Open Access
Add to BibSonomyAbstract
We report on the development of an optically-pumped vertical external-cavity surface-emitting laser emitting near 1120 nm using strain compensated quantum wells. The development is motivated by the need to achieve narrow linewidth emission at ~280 nm via fourth harmonic generation, which is required to cool Mg+ ions. The gain mirror had a top-emitting geometry, was grown by molecular beam epitaxy and comprised GaInAs/GaAs quantum wells strain compensated by GaAsP layers; the strain compensation was instrumental for achieving a dislocation free epitaxial structure without dark lines. We demonstrate VECSEL operation at a fundamental wavelength close to 1118 nm with a linewidth of less than 300 kHz. Using a lithium triborate crystal we achieved frequency doubling to ~559 nm with an output power of 1.1W.
©2012 Optical Society of America
1. Introduction
Laser cooling [1] and trapping of atoms or ions require narrow linewidth (< 1 MHz), frequency-stable light sources that operate in resonance with specific atomic transitions. Laser cooling is commonly performed using either dye or fiber lasers. Dye lasers offer a wide range of wavelengths from the visible to near infrared but the high cost of pump lasers, the size, and the need of constant dye solution maintenance are a drawback. Fiber lasers are capable of generating high output power with excellent beam quality but only at the distinct energy bands determined by the pump-laser-active ions. This wavelength limitation can be overcome by placing the fiber laser as part of an optical parametric oscillator (OPO). The complexity of such OPOs, however, increases their manufacturing costs impeding their implementation to various applications. The drawbacks associated with the current laser solutions used in atom cooling could be alleviated by using a relatively novel type of laser architecture, which is the optically-pumped vertical external-cavity surface-emitting laser (VECSEL), also referred to as semiconductor disk laser (SDL) [2]. The VECSELs make use of semiconductor gain media and resemble many of the design features characterising the solid state disk lasers.
By combining the power scaling capability of disk laser concepts with the gain provided by semiconductor structures, VECSELs are capable of generating high lateral mode quality beams with high output powers over a range of emission wavelengths spanning from ultraviolet to mid-infrared [3–6]. These properties, together with the ability to tailor the emission of the semiconductor gain medium, are unique advantages of this laser platform. In addition, optical elements, such as birefringent filters and etalons, can be added to the cavity to filter out any unwanted cavity modes and to allow tuning of the wavelength and narrow bandwidth operation [7].
This report is focused on leveraging the advantages offered by VECSEL concepts to a wavelength range that would enable obtaining 280 nm radiation required for cooling of Mg+ ions. Fourth harmonic generation to deep ultraviolet has been already demonstrated using VECSELs but at wavelengths that are not suitable for cooling of Mg+ ions [8–10]. In particular we describe the design and fabrication of the strain compensated gain mirrors required for achieving narrow linewidth emission in VECSEL with fundamental wavelength at ~1118 nm.
2. Design and fabrication
The structure of the gain mirror is presented in Fig. 1 . It consists of a 25.5-pair AlAs/GaAs DBR and a GaInAs/GaAs/GaAsP gain section. The indium molar fraction of the quantum wells (QWs) was ~31% as estimated by measuring ω-2θ x-ray diffraction (XRD) patterns. The XRD measurements used for estimating the In composition are shown in Fig. 2 , and are also proof of a high structural quality of the gain mirror. The compressive strain associated with each of the 7 nm thick QWs was compensated by a tensile-strained 30-nm-thick GaAsP layer with an estimated phosphorous molar fraction of 13%. The thickness of each tensile GaAsP layer is in fact compensating about 90% of the compressive strain-thickness product of each QW. According to our studies, without strain compensation, the structure would exhibit relaxation of the QWs and consequently have a high density of dark line defects [11]. Compared to the structure we reported in [11], the amount of QWs was increased from 6 to 10, and the gain mirror processing was substantially simplified by switching from flip-chip architecture [2] to top-emitting, i.e. intracavity heat-spreader, approach [12]. Both changes should result in improved capabilities for high power operation. In particular, for the type of output coupling we have used in our experiments, a gain mirror comprising 10 QWs appears to be optimal for high power operation [13].
Fig. 2 Experimental and simulated (004) ω-2θ x-ray diffraction patterns. The fine structure pattern and the narrow diffraction peaks indicate a high crystalline quality.
Besides reducing the efficiency via non-radiative recombination and increased heating, the presence of dislocations is detrimental to laser life time [14] and therefore the strain compensation is mandatory for gain mirrors incorporating a relatively high amount of In. GaAsP strain compensation has been previously utilized in MOVPE grown GaInAs/GaAs gain mirror [10]. However, our strain compensation strategy differs in terms of GaAsP composition and thickness from the one described in [10], where only slightly strained GaAs0.97P0.03 layers have been utilized. Moreover, the precise thicknesses of the GaAs0.97P0.03 layers are not revealed in [10] but the small amount of phosphorus implies much larger thickness (about a factor of 4), if similar level of strain compensation is targeted. Most likely, significant amount of pump light is absorbed in such thick strain compensation layers with low P content, whereas in our case, the pump light is absorbed mainly in the GaAs barrier layers.
The optical thickness of the gain region was chosen such that the etalon formed between the DBR and the gain mirror surface would be anti-resonant with the laser operation wavelength. This configuration provides a wide gain bandwidth and a wider wavelength tuning range. The structure was grown by solid-source molecular beam epitaxy (SSMBE) on a (001) n-GaAs substrate. To reach high luminescence efficiency, the QW growth temperature was optimized by fabricating several samples containing one QW. The quality assessment was performed by comparing the intensities of their photoluminescence (PL) signals. The optimization yielded a growth temperature of 545 °C, measured by the thermocouple mounted on the substrate holder. This corresponds to a pyrometer reading of about 500 °C. The PL peak intensity versus thermocouple reading is presented in Fig. 3 . Using the room temperature PL setup available, we could not detect any PL signal for the extreme growth temperatures of T = 370 °C and 565 °C (thermocouple reading). A possible explanation for the poor PL signal at the high temperature is the relaxation of the GaInAs layer leading to growth mode transition from 2D to 3D and subsequent generation of misfit dislocations in the crystal [15]. We believe that the poor PL signal at the low growth temperatures is caused by structural defects such as arsenic antisites and gallium vacancies [16].
Fig. 3 Photoluminescence peak intensity from single-QW samples grown at different temperatures. The 370 °C and 565 °C data point were excluded from the figure since no emission was obtained at these temperatures using the available room temperature PL system.
The strain compensation and GaAs barrier layers were grown at a typical growth temperature of 580–590 °C (pyrometer measurements). The growth was stopped between these layers and the QWs in order to ramp the temperature to adequate levels ensuring high quality growth. The stop band of the DBR and the surface PL peak are centered at 1123 nm and 1121 nm, respectively, as shown in Fig. 4 . The success of the strain compensation was investigated by measuring a PL map of the gain mirror that would point out any dark line or dark spot defects. The measurement revealed a very low dark spot density and absence of dark lines. Most likely, these dark spot defects originated from the substrates. The ω-2θ x-ray diffraction spectra, shown in Fig. 2, did not reveal any signs of strain relaxation or diffusion scattering from crystal defects.
Fig. 4 Normalized surface photoluminescence (PL) and reflectance curves recorded from the gain mirror after growth. The dip in the reflectivity characteristic is caused by QW absorption.
After the epitaxial growth and material characterizations, a 2.5 mm × 2.5 mm chip was cleaved from the wafer and capillary bonded [17] on the epi-side to a 3 mm × 3 mm diamond heat spreader that has a 2° wedge. Water was used to pull the two surfaces into contact and eventually to bond them together by intermolecular surface forces. A slight pressure was applied to the chip-diamond interface during and after bonding to ensure firm contact. The single-crystal synthetic diamond had a surface roughness (Ra) below 5 nm and a surface flatness of <5 interference fringes in reflection at 632.8 nm. The surface roughness of the semiconductor chip was determined by using an atomic force microscope from a 3 µm × 3 µm sized scan area. The measurement was performed at two different locations on the gain chip, yielding a roughness of ~0.17 nm (Ra-value). The wedged geometry of the diamond heat spreader alleviates the etalon effects that would otherwise perturb the laser spectrum. The heat spreader was clamped with an intervening indium foil to a cooled copper sample holder (Fig. 5 ). An antireflective coating was deposited on the air-diamond interface in order to reduce the intracavity reflection at the lasing wavelength and further weaken the subcavity etalon effect. The antireflective coating comprised altogether two layers, a TiO2 layer and a SiO2 layer. The reflectance minimum was targeted at the 1120 nm laser wavelength but the coating provided also significantly reduced reflectivity (R<12%) at the pump emission wavelength of 808 nm for an 8° incidence angle.
Fig. 5 Schematic showing the gain mirror attachment to the wedged diamond heat spreader and the diamond attachment to the copper mount with indium.
3. Laser setups and characterization
The gain mirror was first used for generating narrow linewidth light close to 559.271 nm via second harmonic generation (SHG), which can be further frequency doubled to reach the 24Mg+ D2 transition wavelength of 279.635 nm [18]. For this purpose, we assembled a cavity as depicted in Fig. 6 . The cavity comprised a 3 mm thick birefringent (BRF) filter and a 1 mm thick uncoated yttrium aluminum garnet (YAG) etalon for wavelength tuning and linewidth narrowing. A 15 mm long rectangular lithium triborate crystal (LBO) crystal was used for SHG. The LBO crystal was antireflective coated for 1120 nm and 560 nm wavelengths, and was placed about 5 mm away from the 4 mm thick fused silica cavity end mirror. To achieve efficient nonlinear conversion the LBO crystal was heated at about 88 °C. The distance from the gain mirror to the folding mirror was 263 mm while the distance from the folding mirror to the flat end mirror was 108 mm. The mode size (1/e2 diameter) on the semiconductor chip was estimated to be 273 µm in horizontal direction and 308 µm in vertical direction. The angle of the folding mirror was ~11°. The coating of the cavity mirrors were chosen such that the light beam at the fundamental frequency could not escape the cavity while the frequency doubled light could exit the cavity only through the cavity folding mirror. The end mirror coating was also designed to minimize the phase shift between the fundamental and second harmonic frequency light. The laser was pumped with an 808 nm pump laser that was fiber coupled to a 200 µm core fiber and focused to a spot diameter (1/e2 diameter) of ~390 µm.
Up to 1.1 W was achieved at 560 nm with an incident pump power of 30 W, which should enable achieving 100–200 mW at 280 nm [10], high enough for the ion-cooling experiments. The wavelength of the laser was determined by comparing the number of interference oscillations with those of a HeNe laser in a home-built Michelson interferometer. The output wavelength of the laser was adjusted by rotating the BRF. While it was possible to get within ~50 pm of the desired wavelength of 559.271 nm, it was found that the laser wavelength jumped by ~170 pm when tuning the BRF. We suspect the jumps to be associated with etalons formed by the optical elements of the cavity. Because of this wavelength jump we were not able to perform accurate linewidth measurement; the wavelength could not be tuned close enough to emission of a reference laser to observe the beat signal.
In order to eliminate the wavelength discriminating behavior and to measure the laser linewidth, a straight cavity without the LBO crystal was constructed (see Fig. 7 ). Also the etalon was omitted because it was not needed for achieving single frequency operation. In this configuration, the laser would operate in a single-frequency mode depending on the BRF angle and the pump power. The laser still exhibited wavelength jumps on the scanning Fabry-Perot etalon when the cavity length was varied or the BRF angles were adjusted. However, the step was measured to be as small as 7.6 pm, which was the same as the wavelength step size of the wavemeter we used for measuring the emission wavelength. At 1118.5 nm, 7.6 pm corresponds to 1.8 GHz, whereas the cavity free spectral range was 0.82 GHz, which is below the resolution of the wavemeter. Therefore, we could not verify whether the jumps observed by the scanning Fabry-Perot etalon were longitudinal modes of the cavity. By simplifying the cavity, we were able to significantly improve the tuning of the laser by reducing the magnitude of the wavelength jumps.
Figure 8(a) shows the power conversion graph of the laser corresponding to a temperature of the cooling water set to 15 °C. The kinks in the graph are most likely due to sudden changes to higher transversal mode operation or a sudden wavelength jump. The maximum output power was limited by the available pump power. The maximum optical conversion efficiency was about 25%. Figure 8(b) shows the output power versus emission wavelength when the wavelength was tuned by rotating the BRF and the pump power was kept constant at 30 W. The total tuning range reached was 31 nm. During the tuning experiment, the laser operated mostly in a single frequency mode, i.e., it would occasionally emit multiple frequencies as observed with the scanning Fabry-Perot etalon. To measure the emission linewidth, the laser was tuned close to the desired wavelength and pumped with 9 W to reach about 0.8 W of output power and a beat measurement between the laser and a commercial fiber laser (Menlo Systems) was performed. The assessment of linewidth at higher output power was not possible because the chiller used for cooling the gain mirror did not have adequate capacity to keep the temperature constant for the period required to carry out the measurement. The beams from both lasers were directed via a single mode fiber to a single high-bandwidth photodetector and the beat signal was recorded using a spectrum analyzer (Agilent). The measured beat signal is shown in Fig. 9 . The beat width was below the 300 kHz resolution bandwidth of the spectrum analyzer. The use of narrower measurement bandwidth was not possible because it would require a longer sweep time during which the wavelength of the laser drifted outside the measurement window. The wavelength drift could have been eliminated or reduced significantly by introducing active wavelength stabilization to the cavity [19]. This, however, could not be implemented to the cavity in this experiment but will be implemented in the future design of the laser cavity.
4. Conclusion
We have shown that GaInAs/GaAs-based VECSELs are a promising technology to cool Mg+ ions. According to our experiments, strain compensation and a fairly low growth temperature were necessary to reach high crystalline quality for a gain mirror comprising 10 GaInAs QWs. In particular, we have identified an optimal growth temperature of 500 °C (pyrometer reading) for the QWs. The strain compensation was achieved by using 30-nm-thick GaAsP layers with a P concentration of 13%. We were able to reach ~5.5 W at 1118 nm with 30 W of pump power. A linewidth of less than 300 kHz could be measured for an output power of 0.8 W at the fundamental wavelength of 1118 nm with 9 W of pump power. Linewidth measurements at higher pump power were not possible due to the technical limitations of the cooling system available. Furthermore, we were able to demonstrate 1.1 W output power at 559 nm with 30 W of pump power by second harmonic intracavity conversion. Our results prove that it would be possible to generate continuous tuning at the half-frequency of Mg+ ion photoionization wavelength using VECSEL technology.
Acknowledgments
This work was supported by the US National Institute of Standards and Technology, the Academy of Finland (grant 128364) and the graduate school of the Tampere University of Technology. The authors are grateful to Dr. Härkönen for evaporating the antireflective coating on the diamond heat spreader for the VECSEL and Dr. Sandalphon for technical assistance with the laser system.
References and links
1. T. W. Hänsch and A. L. Schawlow, “Cooling of gases by laser radiation,” Opt. Commun. 13(1), 68–69 (1975). [CrossRef]
2. M. Kuznetsov, F. Hakimi, R. Sprague, and A. Mooradian, “High-power (>0.5-W CW) diode-pumped vertical-external-cavity surface-emitting semiconductor lasers with circular TEM00 beams,” IEEE Photon. Technol. Lett. 9(8), 1063–1065 (1997). [CrossRef]
3. M. Rahim, F. Felder, M. Fill, and H. Zogg, “Optically pumped 5 µm IV-VI VECSEL with Al-heat spreader,” Opt. Lett. 33(24), 3010–3012 (2008). [CrossRef] [PubMed]
4. A. Härkönen, M. Guina, O. Okhotnikov, K. Rößner, M. Hümmer, T. Lehnhardt, M. Müller, A. Forchel, and M. Fischer, “1-W antimonide-based vertical external cavity surface emitting laser operating at 2-µm,” Opt. Express 14(14), 6479–6484 (2006). [CrossRef] [PubMed]
5. 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,” IEEE Photon. Technol. Lett. 22(9), 661–663 (2010). [CrossRef]
6. J. E. Hastie, L. G. Morton, A. J. Kemp, M. D. Dawson, A. B. Krysa, and J. S. Roberts, “Tunable ultraviolet output from an intracavity frequency-doubled red vertical-external-cavity surface-emitting laser,” Appl. Phys. Lett. 89(6), 061114 (2006). [CrossRef]
7. T. Leinonen, S. Ranta, M. Tavast, M. Guina, and R. J. Epstein, “Narrow Linewidth 1120 nm Semiconductor Disk Laser Based on strain compensated GaInAs quantum wells,” in Lasers, Sources, and Related Photonic Devices, OSA Technical Digest (CD) (Optical Society of America, 2012), paper AW4A.18.
8. Y. Kaneda, J. M. Yarborough, L. Li, N. Peyghambarian, L. Fan, C. Hessenius, M. Fallahi, J. Hader, J. V. Moloney, Y. Honda, M. Nishioka, Y. Shimizu, K. Miyazono, H. Shimatani, M. Yoshimura, Y. Mori, Y. Kitaoka, and T. Sasaki, “Continuous-wave all-solid-state 244 nm deep-ultraviolet laser source by fourth-harmonic generation of an optically pumped semiconductor laser using CsLiB6O10 in an external resonator,” Opt. Lett. 33(15), 1705–1707 (2008). [CrossRef] [PubMed]
9. J. Paul, Y. Kaneda, T. Wang, C. Lytle, J. Moloney, and J. Jones, “Precision Spectroscopy of Atomic Mercury in the Deep Ultraviolet Based on Fourth-Harmonic Generation from an Optically Pumped External-Cavity Semiconductor Laser,” in Conference on Lasers and Electro-Optics, OSA Technical Digest (CD) (Optical Society of America, 2010), paper CTuS6.
10. Y. Kaneda, M. Fallahi, J. Hader, J. V. Moloney, S. W. Koch, B. Kunert, and W. Stoltz, “Continuous-wave single-frequency 295 nm laser source by a frequency-quadrupled optically pumped semiconductor laser,” Opt. Lett. 34(22), 3511–3513 (2009). [CrossRef] [PubMed]
11. S. Ranta, T. Hakkarainen, M. Tavast, J. Lindfors, T. Leinonen, and M. Guina, “Strain compensated 1120 nm GaInAs/GaAs vertical external-cavity surface-emitting laser grown by molecular beam epitaxy,” J. Cryst. Growth 335(1), 4–9 (2011). [CrossRef]
12. W. Alford, T. Raymond, and A. Allerman, “High power and good beam quality at 980 nm from a vertical external-cavity surface-emitting laser,” J. Opt. Soc. Am. B 19(4), 663–666 (2002). [CrossRef]
13. 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,” IEEE J. Sel. Top. Quantum Electron. 5(3), 561–573 (1999). [CrossRef]
14. J. Jiménez, “Laser diode reliability: crystal defects and degradation modes,” C. R. Phys. 4(6), 663–673 (2003). [CrossRef]
15. S. M. Wang, T. G. Andersson, and M. J. Ekenstedt, “Temperature‐dependent transition from two‐dimensional to three‐dimensional growth in highly strained InxGa1-xAs/GaAs (0.36≤x≤1) single quantum wells,” Appl. Phys. Lett. 61(26), 3139–3141 (1992). [CrossRef]
16. X. Liu, A. Prasad, J. Nishio, E. R. Weber, Z. Liliental-Weber, and W. Walukiewicz, “Native point defects in low-temperature-grown GaAs,” Appl. Phys. Lett. 67(2), 279–281 (1995). [CrossRef]
17. Z. L. Liau, “Semiconductor wafer bonding via liquid capillarity,” Appl. Phys. Lett. 77(5), 651–653 (2000). [CrossRef]
18. M. Herrmann, V. Batteiger, S. Knünz, G. Saathoff, Th. Udem, and T. W. Hänsch, “Frequency metrology on single trapped ions in the weak binding limit: the 3s(1/2)-3p(3/2) transition in 24Mg+,” Phys. Rev. Lett. 102(1), 013006 (2009). [CrossRef] [PubMed]
19. M. A. Holm, D. Burns, A. I. Ferguson, and M. D. Dawson, “Actively stabilized single-frequency vertical-external-cavity AlGaAs laser,” IEEE Photon. Technol. Lett. 11(12), 1551–1553 (1999). [CrossRef]
