Transform-limited pulses as short as 290 fs at 1036 nm are generated by a diode-pumped semiconductor disk laser. The all-semiconductor laser employs a graded-gap-barrier design in the gain section. A fast saturable absorber mirror serves as a passive mode-locker. No further elements for internal or external dispersion control are required.
© 2008 Optical Society of America
Diode-pumped semiconductor disk lasers (SCDLs), also known as optically pumped semiconductor vertical-external-cavity surface-emitting lasers (OPS-VECSELs), challenge conventional solid-state lasers based on dielectric crystals as continuous-wave (cw) and short-pulse light sources: High output power can be achieved with nearly diffraction-limited beam quality [1,2]. Simple and compact all-semiconductor laser designs can be realized. A very broad range of visible and infrared wavelengths can be obtained by appropriate choice of semiconductor materials [3,4]. Intracavity harmonic generation provides further options [2,5,6]. Combining an SCDL gain element with a semiconductor saturable absorber mirror (SAM) as a mode-locker permits short-pulse generation. Pulse repetition rates up to 50 GHz have been demonstrated .
Three mechanisms are considered as major contributions to the pulse formation process in mode-locked SCDLs: Saturable absorption and gain, the AC Stark effect in the SAM, and the interplay of dispersion and self-phase modulation (SPM) in the absorber and gain material. If the effects balance each other, this will lead to a soliton-like pulse shaping process . In this way, pulses with a duration of about 480 fs at multi-GHz repetition rates have been achieved directly from optically pumped SCDLs near 1035 nm [9,10].
Here, we report on passively mode-locked laser operation of SCDLs based on gain sections with graded-index designs. We demonstrate pulse durations of 290 fs directly from the laser oscillator. The pulses are shorter than the 310-fs pulses at a wavelength of ≈1550 nm with a time-bandwidth product of 0.41 from a self-mode-locked edge-emitting diode laser presented in  and set - to the best of our knowledge - a new record for semiconductor lasers without external pulse compression.
2. Design of the semiconductor elements
The SCDL gain section typically consists of alternating layers of semiconductor materials, namely quantum wells (QWs), spacer layers, and strain compensating layers, leading to specific profiles of the bandgap and refractive index. The pump radiation is mainly absorbed by the spacer layers, and subsequently the generated carriers drift into the QWs. With SCDLs one speaks of a “graded-index” (GRIN) design, if the spacers exhibit a varying composition (in our case, a graded Al-content). The compositional grading results in a graded bandgap and thereby in a quasi-electric field, which promotes the carrier drift towards the QWs . Figure 1 schematically shows our GRIN gain section in comparison to a step index (STIN) architecture. The QWs are placed into the electric-field maxima of the interference pattern formed by the incident and reflected light beam (see red curves in Fig. 1), thus increasing the longitudinal confinement and consequently the gain. The Bragg mirror is low-reflective for the pump radiation; therefore, the respective interference pattern is not relevant. By introduction of a double quantum well and omitting the last but one QW in our structure, the amount of pump power provided to each QW is modified. The modification, i.e., unequal pumping of the QWs, results in an inhomogeneous broadening of the gain spectrum of the structure, supporting shorter pulses.
We compared the cw performance of 6-QW GRIN and STIN semiconductor disk lasers and found GRIN-structure gain media to be superior to STIN media at high pump intensities, showing a constant slope efficiency where the STIN-structure output power rolled off. We assigned this to the lack of pump absorption saturation within the GRIN structure, because the free-carrier concentration in the spacers is lower due to the faster transport towards the QWs . In mode-locked experiments with a STIN gain medium, we observed soliton-like pulses with a duration of 590 fs . In this paper, we present 290-fs pulses achieved using a GRIN structure. How far this result was enabled by the design of the gain medium is currently subject of investigation.
The GRIN SCDL structure was grown by metalorganic vapor phase epitaxy (MOVPE) on a 2” GaAs (001) substrate off-oriented 2° towards . The sections were grown in an order which enables subsequent removal of the substrate to achieve minimal thickness of the laser medium and thereby efficient heat removal. The structure consists of a gain section containing 4 strain-balanced InGaAs/(Al)GaAs/AlGaAsP QWs designed for an operating wavelength of 1040 nm and a distributed Bragg mirror (DBR) comprising 25 AlAs/GaAs quarter-wave layer pairs, which was grown last. The growth temperatures of gain section and DBR were 650° and 700° C, respectively. The QWs of the GRIN structure are surrounded by barriers graded from GaAs to Al0.2Ga0.8As. The wafer was cut into 2×2 mm2 pieces, which were mounted mirror-side down (“bottom up”) on CuW heat sinks using AuSn solder. Afterwards the substrate was etched off with an aqueous solution of sulphuric acid and hydrogen peroxide. A ≈300-nm-thick In0.48Ga0.52P layer served as an etch-stopper. Finally, an antireflective dielectric coating was applied and the structure was mounted on a copper heat sink.
The small number of QWs has two consequences: On the one hand, also due to the low absorption of AlGaAs at 840 nm, only ≈20% of the incident pump radiation are absorbed in QWs and barriers. This leads to a lasing threshold at around 1 W of incident pump light and a relatively low gain and efficiency. However, this is not an issue here, since we are not aiming at high output powers. On the other hand, by a thin structure we reduce disturbing effects like filtering by longitudinal confinement, material dispersion, or structural dispersion from residual reflectivity (partly due to imperfection of the AR coating). In fact, the group delay dispersion of the unpumped gain element was calculated to be <200 fs2 at the lasing wavelength. This was confirmed by measurements; a more precise determination was not possible due to the resolution limit of our white-light interferometer.
The SAM for mode-locking was grown in a similar fashion, but in the usual order, i.e., with the DBR directly on the substrate, followed by the absorbing section. This section consists of a single 10-nm-thin InGaAs quantum well embedded in GaAs layers. The top GaAs layer has a thickness of only 2 nm, so that one speaks of a “near-surface” SAM. A SiN layer was added for surface protection, which also acts as an antireflection coating. While the DBR was grown at 770° C, the absorber section was grown at 510°C to accelerate the carrier recombination. For such SAMs we measured interband relaxation times of typically less than 5 ps . For the SAM used here we recorded a relatively broad photoluminescence spectrum with a FWHM of about 60 nm. For a similar SAM we determined a 1/e-saturation fluence of about 10 µJ/cm2, measured with 150-fs pulses at 1060 nm, which in that case meant about 10 meV below the assumed spectral position of the excitonic resonance.
3. Laser operation
3.1 Experimental setup
We studied a V-shaped nearly hemispherical three-mirror resonator with the SCDL gain structure at the folding point as shown in Fig. 2. A spherical mirror (radius of curvature: -5 cm) with a transmission of 0.5% at 1040 nm acts as the output coupler, located at the end of the longer arm (length: ≈43 mm). The other cavity arm (length: ≈7 mm) was terminated by the SAM used as a passive mode-locker. The angle defined by the two arms was approximately 50°. The resonator was operated near the stability limit to realize a small spot size (estimated waist: 20–30 µm) on the SAM. The laser was optically pumped by a 60-µm broad-stripe laser diode. The single-emitter diode generated as much as 4 W of output power at 840 nm with a spectral linewidth of ≈2 nm. The astigmatic emission of the diode laser was collimated by two crossed cylindrical micro-lenses and focused by a f=62.9-mm lens to a beam waist of ≈100 µm at the position of the SCDL-chip. The temperatures of gain element and SAM were adjusted by Peltier elements to 15 and 61 °C, respectively.
3.2 Soliton-like mode locking performance
Figure 3(a) shows the autocorrelation trace of the shortest pulses generated by the SCDL using the GRIN gain element. The measurement was performed with an APE PulseCheck autocorrelator. The fit, assuming a sech2-pulse shape, reveals that pulses as short as 290 fs at a repetition rate of 3 GHz were obtained directly from the oscillator. The optical spectrum centered at 1036 nm, which is only slightly assymetric, displays a FWHM of Δλ ≈3.9 nm (Fig. 3(b)). This results in a time-bandwidth product of about 0.32, corresponding to a nearly transform-limited sech2-pulse.
We recorded the radio frequency spectrum, indicating a 65-dBc extinction ratio of the fundamental beat note at 3.013 GHz without spurious modulations, as measured with 1 kHz resolution bandwidth by a Rhode & Schwarz spectrum analyzer (Fig. 4(a)). From this and wide-span measurements up to 7 GHz (Fig. 4(b)) we deduce that the laser showed stable single-pulse operation.
The average output power was 10 mW for an incident pump power of 1.7 W. From this, we estimate pulse fluences of roughly 2–3 µJ/cm2 and 20–50 µJ/cm2 on the gain element and saturable absorber, respectively. An exact estimation of the fluences is prevented by an error in the spot size determination, resulting from the unknown curvature of the gain medium, which was introduced by the soldering process . Assuming that the saturation fluence of our SAM is approximately 10 µJ/cm2, the pulse fluence was 2–5 times higher. However, one must take into account that saturation fluences depend on the laser wavelength and are of limited usefulness, unless the pulse duration is much shorter than the time scales of all relaxation components of the saturable absorber. Femtosecond pulse generation occured, if the laser photon energy was approximately 10 meV below the assumed position of the exciton resonance of the absorber, which was estimated from SAM photoluminescence measurements. To maintain soliton-like mode locking, only small variations of laser wavelength and exciton position were permitted; i.e., the temperature of each semiconductor element could be varied by a few Kelvin.
The minimum achievable pulse duration of 290 fs was observed with p-polarization of the laser and was limited by an abrupt change to s-polarization, if the pump power was increased above 1.7 W. While operated at 1.7 W of pump power, the laser would occasionally change its polarization; at 1.65 W, operation was stable at least on an hour scale, with pulses slightly longer than 300 fs. s-polarization corresponded to a regime of non-soliton-like pulse shaping, generating chirped few-ps pulses. Further increase of the pump power caused a second transition into another soliton-like regime, characterized by the original p-polarization and multiple fs-pulses circulating in the laser cavity.
The polarization change is observed, as there is no strongly polarization-selective element inside the laser resonator. p-polarization is not preferred considering the higher losses experienced at the Bragg reflector of the gain medium; the difference in reflection for s- and p-polarization is only <0.1%, however. Depending on the operation regime, this value can be exceeded by a gain difference that results from growth anisotropy or from tensions generated by the soldering process. Multiple-pulsing instabilities are typical for a system with soliton-like pulse shaping . The issue becomes the more important, the more of the gain bandwidth is utilized, i.e. if minimum pulse durations are aimed at like in our case. Pulses split, because for higher pulse energies, the reduction of the SAM absorber loss by stronger saturation would become smaller than the increase of gain filtering loss experienced by the spectrally broader pulse. For a semiconductor laser, the issue of multiple pulsing is enhanced by the short interband relaxation time , which is of the same order as the cavity round-trip time, i.e., the laser can reduce loss from spontaneous recombination by decreasing the time interval between pulses. To obtain shorter single pulses, the gain bandwidth and the ratio of interband relaxation time to round-trip time should be increased.
A pulse tail (“continuum”) is another phenomenon known with soliton-like mode locking and can be found with many dielectric-gain-media lasers and all SCDLs, e.g. in [7,9,10,14]. The actual soliton-like pulse optimizes its gain by shedding radiation. In our case, the tail is very weak and observable only as a slight asymmetry of the optical spectrum (Fig. 3(b)). A tail or afterpulses can be promoted by too slow carrier dynamics in the semiconductor elements, as pulses become shorter than the respective recovery times. In this context, the gain medium may profit from the faster carrier transport in the GRIN structure, as the barriers serve less as a carrier reservoir. Continuum generation should be countered by designing and operating the saturable absorber for fast response.
The energy separation between laser phonons and the band edge or the exciton resonance of the SAM is an important parameter determining the modulation depth and the ratio of fast to slow components of the absorber response. The absorption is saturated via band filling; relaxation times are given by ionization of excitons, thermalization within the conduction band and with the lattice (fast, sub-ps time scales) and by interband relaxation (slow, ps time scale). The AC Stark effect was proposed as a another, practically instantaneous component , as the exciton resonance is Stark-shifted by the electrical field strength of the pulse. However, a thorough investigation of the saturation dynamics of the semiconductor elements in mode-locked SCDLs has not been performed so far.
We demonstrated a passively mode-locked diode-pumped semiconductor disk laser utilizing a graded-index 4-QW gain structure, that generated practically chirp-free pulses with a duration of 290 fs. To the best of our knowledge, these are the shortest pulses obtained directly from a semiconductor laser. For this result, we found the appropriate energy separation between the laser photon energy and the estimated exciton resonance to be essential, i.e., the first had to be ≈10 meV below the latter.
The authors like to thank Thomas Roos and his team at Ferdinand-Braun-Institute for excellent mechanical support. This work was funded by the German Bundesministerium für Bildung und Forschung under Grant No. 13N8570.
References and links
1. 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, 561–573 (1999). [CrossRef]
2. J. L. Chilla, S. D. Butterworth, A. Zeitschel, J. P. Charles, A. L. Caprara, M. K. Reed, and L. Spinelli, “High power optically pumped semiconductor lasers,” in Solid State Lasers XIII: Technology and Devices, R. Scheps and H. J. Hoffman, eds., Proc. SPIE 5332, 143–150 (2004). [CrossRef]
3. J. E. Hastie, S. Calvez, M. D. Dawson, T. Leinonen, A. Laakso, J. Lyytikäinen, and M. Pessa, “High power CW red VECSEL with linearly polarized TEM00 output beam,” Opt. Express 13, 77–81 (2004). [CrossRef]
4. L. Cerutti, A. Garnache, A. Ouvrard, and F. Genty, “High temperature continuous wave operation of Sb-based vertical external cavity surface emitting laser near 2.3 µm,” J. Cryst. Growth 268, 128–134 (2004). [CrossRef]
5. 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, 061114 (2006). [CrossRef]
6. E. U. Rafailov, W. Sibbett, A. Mooradian, J. G. McInerney, H. Karlsson, S. Wang, and F. Laurell, “Efficient frequency doubling of a vertical-extended-cavity surface-emitting laser diode by use of a periodically poled KTP crystal,” Opt. Lett. 28, 2091–2093 (2003). [CrossRef] [PubMed]
7. U. Keller and A. C. Tropper, “Passively modelocked surface-emitting semiconductor lasers,” Phys. Rep. 429, 67–120 (2006). [CrossRef]
8. R. Paschotta, R. Häring, A. Garnache, S. Hoogland, A. C. Tropper, and U. Keller, “Soliton-like pulse-shaping mechanism in passively mode-locked surface-emitting semiconductor lasers,” Appl. Phys. B 75, 445–451 (2002). [CrossRef]
9. A. Garnache, S. Hoogland, A. C. Tropper, I. Sagnes, G. Saint-Girons, and J. S. Roberts, “Sub-500-fs soliton-like pulse in a passively mode-locked broadband surface-emitting laser with 100 mW average power,” Appl. Phys. Lett. 80, 3892–3894 (2002). [CrossRef]
10. S. Hoogland, A. Garnache, I. Sagnes, J. S. Roberts, and A. C. Tropper, “10-GHz Train of Sub-500-fs Optical Soliton-Like Pulses From a Surface-Emitting Semiconductor Laser,” IEEE Photon. Technol. Lett. 17, 267–269 (2005). [CrossRef]
11. Y. Nomura, S. Ochi, N. Tomita, K. Akiyama, T. Isu, T. Takiguchi, and H. Higuchi, “Mode locking in Fabry-Perot semiconductor lasers,” Phys. Rev. A 65, 043807 (2002). [CrossRef]
12. H. Kroemer, “Quasi-Electric Fields and Band Offsets: Teaching Electrons New Tricks (Nobel Lecture),” ChemPhysChem 2, 490–499 (2001). [CrossRef]
13. F. Saas, V. Talalaev, U. Griebner, J. W. Tomm, M. Zorn, and M. Weyers, “Optically pumped semiconductor disk laser with graded and step indices,” Appl. Phys. Lett. 89, 151120 (2006). [CrossRef]
14. F. Saas, G. Steinmeyer, U. Griebner, M. Zorn, and M. Weyers, “Exciton resonance tuning for the generation of sub-picosecond pulses from a mode-locked semiconductor disk laser,” Appl. Phys. Lett. 89, 141107 (2006). [CrossRef]
15. U. Zeimer, J. Grenzer, D. Korn, S. Döring, M. Zorn, W. Pittroff, U. Pietsch, F. Saas, and M. Weyers, “X-ray diffraction spot mapping - a tool to study structural properties of semiconductor disk laser devices,” phys. stat. sol. (a) 204, 2753–2759 (2007). [CrossRef]
16. M. J. Lederer, B. Luther-Davies, H. H. Tan, C. Jagadish, N. N. Akhmediev, and J. M. Soto-Crespo, “Multipulse operation of a Ti:sapphire laser mode locked by an ion-implanted semiconductor saturable-absorber mirror,” J. Opt. Soc. Am. B 16, 895–904 (1999). [CrossRef]
17. Esa J. Saarinen, Robert Herda, and Oleg G. Okhotnikov, “Dynamics of pulse formation in mode-locked semiconductor disk lasers,” J. Opt. Soc. Am. B 24, 2784–2790 (2007). [CrossRef]
18. S. Tsuda, W. H. Knox, S. T. Cundiff, W. Y. Jan, and J. E. Cunningham, “Mode-Locking Ultrafast Solid-State Lasers with Saturable Bragg Reflectors,” IEEE J. Sel. Top. Quantum Electron. 2, 454–464 (1996). [CrossRef]