For years, the quantum optics community has been testing the limits on laser size: How small can you go? Thanks to advances in semiconductor crystal growth and fabrication techniques, the answer is now wavelength-scale.
Researchers H.-G.Park, S.-H. Kim, M.-K. Seo and Y.-H. Lee are working to develop the smallest possible laser.
The laser is shrinking. Like other modern technologies, lasers have been found to hold many attractive capabilities when applied to the nano realm. For example, ultra-small lasers hold promise as light sources for photonic integrated circuits that require minimal thermal overhead.
In addition, researchers have been able to use nanolasers in miniature semiconductor cavities to explore fundamental issues in quantum optics. When an atom is placed in a nanocavity, its spontaneous emission lifetime is strongly altered—a property that is useful for realizing the on-demand single photon source on a chip.
Using nanolasers to strongly confine photons in a photonic crystal resonator is a particularly attractive prospect. The cavity of a photonic crystal, composed of periodic dielectric materials, can have both a high quality (Q) factor and a small mode volume, which is advantageous for ultralow-threshold lasing. Once this small cavity is made to support only one resonant mode and emit only the desired spontaneous emission, the ultimate thresholdless laser may be possible, with the help of the photonic bandgap.
In recent years, researchers have reported various optically pumped ultra-small photonic crystal lasers or electrically pumped photonic crystal band edge lasers with relatively large modal volumes (Science 293, 1123 and Science 302, 1374). However, a stand-alone, simple, ultrasmall photon source that operates with a minimum power budget is feasible only if one activates the wavelength-scale laser electrically.
Optically pumped photonic crystal lasers
Researchers have widely adapted the two-dimensional photonic crystal slab as a basic building block for photonic devices such as lasers, waveguides and filters. When a photonic crystal cavity is formed on a 2D semiconductor slab, photons tend to localize in the proximity of the photonic crystal resonator due to the effects of the photonic bandgap and total internal reflection.
A group of Caltech researchers demonstrated the first photonic crystal laser in an indium gallium arsenic phosphide (InGaAsP) material system through optical pumping (Science 284, 1819). Strained quantum wells provided an optical gain near 1,550 nm. The researchers observed pulsed lasing action by InGaAs laser diode pumping at low temperatures.
After achieving this feat, investigators unveiled photonic crystal lasers with various interesting properties. For example, some reported continuous-wave optically pumped photonic crystal lasers at room temperature (Appl. Phys. Lett., 79, 3032, and Opt. Express 15, 7506). Others demonstrated various photonic crystal resonators with very high Q factors and very small modal volumes. For example, Q factors in excess of 1,000,000, and modal volumes smaller than 0.019 μm3, were found in optimized photonic crystal cavities, respectively, by separate research groups. K. Nozaki et al. achieved an effective threshold pump power smaller than 2 μW from single-mode photonic crystal lasers.
(Top) Schematic of a 2D photonic crystal laser (from Science 284, 1819). (Bottom) Various high-Q ultra-small photonic crystal cavities (from Appl. Phys. Lett. 79, 3032; Opt. Express 15, 7506; Nature Mater. 4, 207).
Electrically driven photonic crystal band-edge lasers
The electrical pumping scheme was first realized in a relatively-large-volume photonic crystal cavity. In 2001, the Noda group at Kyoto University used wafer fusion techniques to fabricate 2D photonic crystal band edge lasers near 1,550 nm (Science 293, 1123). The photonic crystal defined on an n-InP substrate was fused with the other wafer containing an InGaAsP/InP multiple-quantum well active layer. To reduce the nonradiative surface recombination, researchers inscribed photonic crystal patterns that do not contain quantum wells on the wafer. Under pulsed current pumping conditions at room temperature, they achieved 2D lasing oscillation in a plane of photonic crystals. Photons were coupled out to the direction normal to the substrate due to first-order diffraction. This surface-emitting laser device produced output power larger than 20 mW.
Also in 2003, Colombelli et al. at Bell Laboratory demonstrated a surface-emitting quantum cascade photonic crystal laser by combining two device concepts (Science 302, 1374). They used a high-index-contrast 2D photonic crystal structure to form a micro-resonator that simultaneously provides the optical feedback needed for laser action and the vertical out-coupling from the semiconductor surface. The top metallic contact allowed electrical current injection and also provided vertical confinement.
(a) Schematic of the electrically driven single-cell photonic laser (Science 305, 1444). (b) Scanning electron microscope image of a fabricated pedestal structure for current injection (Science 305, 1444). (c) Controlled size of the central post (Appl. Phys. Lett. 90, 171122).
Single-cell photonic crystal lasers by current injection
Scientists have suggested various photonic crystal structures, and they have tried to implement a single-cell free-standing slab photonic crystal laser. However, one of the most challenging issues they have faced is how to make—and where to position—the electrical contact post.
The point of maximum symmetry is typically chosen for carrier injection. The size of the current post must be small enough not to spoil the Q factor of the photonic crystal resonator too much, yet large enough to supply electric current smoothly. In addition, it’s important to keep in mind that carriers recombine and electroluminescence (EL) emits near the p-contact region due to low mobility of the holes. In collaboration with our colleagues, we placed the submicron-size p-contact post at the center of the photonic crystal slab resonator on top of the p-type InP substrate (Science 305, 1444). In the figure on the right, for n-contact, a ring-shaped ohmic metallic layer was deposited on the n-type photonic crystal slab surface.
In this heterojunction n-i-p structure, we supplied high-mobility electrons laterally from the top circular electrode, after they had traveled a few microns of distance. Holes were injected directly through the bottom post and stayed in the proximity of the central post, owing to their low mobility. Accordingly, electrons and holes tended to recombine radiatively near the central post, thereby achieving efficient coupling of the electroluminescent photons and the resonant mode profile.
Since the HCl wet-etching selectivity is excellent in an InGaAsP/InP material system, the post size could be successfully and gradually controlled and reduced by increasing the wet-etching temperature from 10° C to room temperature (Appl. Phys. Lett. 90, 171122). In addition, because the InP layer underneath the larger air-hole is etched away faster, the heterogeneous photonic crystal lattices of different air-hole sizes could be introduced to improve the position and size of the central InP post.
(a) Scanning electron microscope image of the fabricated photonic crystal cavity (top view). The lasing images captured by infrared camera of (b) the monopole mode and (c) the hexapole mode. The calculated vertical-component of Poynting vector of (d) the monopole mode and (e) the hexapole mode. (f) L-I and (g) I-V characteristics of the monopole-mode laser (Science 305, 1444 and Appl. Phys. Lett. 90, 171122).
Generally, the optical characteristics of a 2D slab photonic crystal cavity will be sensitively affected by the introduction of additional loss channels. However, the resonant mode with central intensity node can still have a high Q factor. In order to selectively excite a high-Q resonant mode, one must have a complete understanding of the resonant modes that are available in a single-cell triangular lattice photonic crystal cavity. Under pulsed current injection conditions at room temperature, we observed lasing actions—from both the monopole and the hexapole modes—in samples that were optimally designed for respective modes.
Parts (b) and (c) of the figure on the right show the near-field images of these two resonant modes captured from the top by an infrared (IR) camera. Both images clearly exhibit a central intensity minimum. In addition, we confirmed the monopole- or hexapole-mode operations by comparing the measured resonant frequencies with those obtained from the 3D finite-difference time-domain (FDTD) calculation, as shown in (d) and (e), respectively. We used numerical structural input data obtained directly from the scanning electron microscope image in the FDTD computation in order to include all the fabrication imperfections. Calculated mode profiles compared well with the measured ones.
We observed a low-threshold current of about 260 μA in the monopole-mode laser from the peak output intensity, as shown in (f). Considering that there are nonnegligible current leakage paths in the structure, the actual threshold current may be smaller. The soft turn-on shoulder near the threshold implies a large spontaneous emission factor.
We estimated a spontaneous emission factor of 0.25 by comparing spectrally integrated output intensities with L-I curves obtained from the rate equations. This value is considerably larger than those previously reported from the semiconductor nanolasers and is attributable to the effective carrier localization by electrical pumping together with the nondegeneracy and the small modal volume.
The hexapole mode reported a lower threshold of about 100 μA, thanks to the Q factor higher than that of the monopole mode. On the other hand, the Q factors of the quadrupole and dipole modes were dramatically degraded by the introduction of the central post at the antinode site. Hence, these modes were effectively suppressed from lasing.
The typical electrical characteristics of the monopole-mode photonic crystal laser are shown in (g). The turn-on voltage is less than 1.0 V and the electrical resistance is 2.2 kΩ. The relatively high resistance is mainly due to the sub-micrometer size of the p-InP post and in part due to the lateral distance between the n-electrode and the center. The current leakage can be attributed to the nonradiative recombination at the air-semiconductor air hole interfaces and at the edge of the mesa.
Electrically driven metal-coated gold-finger nanolasers
Recently, M.T. Hill and colleagues reported a metal-coated ultrasmall semiconductor laser (Nature Photon. 1, 589). Their use of a metallic cavity is of great interest because metal can tightly confine photons within a sub-wavelength-scale volume. However, the ohmic losses have been prohibitively high for laser operation in this type of nano-cavity.
Martin et al. at COBRA Research Institute have successfully demonstrated the first laser operation in an electrically pumped gold-coated nanocavity with dimensions of 200 to 300 nm. This ultrasmall laser exhibits a low threshold current smaller than 10 μA. The metal cavity structure consists of an InP/InGaAs/InP pillar with a conventional double heterostructure surrounded by a thin insulating silicon nitride layer encapsulated by gold.
The surrounding gold forms a metallic resonator around the InGaAs active layer emitting at telecommunication wavelengths. Electrons are injected through the top of the pillar. And holes are injected through the p-InGaAsP layer connected to a large-area lateral contact.
The smallest dielectric photonic crystal cavity has a mode volume of ~2 (λ/2n)3 (Opt. Express 15, 7506). This metal nano-cavity has a much smaller mode volume of 0.38 (λ/2n)3. However, the Q factor of the metal cavity is very small at room temperature, and lasing was not observed.
However, at low temperatures (10 to 77 K) where metallic losses are reduced, unambiguous lasing operation was achieved. Finite difference time domain computations of the fabricated pillar show Q = 268 for 10 K and Q = 144 for 77 K. In fact, if one is able to realize a low-loss silver nanocavity, higher temperature lasing is expected. The realization of a metallic nanocavity with a very small modal volume may create new opportunities for coherent all-optical processing in ultra-small plasmonic circuits.
(Left) Schematic of the metal-coated gold-finger laser. (Right) Scanning electron microscope image of the fabricated structure (Nature Photon. 1, 589).
The electrically driven wavelength-scale photonic crystal laser with a large Purcell factor represents a meaningful step toward the ultimate light source. However, many issues must still be addressed. For example, the quantum wells that are widely used in photonic crystal lasers as an active material must be replaced by other active semiconductor materials with a narrower emission linewidth, such as quantum dots or nanowires. The incorporation of quantum dots would reduce non-radiative recombination and thus the lasing threshold.
In addition, the system that combines a single quantum dot with a high-Q photonic crystal cavity could be a good platform for the study of strong-coupling effects (Nature 432, 200 and Nature 432, 197). A semiconductor nanowire would also be an attractive material for forming a novel hybrid nanostructure. The typical optical nanowire is made of a single-crystal semiconductor with a diameter of tens of nanometers and lengths up to tens of micrometers.
Hybrid structures that merge nanowire building blocks with top-down fabricated photonic crystals can offer a solution to couple light into and out of nanowires (Nano Lett. 6, 11). Efficient light localization and light suppression are expected from the hybrid semiconductor nanowire photonic crystal structure.
The other important issue is the overlap of the cavity resonance and the quantum dot resonance, spectrally and spatially. The resonant behaviors of a photonic crystal laser depend critically on the precision of nanofabrication. In general, the resonance of a photonic crystal cavity is spectrally deviated from the design value, owing to the unavoidable imperfections accumulated during crystal growth and fabrication.
Researchers have proposed several post-processing fine spectral tuning schemes (Appl. Phys. Lett. 87, 021108). For instance, there have been reports of wet chemical digital etching and atomic force microscope nano-oxidation techniques in the photonic crystal structures. Even after this spectral fine tuning, the quantum dot must still be placed at the antinode of a resonance in order to maximize the interaction of two resonances.
The recently demonstrated microfiber-based reconfigurable photonic crystal resonator could be one of the viable solutions (Opt. Express 15, 17241). Here, the position of a photonic crystal cavity can be repeatedly defined by relocating a curved microfiber along a linear photonic crystal waveguide. However, researchers are actively pursuing more flexible nano-tuning techniques.
One of the challenges is collecting valuable photons confined within a wavelength-scale high-Q resonator into free space or a fiber. S.-H. Kim et al. have demonstrated efficient vertical beaming based on the symmetry perturbation of a specific resonant mode (Phys. Rev. B 73, 235117). Localized photons have also been shown to be funneled horizontally into a microfiber or a neighboring low-loss photonic crystal waveguide. The microfiber-coupled photonic crystal laser system is another example toward this end.
The “practical” single photon source is expected to be first realized through optical pumping. For example, when the background emissions are minimized, the resonant pumping of a selected quantum dot has demonstrated a respectable quantum efficiency of 97 percent (Phys. Rev. Lett. 98, 117402). S. Strauf and colleagues have also recently reported the single photon source with a repetition rate up to 31 MHz and extraction efficiency of 38 percent using negatively charged excitons (Nature Photon. 1, 704).
Figuring out how to operate these lasers at higher temperatures is also a nontrivial problem. For the electrically driven version, the critical issue of the digital carrier injection must also be addressed. Although there are formidable obstacles to overcome, we believe that the ultimate single photon source could be realized in the photonic crystal cavity in the not-so-distant future.
[Hong-Gyu Park is with the department of physics at Korea University in Seoul. Se-Heon Kim is with the department of chemical and biomolecular engineering at KAIST. Min-Kyo Seo and Yong-Hee Lee are with the department of physics at KAIST in Daejeon, South Korea.]
References and Resources
>> O. Painter et al. “Two-dimensional photonic band-gap defect mode laser,” Science 284, 1819 (1999).
>> S. Noda et al. “Polarization mode control of two-dimensional photonic crystal laser by unit cell structure design,” Science 293, 1123 (2001).
>> H.-G. Park et al. “Nondegenerate monopole-mode two-dimensional photonic band gap laser,” Appl. Phys. Lett. 79, 3032 (2001).
>> R. Colombelli et al. “Quantum cascade surface-emitting photonic crystal laser,” Science 302, 1374 (2003).
>> H.-G. Park et al. “Electrically driven single-cell photonic crystal laser,” Science 305, 1444 (2004).
>> J.P. Reithmaier et al. “Strong coupling in a single quantum dot-semiconductor microcavity system,” Nature 432, 197 (2004); T. Yoshie et al. “Vacuum Rabi splitting with a single quantum dot in a photonic crystal nanocavity,” ibid., 200 (2004).
>> K. Hennessy et al. “Tuning photonic crystal nanocavity modes by wet chemical digital etching,” Appl. Phys. Lett. 87, 021108 (2005).
>> B.-S. Song et al. “Ultra-high-Q photonic doubleheterostructure nanocavity,” Nature Mater. 4, 207 (2005).
>> C. J. Barrelet et al. “Hybrid single nanowire photonic crystal and microresonator structures,” Nano Lett. 6, 11 (2006).
>> S.-H. Kim et al. “Vertical beaming of wavelength-scale photonic crystal resonators,” Phys. Rev. B 73, 235117 (2006).
>> M.-K. Kim et al. “Reconfigurable microfiber-coupled photonic crystal resonator,” Opt. Express 15, 17241 (2007).
>> M.T. Hill et al. “Lasing in metallic-coated nanocavities,” Nature Photon. 1, 589 (2007).
>> K. Nozaki et al. “Room temperature continuous wave operation and controlled spontaneous emission in ultrasmall photonic crystal nanolasers,” Opt. Express 15, 7506 (2007).
>> D. Press et al. “Photon antibunching from a single quantum-dot-microcavity system in the strong coupling regime,” Phys. Rev. Lett. 98, 117402 (2007).
>> M.-K. Seo et al. “Low threshold current single-cell hexapole mode photonic crystal laser,” Appl. Phys. Lett. 90, 171122 (2007).
>> S. Strauf et al. “High-frequency single-photon source with polarization control,” Nature Photon. 1, 704 (2007).