We have observed laser action from optically-pumped InAs-quantum-dots embedded in a line-defect waveguide in an air-bridge type GaAs-photonic-crystal slab (an array of air-holes). The lasing is found to occur without any optical cavity such as a set of Fabry-Perot mirrors. Comparison of the observed transmittance spectrum with the calculated band dispersion of the W3 defect-mode enables us to specify the lasing wavelength as that at the band edge. From this fact it follows that distributed feedback mechanism at the band edge with a vanishingly small group-velocity should be responsible for the present lasing. Usefulness of this kind of compact laser in a future ultrafast planar photonic integrated circuit is discussed.
©2004 Optical Society of America
Three (3D)- and two (2D)-dimensional photonic crystals (PCs) are very suited for controlling radiation field and propagation characteristics of light [1, 2]. As an important application, it is expected that an ultrafast and ultra-miniature planar photonic integrated circuit (PIC) should be developed on the basis of 2D PC slab waveguides (PCS-WGs) composing of the line-defect . In this connection, a PCS-WG with quantum dots (QDs) embedded in a selective area should be attractive , because such a WG provides the PIC with active and passive elements monolithically integrated in the same platform. For example, a small PCS-WG laser containing such QDs as light emitters may be feasible without any Fabry-Perot mirrors, but utilizing a vanishingly small group velocity V g, e.g., in the vicinity of the band edge [5–7].
As the first step toward such a QD laser, we fabricated GaAs-based line-defect-PCS-WGs containing InAs-QDs in the entire region. In this Letter we report that lasing action without any optical cavity such as one comprising a set of Fabry-Perot mirrors has been successfully observed from those samples by optical pumping. Laser emission is observed to come out from one cleaved edge of the PCS-WG. In order to gain an insight into the lasing mechanism, we have compared the observed transmission spectrum of the sample with the calculated line-defect band for the guided modes. As a result, it is found that the lasing wavelength corresponds to that in the vicinity of the Brillouin-zone (BZ)-edge of the even-symmetry line-defect band, where V g is vanishingly small. This fact indicates that the distributed feedback mechanism should be responsible for the present lasing.
Both triple(W3)- and single(W1)-line-defect WGs of 600 µm in length, which are surrounded by 10-rows of air-holes on both sides, were fabricated in an air-bridge type GaAs PC composed of air-holes with a 2D triangular lattice structure; the fabrication method is similar to the previous one . Scanning electron microscope (SEM) images of those samples are shown in Fig. 1.
Three layers of InAs QDs with the density of 3.2×1010 cm-2 were embedded in the 280-nm-thick GaAs core layer. By using a special site-control method, we have successfully prepared the QDs with relatively small size-distribution, which is reflected in a narrow spectral-width of 35 nm in the inhomogeneously-broadened emission spectrum from the lowest-quantum-confined state; in this case the spectrum was observed from the top surface of the sample, unlike the case which will be described later. The shape of QDs is a spheroid with the rotational axis perpendicular to the slab plane. The length of the axis, which is much smaller than other two, is adjusted to be such that the central wavelength of the emission is located around 1300 nm. Adequate values of the lattice constant a and the hole diameter 2r were adopted so as for the energy region showing V g=0 for the line-defect guided mode to overlap with the emission spectrum; actually, several similar samples were fabricated with a-values varied systematically, considering inevitable fluctuations of parameters involved in the fabricated sample.
3. Experimental method
For optical pumping, a Q-switched YAG-laser of 1.06 µm in wavelength with the pulse-duration of 30 ns operated at the repetition rate of 1 kHz was employed. Namely, pulsed beam from the laser was focused onto the top surface of the PCS-WG with the spot diameter varied in the range from 20 to 100 µm. The spot position was adjusted to cover one of the cleaved edges of a PCS-WG sample, unless otherwise noted. Emission light emerging from this edge was picked up with a polarization-preserving optical fiber with the core diameter of 9.9 µm and was fed to a monochromator equipped with a grating of 150 grooves/mm. Then, it was detected with a cooled multichannel InGaAs detector [9, 10].
In this paper, we describe mainly the result for a W3 line-defect PCS-WG of a=315 nm and 2r/a=0.56; this is because light propagation loss in samples with QDs is smaller in W3 samples than W1 samples, allowing us to more easily observe lasing action for the former sample.
Figure 2 shows the result for the case where the line-defect portion 12 µm apart from the cleaved edge was pumped with the spot diameter of 25 µm. The spontaneous emission spectrum observed in the same way as described above, which is predominantly polarized parallel to the slab plane, is shown in Fig. 2(a). The intensity of emission polarized in parallel to the plane was observed to be, roughly speaking, by a factor of 5 larger than that polarized perpendicularly to the plane. The spectral shape as well as the peak wavelength of 1287 nm was essentially the same between the two. It is clearly seen in Fig. 2 that the emission spectrum varies drastically as pumping fluence is increased. Namely, above a threshold pumping fluence of (2.1±0.4)×105 W/cm2 a narrow spectral line manifests itself at the position of 1281 nm, which shifts by 6 nm from the peak wavelength of spontaneous emission spectrum of 1287 nm.
In Fig. 3 is shown a plot of the relative emission intensity at 1281 nm as a function of pump-fluence, which indicates a clear nonlinear behavior characteristic of lasing action. This laser line is found to be polarized parallel to the slab plane; in other words, the TE-like mode is responsible for the present lasing. Variation of the relative intensity observed with shifting the optical-fiber position in parallel to the cleaved sample edge, reveals that the laser light comes out from the edge of the line-defect. We also examined the dependence of the threshold fluence on the spot diameter, with the incident power kept constant, for the case where the spot covers the line-defect. The result reveals that it decreases as the diameter increases.
Next, it is found that lasing still occurs when the position of the focused spot for pumping was moved, with the spot diameter kept unchanged, to the position 200 µm apart from the edge. Importantly, the lasing wavelength remains unchanged, and the threshold fluence becomes approximately one and half larger for this case than for the case including one edge. The observed intensity of lasing light at pumping fluence 9.0×105 W/cm2 for the former becomes considerably weaker compared to that at 6.0×105 W/cm2 for the latter case, as it should be; this is because in the former case, QDs not optically excited cause laser light attenuated before reaching the edge. This fact is considered to indicate that the present laser action occurs without any Fabry-Perot mirrors.
In order to unravel the mechanism of the present lasing, we have observed the transmission spectrum of the same sample, and have compared it with the guided modes of the calculated line-defect band. The observed spectrum (red line) for the W3 sample is shown in Fig. 4, where the spectrum (black line) for an identical sample without QDs is also shown for comparison. The latter spectrum is made red-shifted by 28 nm, considering slight difference of the sample depth between the two, i.e., 280 nm for the sample with QDs against 250 nm for that without QDs; the calculated energy difference of 28 nm for the long-wavelength edge of the band gap between 250-nm and 280-nm thick samples is adopted as the shift quantity. In the left of Fig. 5 is shown the band structure of the W3 line-defect modes, which was calculated by using two-dimensional (2D) finite-difference-time-domain (FDTD) method assuming an effective refractive index; for comparison as well as for later use, that for the W1 line-defect modes is also shown in the right of Fig. 5. In the present W3 case, there exist six line-defect bands in the band gap, including one for the fundamental refractive-like modes; those can be classified into the even and odd modes, which are symmetric and anti-symmetric, respectively, with respect to the mirror plane parallel to the propagation direction and normal to the slab plane.
First, let us see that the observed spectrum of the sample with QDs can be well explained in terms of the calculated band structure. In doing so, the spectrum of that without QDs is very helpful, since it provides us information about the transmission region due to the line-defect guided modes. It is noted that the latter spectrum is recognized to be basically consistent with the calculated band structure, in particular, concerning the observed high transmission region from 1220 to 1350 nm; note that this region should correspond to the energy region for the guided modes below the air light-line.
Now, as for the spectrum of the sample with QDs, first we point out that a value of 1355 nm observed as the long-wavelength edge of the high transmission region agrees well with the calculated one 1346 nm; note that this value corresponds to the upper-most energy of the lower slab band. A comparison of the spectrum with that without QDs reveals that, as wavelength decreases, the high transmittance region, which is supposed to be from 1355 to 1220 nm, starts decreasing from around 1330 nm, reaches the minimum at 1290 nm and then recovering again to some extent up to 1245 nm. This feature is well interpreted as arising from the absorption due to QDs; see the emission spectrum in Fig. 2(a). It is remarked that a comparison between the two spectra also provides information about the magnitude of attenuation due to absorption of QDs, which is approximately 15 dB per 600 µm at 1290 nm.
From the fact that the spectrum of the sample with QDs can be well explained in terms of the band structure shown in Fig. 5, one can deduce what energy position of the W3 line-defect band the laser wavelength corresponds to. Thus we know that the observed lasing wavelength of 1281 nm is close to the line-defect-band edge value 1275 nm of the second-lowest even mode, as marked by an arrow. From this fact we are led to conclude that the present lasing should take place due to distributed feedback mechanism in the vicinity of the BZ boundary of the line-defect band.
We have also observed lasing in a few other similar W3 samples with parameter values slightly different from those of the above sample. The lasing behavior is basically similar to the case described thus far, except for lasing wavelength. Namely, the observed lasing wavelengths were 1287, 1293 and 1304 nm. This result is considered to be quite reasonable because the BZ-edge energy varies from sample to sample.
Here we briefly mention the preliminary result for the case of W1 PCS-WG, although we have not examined many W1 samples yet. We have already observed lasing action at 1301 nm in one sample of a=330 nm and 2r/a=0.60. In this case, the threshold pumping fluence observed under similar pumping condition is larger to some extent than that of the W3 sample. This result is likely to be caused by larger light propagation loss in W1 than W3 samples, which was directly observed.
Now, let us discuss the result from a few points of view. First, concerning the lasing in the line-defect PCS WG, very recently a pioneering work utilizing InGaAsP multiple quantum-wells (MQWs) as light emitters and W1 samples of shorter length (18 µm) has been reported , where it is confirmed that the lasing occurs at the BZ-edge of the even-symmetry guided mode; the lasing mechanism is also elucidated . Concerning the lasing mechanism, the present result is considered to be consistent with this MQWs laser. In this connection, it is remarked that lasing has also been observed for composite defects composing of bend or branch, and line-defect based on PCS WGs with InGaAsP MQWs embedded ; the lasing is interpreted as arising from a localized mode with the resonant frequency slightly lower than the band-edge frequency for the even-symmetry W1 line-defect mode. It is noted that the present lasing should occur due not to a localized mode caused by fluctuations of lattice parameters, but to the distributed feedback mechanism. This statement is supported by the fact that the lasing wavelength is the same between the cases where the region including one edge and that far from the edge were optically pumped, as already described; furthermore, the fact that in the latter case, the intensity of the output laser observed in front of the edge was not that much weak, as compared to that in the former case, also seems to support the above statement.
Second, one might think that another lasing should occur in the bulk PC area composing of ten rows of air-holes along the line-defect under study. In this case, lasing is caused by V g=0 at the BZ-edges of the bulk-PC guided modes above and below the first stop band in the Γ-K direction for TM-like mode . Experimentally, however, no such lasing was observed. Presumably, the reason is either that those wavelengths are not superimposed with the present spontaneous emission spectrum or that the width of the relevant region (ten rows) is too narrow to cause a ring laser. In addition, that the TM-polarized spontaneous emission is observed to be by a factor of 5 weaker in intensity compared to the TE-polarized one is possible to be partly responsible for this problem.
Finally, from the view point of application, in particular, to the PIC platform, the present laser utilizing QDs should be more practical than those utilizing MQWs, since it is much easier to bury QDs only in a selective area. Notice that this kind of laser can be directly connected to a transparent passive PCS WG without QDs; such a structure can be fabricated in a rather simple way, i.e., without selective/overgrowth technique. It is noted that light wave from such a laser can be made not reflected back to the QDs-lasing region by using a PCS-WG-based directional coupler recently developed by our group . In this case, it is better to use the W1 line-defect PCS-WG with QDs, since the W1 straight and bend PCS-WGs have already been demonstrated to show superior characteristics in terms of propagation loss and band-width [10, 11]. By designing carefully a W1 sample so as for the zone-edge wavelength to be almost superimposed with the peak emission wavelength on one hand, and by burying QDs as homogeneously as possible to reduce the propagation loss on the other hand, very recently we have already developed such a sample of good quality. We have also succeeded in fabricating such a sample as embedding QDs only in a selective area. An attempt to observe laser action in this kind of sample is now in progress.
We have successfully observed lasing action without any optical cavity from optically-pumped InAs-QDs embedded in the PC-slab line-defect. By comparing carefully the transmittance spectrum observed for the same sample in a broad wavelength range with the calculated band structure, we have specified the wavelength of the present lasing as that in the flat region of the line-defect band. As a consequence, it is concluded that the small group-velocity causes, or the distributed feedback mechanism should cause the present lasing. The present InAs-QDs-based laser should be served as a new type of compact laser source in the future ultrasmall planar integrated optical circuit.
This work was conducted in a framework of the Femtosecond Technology Project sponsored by The New Energy and Industrial Technology Development Organization (NEDO) of Japan.
References and links
2. J. D. Joannopoulos, R. D. Meade, and J. N. Winn, Photonic Crystals; Molding the Flow of Light (Princeton University Press, Princeton, 1995).
3. Y. Sugimoto, N. Ikeda, N. Carlsson, N. Kawai, K. Inoue, and K. Asakawa, “Light propagation characteristics of photonic crystal waveguide for miniaturized ultra-fast optical-pulse control/delay devices,” in Photonic Technology in the 21st Century, Proc. SPIE4598, 58–72 (2002).
4. S. Kohmoto, H. Nakamura, T. Ishikawa, and K. Asakawa, “Site-controlled self-organization of individual InAs quantum dots by scanning probe-assisted nanolithography,” Appl. Phys. Lett. 75, 3488–3490 (1999). [CrossRef]
5. K. Inoue, M. Sasada, J. Kawamata, K. Sakoda, and J. W. Haus, “A Two-Dimensional Photonic Crystal Laser,” Jpn. J. Appl. Phys. 38, L157–L159 (1999). [CrossRef]
6. M. Meier, A. Mekis, A. Dodabalapur, A. Timko, R. E. Slisher, J. D. Joannopoulos, and O. Nalamasu, “Laser action from two-dimensional distributed feedback in photonic crystals,” Appl. Phys. Lett. 74, 7–9 (1999). [CrossRef]
7. M. Notomi, H. Suzuki, and T. Tamamura, “Directional lasing oscillation of two-dimensional organic photonic crystal lasers at several photonic band gaps,” Appl. Phys. Lett. 78, 1325–1327 (2001). [CrossRef]
8. Y. Sugimoto, Y. Tanaka, N. Ikeda, Y. Nakamura, K. Asakawa, and K. Inoue, “Low propagation loss of 0.76 dB/mm in GaAs-based single-line-defect two-dimensional photonic crystal slab waveguides up to 1 cm in length,” Opt. Express 12, 1090–1096 (2004), http://www.opticsexpress.org/abstract.cfm?URI=OPEX-12-6-1090 [CrossRef] [PubMed]
9. K. Inoue, Y. Sugimoto, N. Ikeda, Y. Tanaka, K. Asakawa, H. Sasaki, and K. Ishida, “Ultra-Small Photonic-Crystal-Waveguide-Based Y-Splitters Useful in the Near-Infrared Wavelength Region,” Jpn. J. Appl. Phys. 43, L446–L448 (2004). [CrossRef]
10. K. Inoue, Y. Sugimoto, N. Ikeda, Y. Tanaka, K. Asakawa, T. Maruyama, K. Miyashita, K. Ishida, and Y. Watanabe, “Ultra-Small GaAs-Photonic-Crystal-Waveguide-Based Near-Infrared Components: Fabrication, Guided-Mode Identification, and Estimation of Low-Loss and Broad Band-Width in Straight Waveguides, 60°-Bends, and Y-splitters,” Jpn. J. Appl. Phys , 43, 6112–6124 (2004). [CrossRef]
11. A. Sugitatsu and S. Noda, “Room temperature operation of 2D photonic crystal slab defect-waveguide laser with optical pump,” Electron. Lett. 39, 213–214 (2003). [CrossRef]
12. A. Sugitatsu, T. Asano, and S. Noda, “Characterization of line-defect-waveguide lasers in two-dimensional photonic-crystal slabs”, Appl. Phys. Lett. 84, 5395–5397 (2004). [CrossRef]
13. K. Inoshita and T. Baba, “Lasing at bend, branch and intersection of photonic crystal waveguides,” Electron. Lett. 39, 844–845 (2003). [CrossRef]
14. H. Y. Ryu, S. H. Kwon, Y. J. Lee, and Y. H. Lee, “Very-low threshold photonic band-edge lasers from free-standing triangular photonic crystal slabs,” Appl. Phys. Lett. 80, 3476–3468 (2002). [CrossRef]
15. Y. Sugimoto, Y. Tanaka, N. Ikeda, T. Yang, H. Nakamura, K. Asakawa, K. Inoue, T. Maruyama, K. Miyashita, K. Ishida, and Y. Watanabe, “Design, fabrication, and characterization of coupling-strength-controlled directional coupler based on two-dimensional photonic-crystal slab waveguides,”, Appl. Phys. Lett. 83, 3236–3238 (2003). [CrossRef]