We demonstrate all-optical bit memory operation with photonic crystal (PhC) nanocavities based on an InGaAsP substrate with a band gap at a wavelength of about 1.3 µm. The optical bistability is based on a refractive index modulation caused by carrier-plasma dispersion. The operating energy required for switching is only 30 fJ, and the minimum optical bias power for bistability is 40 µW, which is about one hundred times less than that required for laser based bistable memories.
© 2008 Optical Society of America
In recent years, the increase in device power consumption has become a new bottleneck as regards telecommunication networks along with the rapid increase in communication traffic. For instance, since signal processing in packet routers depends completely on the electric circuit, the power consumption increases rapidly with increases in traffic because of the large number of photoelectric converters equipped with interconnections in the device and the I/O. To overcome this problem, it is thought that an optical memory must be used for packet processing to minimize the use of optical to electrical conversion .
There has been a lot of research on optical buffer memories based on long optical fibers, and optical bit memories based on semiconductor bistable lasers [2-4]. However, the integration of the former is impossible, and the power consumption of the demonstrated laser memories is fairly large. Moreover, it is difficult to integrate these memories on one small chip. Therefore, the development of a micro optical memory with the potential for large-scale integration on one chip has become a critical issue in the field of optical information processing.
Photonic crystal (PhC) is a promising candidate as a platform on which to construct devices with dimensions of several wavelengths for future photonic integrated circuits. A PhC in a photonic band gap functions as a light insulator. A line defect and a point-defect cavity in a PhC form a waveguide and a resonator, respectively. We have focused on a coupled resonator-waveguide system on a PhC platform. The cavity mode volume in the PhC is extremely small at 0.1~0.2 µm3, and such a small size is difficult to realize without a PhC. Moreover, it is effectively coupled with a single-mode PhC waveguide (WG). The photon density in the cavity can be extremely high with a low input power owing to its small size, and this results in a large optical nonlinearity [5,6]. As a result, we can realize an ultrasmall alloptical bistable memory. Moreover, it is easy to integrate these elements on a one-chip PhC platform as a future all-optical random access memory (RAM) system.
We have already realized all-optical bistable memory operation using a Si-based PhC nanocavity . The power consumption for the operation is very low at 0.4 mW, but the memory holding time is limited to less than 2.5 ns. This is due to accumulated heat in the cavity. The dominant optical nonlinearities in our memory derive from a carrier plasma effect and a thermo-optic (TO) effect. Optically generated carriers reduce the refractive index of the cavity with a relaxation time of around 100 ps. On the other hand, accumulated heat caused by carrier relaxation increases the refractive index of the cavity with a sub microsecond relaxation time. That is, the memory function formed by the carrier plasma effect is gradually eliminated by the TO effect. This is why the Si-PhC memory holding time is very short.
To solve this problem, we have substituted InGaAsP for Si. The advantage of InGaAsP is that the refractive index change caused by the carrier plasma effect is larger than with Si [8,9] because the value is inversely proportional to the electron mass. Moreover, since the two-photon absorption (TPA) [10,11] is much larger than that of Si , the input light generates a carrier more effectively in InGaAsP-PhC than in Si-PhC. Namely, an InGaAsP-PhC cavity can function as an optical memory with much lower power than a Si-PhC cavity, and the memory function of an InGaAsP-PhC cavity based on the carrier plasma effect is not easily eliminated by the TO effect.
The KAIST group has already realized an all-optical bistable memory based on an InGaAsP-PhC nanocavity . Since the band gap of their InGaAsP slab is very close to the operation wavelength of 1.55 µm, the nanocavity can function as an optical memory with a very low power of 30 µW because of the large linear absorption. However, the absorption is too large for constructing WGs in the PhC. So they inputted light directly into the PhC cavity from the optical fiber. It is an excellent technique with which to measure the cavity characteristics, but it is not suitable for composing an optical integrated circuit based on a photonic crystal platform. Moreover, although they achieved bistable memory operation whereby bit information can be written by writing pulses, it has not yet been proven that the information can be read out by a reading pulse or that it does not disappear after it is read, which are indispensable bit memory functions for an optical RAM system.
In this report, we adopt an InGaAsP slab with a band gap whose wavelength is around 1.3 µm as a PhC core to suppress the optical absorption in the PhC-WG and achieve the bit memory operation of the PhC nanocavity.
2. Experiments and discussions
Figure 1 shows a scanning electron micrograph of our PhC fabricated on an InGaAsP substrate by a combination of electron beam lithography and ICP dry etching. The air hole size and core thickness are both 200 nm. We adopted a width-tuned line defect cavity so that the air holes surrounding the cavity were shifted away from the center of the line defect [14,15]. The cavity mode volume is very small at 0.16 µm3 with a calculated Q-factor of over 10 million, and we have realized a Q-factor of 1,800,000 in a Si-based PhC nanocavity . A light is inputted from a lensed fiber to a PhC-WG with a core of 1.05 W through an input/output waveguide with a 3 µm wide PhC line-defect and a spot size converter with a 15 µm long elliptic arc . Here W is the basic line-defect width, which is defined as the distance between the center of adjacent holes of a 3 and a is the lattice constant of the PhC.
The resonant wavelength of the cavity is proportional to the hole distance. We tuned it from 420 to 440 nm while keeping the hole size constant. Figure 2 shows Q-factors for several distances. The Q-factors around a wavelength of 1.56 µm are over 100,000, and as far as we know, the maximum value of 130,000 is the largest Q-factor ever reported for an InP-based PhC nanocavity. The Q-factor of the device is determined by the intrinsic Q-factor of the cavity and the coupling efficiency between the cavity and the waveguides. The tendency for the Q-factor to decrease reveals that the coupling becomes stronger as the wavelength increases. This is due to the elongation of the cavity mode. So the intrinsic Q-factor is larger than 130,000 in all cases. The merit of using low loaded Q samples based on a high intrinsic Q is that they have high transmittance, and so we used them in the experiments.
Figure 3 shows the hysteresis responses for several wavelengths. The wavelength is detuned from the center of the resonant wavelength of 1585.1151 nm to a shorter wavelength because the optically generated carrier reduces the refractive index of the cavity and shifts the resonance toward a shorter wavelength. The input signals are triangular pulses with a 100 ns base. The hysteresis response appears at an input power of only a few tens of µW. This means that the InGaAsP-PhC nanocavity functions as an all-optical bistable memory operating at a lower power than Si . The jump-down power of the hysteresis loop is shifted toward the higher power direction as a result of the accumulated heat in the cavity. The jump-down power is determined by measuring the time at which the ON state cavity turns abruptly OFF as the input power is gradually decreased. But since heat is gradually accumulated in the cavity over a period of 100 ns and the TO effect breaks the bistable condition based on the carrier plasma effect, the cavity turns OFF earlier than expected, which leads to an overestimation of the jump-down power. This means that the bistability can be formed even with a very low bias power and before the heat is accumulated sufficiently in the cavity.
Figure 4(a) shows the all-optical bistable memory operation of an InGaAsP PhC nanocavity. We used the input CW light as a bias and set the power slightly below the bistable threshold power, and we input a set pulse and a reset pulse at 0 and 50 ns, respectively. The bias and the set pulse were generated by the same laser source and were formed by the same optical modulator. The reset pulse was formed by cutting a bias for 500 ps. If there is no set pulse, the output signal level remains low. When a set pulse is applied, the output signal turns ON and stays in the ON state until the bias light is reset. The lowest bias power for bistability is 40 µW, which is ten times lower than that of Si-PhC  and a hundred times lower than that of bistable laser based memories [2-4]. Moreover, the cavity can respond to a very low energy pulse of 30 fJ with a 100 ps width. It should be noted that although our InGaAsP slab is almost transparent in the C-L band, the operating bias power is comparable to that reported by the KAIST group . This is because the Q-factor of 53,000 is much larger than theirs (2,200).
The detuning value at the lowest bias power is -0.190 nm. According to Fig. 3, the jump-down point of the hysteresis loop at this detuning is higher than 40 µW, this means that the cavity cannot function as a bistable memory with a 40 µW bias, which does not correspond to the result in Fig. 4(a). This is because of the accumulated heat in the cavity as described above. The result shows that a bistable condition can be formed with a 40 µW bias power before the heat has accumulated sufficiently in the cavity.
The memory holding time in Fig. 4(a) is 85 ns and can be improved by increasing the bias power. The longest memory time is 150 ns when the bias power is 250 µW as shown in Fig. 4 (b). With Si-PhC, the memory holding time is 1 ns with a minimum bias power of 400 µW. The holding time can be improved but will always be shorter than 2.5 ns even if the bias power reaches 750 µW. Our result shows that an InGaAsP-PhC nanocavity can achieve a long memory holding time that is 60 times longer than that with Si even at one third of the bias power. Since the memory holding time is limited by the accumulated heat in the cavity, we must attempt to suppress the heat generation, or to minimize the thermal resistance so that the heat easily escapes from the cavity thus extending the memory holding time to infinity.
Figure 5 shows the reading and writing operations of the bit memory. Here, we used two resonant modes . One is same as the mode used in Figs. 3 and 4, and is employed as a bias and a writing (W) pulse. They are in phase with each other as in Fig. 4. The resonance is 1585.1151 nm and the detuning value of the bias and the W-pulse is -0.460 nm. The other mode is a higher order mode, and is used as a reading (R) pulse. The wavelength of the R-pulse is tuned to 1576.373 nm. We set the W/R pulse powers sufficiently above/below the bistable threshold power. The transmittance of the cavity is low when the cavity is OFF. Once the cavity is turned ON, the transmittance becomes high until the cavity is turned OFF. Since the R-pulse power is insufficient to turn the cavity ON, R-pulses do not pass though the cavity while the cavity is OFF. After the W-pulse has been used to turn the cavity ON, the input Rpulses pass through the cavity. Therefore, this figure shows that the ON/OFF state of the cavity can be read out by the R-pulse. The concept of using an extra resonant mode alleviates the problem of the R-pulse and the bias having to be in phase with each other when their wavelengths are the same. If this is not done, there is a possibility that the R-pulse turns the memory OFF, because the R-pulses weaken the bias power. It is also possible to alleviate the phase problem between the W-pulse and the bias. This technique makes the memory system simple because it eliminates the many steps required to make the input light signals in phase with the bias.
We developed an ultra high-Q InGaAsP-PhC nanocavity. The Q-factor is 130,000, which is the largest value ever reported for an InP based PhC nanocavity. And we demonstrated that this nanocavity can function as an all-optical bit memory operating at very low power. The power is ten times lower than that for a Si-PhC memory and a hundred times lower than that for a laser based optical memory. In addition, we improved the memory time of the PhC nanocavity from 2.5 to 150 ns. We believe that a PhC platform in which many nanocavities are cascaded or integrated on one chip is a promising candidate as a bit memory cell array for future all-optical packet switching systems.
We thank Professor K. Kitayama for fruitful discussions and we are grateful to Y. Itaya for his continuous encouragement. This work was supported by the National Institute of Information and Communications Technology (NICT).
References and links
1. K. Kitayama, S. Arakawa, S. Matsuo, M. Murata, M. Notomi, R. Takahashi, and Y. Itaya, “All-optical RAM-based buffer for packet switch,” in Proceedings of Photonics in Switching 2007, (San Francisco, California, USA, August 2007), Paper SYMP1.3. [CrossRef]
2. M. T. Hill, H. J. S. Dorren, T. de Vries, Xaveer J. M. Leijtens, J. H. den Besten, B. Smalbrugge, Y. S. Oei, H. Binsma, G. D. Khoe, and M. K. Smit, “A fast low-power optical memory based on coupled micro-ring lasers,” Nature 432, 206–208 (2004). [CrossRef]
3. H. Kawaguchi, T. Mori, Y. Sato, and Y. Yamayoshi, “Optical buffer memory using polarization-bistable vertical-cavity surface-emitting lasers,” Jpn. J. Appl. Phys. 45, L894–L897 (2006). [CrossRef]
4. M. Takenaka, M. Raburn, and Y. Nakano, “All-Optical Flip-Flop Multimode Interference Bistable Laser Diode,” IEEE Photon. Technol. Lett. 17, 968–970 (2005). [CrossRef]
5. M. Soljacic, M. Ibanescu, S. G. Johnson, Y. Fink, and J. D. Joannopoulos, “Optimal bistable switching in nonlinear photonic crystals,” Phys. Rev. E 66, 055601(R) (2002). [CrossRef]
6. M. Notomi, A. Shinya, S. Mitsugi, E. Kuramochi, and H. Ryu, “Waveguides, resonators and their coupled elements in photonic crystal slabs,” Opt. Express 12, 1551–1561 (2007). [CrossRef]
7. T. Tanabe, M. Notomi, S. Mitsugi, A. Shinya, and E. Kuramochi, “All-optical switches on a silicon chip realized using photonic crystal nanocavities,” Appl. Phys. Lett. 87, 151112 (2005). [CrossRef]
8. R. A. Soref and B. R. Bennett, “Electrooptical effects in silicon,” IEEE J. Quantum Electron. QE-23 , 123 (1987)
9. B. R. Bennett, R. A. Soref, and J. A. Del Alamo, “Carrier-induced change in refractive index of InP, GaAs, and InGaAsP,” IEEE J. Quantum Electron. 26, 113 (1990). [CrossRef]
10. H.K. Tsang, R. V. Plenty, I. H. White, R. S. Grant, W. Sibbett, J. B. D. Soole, H. P. Leblanc, N. C. Andreadakis, R. Bhat, and M. A. Koza, “Two-photon absorption and self-phase modulation in InGaAsP/InP multi-quantum-well waveguides,” J. Appl. Phys. 70, 3992–3994 (1991). [CrossRef]
11. J. Mork and A. Mecozzi, “Response function for gain and refractive index dynamics in active semiconductor waveguides,” Appl. Phys. Lett. 65, 1736 (1994). [CrossRef]
12. G. W. Rieger, K. S. Virk, and J. F. Young, “Nonlinear propagation of ultrafast 1.5 µm pulses in high-index-contrast silicon-on-insulator waveguides,” Appl. Phys. Lett. 84, 900 (2004). [CrossRef]
13. M-K. Kim, I-K. Hwang, S-H. Kim, H-J. Chang, and Y-H. Lee, “All-optical bistable switching in curved microfiber-coupled photonic crystal resonators,” Appl. Phys. Lett. 90, 161118 (2007). [CrossRef]
14. E. Kuramochi, M. Notomi, S. Mitsugi, A. Shinya, and T. Tanabe, “Ultrahigh-Q photonic crystal nanocavities realized by the local width modulation of a line defect,” Appl. Phys. Lett. 88, 041112 (2006). [CrossRef]
15. T. Tanabe, M. Notomi, E. Kuramochi, A. Shinya, and H. Taniyama, “Trapping and delaying photons for one nanosecond in an ultrasmall high-Q photonic-crystal nanocavity,” Nature Photon. 1, 49 (2007). [CrossRef]
16. E. Kuramochi, T. Tanabe, H. Taniyama, A. Shinya, and M. Notomi, “Experimental demonstration of ultrahigh-Q photonic crystal nanocavities in very thin photonic crystal barriers with air-slots,” in Extended Abstracts of 69th Autumn Meeting of The Jpn. Soc. of Appl. Phys., (Kanagawa Japan, August 2008), Paper 3p-V-17
17. W. K. Burns, A. F. Milton, and A. B. Lee, “Optical waveguide parabolic coupling horns,” Appl. Phys. Lett. 30, 28–30 (1977) [CrossRef]
18. M. Notomi, A. Shinya, S. Mitsugi, G. Kira, E. Kuramochi, and T. Tanabe, “Optical bistable switching action of Si high-Q photonic-crystal nanocavities,” Opt. Express 13, 2678–2687 (2005) [CrossRef]