs. A time delay of 0.66 µs was measured between the pump pulse and the yellow pulse. The reason for the delay is not clear. We suspected that the time delay could be caused by the need for the gain chip to heat up and consequently reach sufficiently low detuning for lasing. However, the delay did not change even when the coolant water temperature was raised to 60°C as would have been expected based on the assumption. This implies that the heating of the gain chip does not play a noticeable role in the pulse on-set delay.
For comparison, power curves for two different pulse widths were measured. First the driver was set to produce 1.5 µs pulses and then to 1.0 µs. The repetition rate was kept at 10 kHz. The peak power of the pulses was calculated by measuring the average power of the output and FWHM pulse width of the pulse waveforms. The peak power measurements for the two different pulse widths are shown in Fig. 8 (a). The CW power curve (average power) is also shown in Fig. 8 (a) for comparison. It can be clearly seen that in pulsed operation higher powers can be achieved due to decreased heating of the semiconductor gain chip. The CW operation shows thermal rollover at around 60 W of absorbed pump power, but the pulsed operation shows a linearly increasing trend, limited by the available peak pump power. Furthermore, the optical-to-optical conversion efficiency of the CW operation at maximum output power was 14% whereas for the pulsed operation the conversion efficiencies were 20% and 21%, for the 1.08 µs and 0.57 µs pulses respectively. The maximum average output power measured for CW operation was 8.5 W and the maximum peak power measured in the pulsed mode was 14.1 W for the 0.57 µs pulse width and 13.8 W for the longer, 1.08 µs pulse width. The maximum average output powers for the pulsed operation were 81 mW and 149 mW, respectively. The pulse width was limited by the long rise time of the driver; at shorter pulse widths the current was cut which decreased the peak pump power. Yellow spectrum for the 0.57 µs pulse is shown in Fig. 8 (b).
High-power efficient operation of a frequency doubled yellow VECSEL has been demonstrated for both CW and pulsed operation. The maximum yellow output power achieved was 20 W for 75 W of absorbed pump power, which yielded an optical-to-optical conversion efficiency of 27%. The maximum conversion of 28% was achieved with a slightly lower output power of 16 W. Wavelength selective components inside the laser cavity allowed to narrow the emission linewidth and tune the operation wavelength of the laser; a tuning bandwidth of ~26 nm was measured. A further narrowing of the linewidth into single longitudinal operation with a different choice of etalon would make the laser suitable for guide star application. During the high power measurements, the VECSEL operated also in single wavelength at 589 nm with ~10 W of output power. However, the operation was unstable.
The operation of the laser was also tested at different gain mirror mount temperatures using another, critically phase matched (CPM) LBO crystal. At mount temperature of 12.8 °C the maximum output power achieved was about 15 W, whereas for the highest temperature, 38.7 °C, it was 7 W. The high power operation even at elevated temperature makes it more attractive for implementation into medical equipment, since it will not require high cooling capacity.
The same CPM LBO was used during pulsed operation, where the thermal effects were reduced so that a thermal rollover was not apparent for the available maximum pump power. The maximum output powers achieved with pulse widths of 0.57 and 1.08 µs were 14.1 and 13.8 W respectively, compared to the CW maximum of 8.5 W in this setup. The power curves for the pulsed operation suggest that even higher powers can be achieved for higher peak pump powers. The pulse widths were limited by the modulation capability of the electronics and pump system time, but the ~1 µs long pulses are already suitable for certain medical application e.g. eye surgery.
This work was financially supported by the EU FP7 project APACOS (315711) and TEKES project Brightlase (230225). The main author would like to acknowledge TES Foundation and Walter Ahlström Foundation for financial support.
References and links
1. O. T. Tan, J. M. Carney, R. Margolis, Y. Seki, J. Boll, R. R. Anderson, and J. A. Parrish, “Histologic responses of port-wine stains treated by argon, carbon dioxide, and tunable dye lasers. a preliminary report,” Arch. Dermatol. 122(9), 1016–1022 (1986). [CrossRef] [PubMed]
2. M. A. Mainster, “Continuous-wave and micropulse 577 nm yellow–orange laser photocoagulation: a laser for all reasons,” Retina Today1–8 (2010).
3. N. Farahani, M. J. Schibler, and L. A. Bentolila, “Stimulated emission depletion (STED) microscopy: from theory to practice,” In Microscopy: Science, Technology, Applications and Education, A. Méndez-Vilas, J. Díaz, Eds. (Formatex Research Center, 2010), pp. 1539–1547.
5. C. E. Max, S. S. Oliver, H. W. Friedman, J. An, K. Avicola, B. W. Beeman, H. D. Bissinger, J. M. Brase, G. V. Erbert, D. T. Gavel, K. Kanz, M. C. Liu, B. Macintosh, K. P. Neeb, J. Patience, and K. E. Waltjen, “Image improvement from sodium-layer laser guide star adaptive optic system,” Science 277(5332), 1649–1652 (1997).
6. L. Toikkanen, A. Härkönen, J. Lyytikäinen, T. Leinonen, A. Laakso, A. Tukiainen, J. Viheriälä, M. Bister, and M. Guina, “Optically pumped edge-emitting GaAs-based laser with direct orange emission,” Photon. Technol. Lett. 26(4), 384–386 (2014). [CrossRef]
7. Y. Yao, Q. Zheng, D. P. Qu, K. Zhou, Y. Liu, and L. Zhao, “All-solid-state continuous-wave frequency doubled Nd:YAG/LBO laser with 1.2 W output power at 561 nm,” Laser Phys. Lett. 7(2), 112–115 (2010).
8. M. Kuznetsov, F. Hakimi, R. Sprague, and A. Mooradian, “Design and characteristics of high power (>0.5-W CW) diode-pumped vertical-external-cavity surface-emitting semiconductor lasers with circular TEM00 beams,” J. Select. Topics Quantum Electron. 5(3), 561–573 (1999). [CrossRef]
9. B. Rudin, A. Rutz, M. Hoffmann, D. J. H. C. Maas, A.-R. Bellancourt, E. Gini, T. Südmeyer, and U. Keller, “Highly efficient optically pumped vertical-emitting semiconductor laser with more than 20 W average output power in a fundamental transverse mode,” Opt. Lett. 33(22), 2719–2721 (2008). [CrossRef] [PubMed]
10. T.-L. Wang, Y. Kaneda, J. M. Yarborough, J. Hader, J. V. Moloney, A. Chernikov, S. Chatterjee, S. W. Koch, B. Kunert, and W. Stolz, “High-power optically pumped semiconductor laser at 1040 nm,” Photon. Technol. Lett. 22(9), 661–663 (2010). [CrossRef]
11. J. Hastie, S. Calvez, M. Dawson, T. Leinonen, A. Laakso, J. Lyytikäinen, and M. Pessa, “High power CW red VECSEL with linearly polarized TEM00 output beam,” Opt. Express 13(1), 77–81 (2005). [CrossRef] [PubMed]
12. S. Ranta, M. Tavast, T. Leinonen, N. Van Lieu, G. Fetzer, and M. Guina, “1180 nm VECSEL with output power beyond 20 W,” Electron. Lett. 49(1), 59–60 (2013). [CrossRef]
13. N. Schulz, M. Rattunde, C. Manz, K. Koehler, C. Wild, J. Wagner, S.-S. Beyertt, U. Brauch, T. Kuebler, and A. Giesen, “Optically pumped GaSb-based VECSEL emitting 0.6 W at 2.3 µm,” Photon. Technol. Lett. 18(9), 1070–1072 (2006). [CrossRef]
15. L. Fan, M. Fallahi, A. Zakharian, J. Hader, J. Moloney, R. Bedford, J. Murray, W. Stolz, and S. Koch, “Extended tunability in a two-chip VECSEL,” Photon. Technol. Lett. 19(8), 544–546 (2007). [CrossRef]
16. C. Borgentum, J. Bengtsson, A. Larsson, F. Demaria, A. Hein, and P. Unger, “Optimization of a broadband gain element for a widely tunable high-power semiconductor disk laser,” Photon. Technol. Lett. 22(13), 78–980 (2010).
17. N. Hempler, J.-M. Hopkins, A. J. Kemp, N. Schulz, M. Rattunde, J. Wagner, M. D. Dawson, and D. Burns, “Pulsed pumping of semiconductor disk lasers,” Opt. Express 15(6), 3247–3256 (2007). [CrossRef] [PubMed]
18. T. Leinonen, V.-M. Korpijärvi, A. Härkönen, and M. Guina, “7.4W yellow GaInNAs-based semiconductor disk laser,” Electron. Lett. 47(20), 1139–1140 (2011). [CrossRef]
19. M. Fallahi, L. Fan, Y. Kaneda, C. Hessenius, J. Hader, H. Li, J. V. Moloney, B. Kunert, W. Stolz, S. W. Koch, J. Murray, and R. Bedford, “5-W yellow laser by intracavity frequency doubling of high-power vertical-external-cavity surface-emitting laser,” Photon. Technol. Lett. 20(20), 1700–1702 (2008). [CrossRef]
20. J. Rautiainen, I. Krestnikov, J. Nikkinen, and O. G. Okhotnikov, “2.5 W orange power by frequency conversion from a dual-gain quantum-dot disk laser,” Opt. Lett. 35(12), 1935–1937 (2010). [CrossRef] [PubMed]