Strong Coupling and Its Dynamic Control of Distant Nanocavities

Susumu Noda (left) and Yoshiya Sato (right)



Authors: Susumu Noda and Yoshiya Sato

Affiliation: Department of Electronic Science and Engineering, Kyoto University, Japan

Noda's Quantum Optoelectronics Laboratory >>

The dynamic manipulation of photons in nanostructures is essential for various applications including advanced photonic circuits, stopping light, and quantum information processing. In particular, the formation and dynamic control of a coupled state among on-chip, photonic nanoelements at arbitrary positions should have great impact on these directions. However, when we couple individual nanocavities, they must be placed in close proximity since the light is confined so tightly in each cavity that their evanescent fields extends only a few microns. With this limitation, a nanocavity can only couple to adjacent nanocavities, which restricts the architecture of the system and makes it difficult to achieve on-demand dynamic control of the coupled states.

In our recent work, we have obtained strong coupling between nanocavities even though they are separated by a large distance (> 80 μm) and achieve dynamic control over their coupling to freeze the photon state on demand. Photonic states can now be separated without being isolated, opening the door to the development of advanced functional photonic circuits in scalable classical and quantum information processing. These results have recently been published in the journal "Nature Photonics"[1]





















Fig. 1.a, Schematic model of two indirectly coupled photonic nanocavities through a waveguide with reflecting boundary walls. b, The resonant spectrum of the isolated individual nanocavities A and B (red line) and their resonant spectrum when coupled to an open waveguide (dashed green line). The green line spectrum has a line width of δ_in, which corresponds to the coupling bandwidth between the nanocavities and the waveguide. c, The resonant spectrum of the bounded waveguide discretized by Fabry-Perot (FP) resonant effect.

At first, we discuss how to realize strong coupling between distant nanocavities. We employ a system as shown in Fig. 1. Two nanocavities (A and B) are connected by a waveguide, which is terminated on both sides by reflecting walls (C and D). The individual nanocavities each have a single resonant mode with the same frequency (see Fig. 1b, red line). The bounded waveguide has many standing wave modes due to Fabry-Perot (FP) resonance (see Fig. 1c). To realize strong cavity-cavity coupling, we theoretically investigated the system in detail and found that all FP waveguide modes should be detuned far from the nanocavity modes (by more than the coupling bandwidth δ_in between the nanocavities and the waveguide) as shown in Fig. 1b and c. Even under such a condition, the nanocavities can still couple to each other indirectly through a forced oscillation of the FP waveguide modes, while concentrating photons in either nanocavity, not the waveguide.

Fig. 2. a, Overall view of the fabricated silicon-based photonic crystal observed by SEM. Two photonic crystal nanocavities (A & B) are placed 202a apart with a line defect waveguide nearby. b, Magnified image of a nanocavity. This is based on a multi-step hetero structure with a1 of 415 nm and a2 of 420 nm. c, Magnified image of the waveguide and a partial reflector. d, e Spectra of the vertically emitted light observed at cavities A (d) and B (e) respectively, obtained by introducing a tunable continuous-wave laser through partial reflector C. f, Time resolved amplitude of vertically emitted light from cavities A & B. A pulse laser with duration of 4 ps and a centre wavelength of ~ 1539.45 nm (width = 1 nm) was introduced through the partial reflector C, and vertically emitted light from each cavity was observed in the time domain by a cross-correlation method.

According to this scheme, we fabricated a silicon-based photonic crystal sample as shown in Fig. 2a-c. Two multistep-hetero nanocavities[2], A and B, with original Q factors of ~ 1 million, were placed 83µm apart, with a line defect waveguide nearby. Both ends of the waveguide were bound by partial reflectors (C and D), formed by narrowing the waveguide’s width. Figure 2d, e show resonant spectra of the fabricated sample observed from nanocavity A and B, respectively. Two resonant peaks with similar intensities were observed from both cavities with exactly the same wavelengths (1539.39 and 1539.54 nm). These peaks correspond to the coupled nanocavity modes. The splitting of the peaks (150 pm) is 50 times larger than the resonant peak’s width (~3pm), indicating that the system is within the strong coupling regime.

Next, we carried out time domain measurements. We excited the coupled cavity modes by introducing a short optical pulse through partial reflector C. The results are shown in Fig. 2f. Clear exchange of photons between the distant nanocavities was observed with a period of ~54 ps. This exchange is seen to continue more than 400 ps, demonstrating the long coherence time of photons in this system. Note that the photon lifetime is much larger than that of the FP waveguide modes (~40ps). This indicates that the photons are predominantly concentrated in the nanocavities rather than in the waveguide when the coupled nanocavity modes are excited.























Fig. 3. Dynamic control of the coupling state between the nanocavities. a, Schematic of the experimental set-up. A control pulse is irradiated into cavity B, causing a dynamic wavelength shift by the carrier plasma effect. b, Time-resolved amplitude of emitted light from cavities A and B, where the control pulse was irradiated into cavity B when the photons populated only cavity A.

The next important step was to demonstrate dynamic control over these coupling states. Because the nanocavities are sufficiently far apart, it is possible to induce a dynamic change in either cavity, completely independently, to control the coupling state. Here, we attempted to induce a wavelength shift of cavity B using a control pulse with duration of 4 ps, and a wavelength of 770 nm at a specified time during the photon exchange. (See Fig. 3a) The control pulse is absorbed by the cavity, generating free carriers, which lowers the refractive index and induces a blueshift of the resonant wavelength [3]. This blueshift cuts off the coupling between cavities A and B, and this decoupled state continues for ~1 ns due to the long carrier lifetime in silicon. Figure 3b shows the results obtained for the control pulse irradiation. The control pulse is irradiated onto the cavity B with the photons populating only cavity A. This figure clearly shows that the photon exchange was stopped successfully, and froze the photon population in the state at the moment of control pulse irradiation. This finding suggests that the behaviour of photons can be controlled even in regions where they are not present which therefore enables remote control of photons. We have also observed similar phenomena even at single photon power levels.

The results obtained in this work are expected to be applicable to various nanophotonic circuits that require distant coupling of on-chip cavities and will become a fundamental building block for areas including the stopping (or slowing) of light and even photonic quantum information processing.

References
[1] Y. Sato, Y. Tanaka, J. Upham, Y. Takahashi, T. Asano, and S. Noda “Strong coupling between distant photonic nanocavities and its dynamic control”. Nature Photonics. 6, 56-61 (2012). Abstract.
[2] Yasushi Takahashi, Yoshinori Tanaka, Hiroyuki Hagino, Tomoyuki Sugiya, Yoshiya Sato, Takashi Asano, and Susumu Noda, "Design and demonstration of high-Q photonic heterostructure nanocavities suitable for integration". Optics Express. 17, 18093-18102 (2009). Abstract.
[3] Yoshinori Tanaka, Jeremy Upham, Takushi Nagashima, Tomoaki Sugiya, Takashi Asano & Susumu Noda, "Dynamic control of the Q factor in a photonic crystal nanocavity". Nature Materials, 6, 862-865 (2007). Abstract.