Electrically Controlled Silicon-based Photonic Crystal Chromatic
Dispersion Compensator with Ultra Low Power Consumption

Ching Eng Pnga,*, Gi Ho Parka, Soon Thor Lima, Er Ping Lia,
Aaron J. Dannerb, Kensuke Ogawac, Yong Tsong Tanc,
a Advanced Photonics and Plasmonics, IHPC, A*Star, 1 Science Park Rd, #01-01, Singapore 117528.
b Department of ECE, National University of Singapore, 4 Engineering. Drive 3, Singapore 117576
c Optics and Electronics Lab., Fujikura Ltd, 1440 Mutsuzaki, Sakura, Chiba, 285-8550, Japan



We show full 3-Dimensional (3D) electrical and optical simulation of a tunable silicon-based Photonic Crystal (PhC) Chromatic Dispersion Compensator (CDC) with high power efficiency and ultra-low power consumption (114nW), operating at a speed of 40.5MHz. The device exploits a structure where the optical field maximum is not in a PhC waveguide, but rather in a hybrid Si3N4/Si/SiO2 structure that will allow greater ease of fiber coupling due to larger mode size and reduced loss. The CDC is broadband, and produces constant 2nd order chromatic dispersion over an optical communication band such as C-band.



There is intense interest in CDCs for high-speed optical-fiber telecommunications as chromatic dispersion plays a significant role in the propagation of short optical pulses. This work presents results on the modeling and accuracy of a silicon-based actively-tunable PhC CDC with ultra-low power consumption (a three order-of-magnitude reduction relative to other reported devices).

Chromatic dispersion is a serious problem in optical fiber communications, and dispersion compensators in the past have made use of various structures including all-pass filters [1, 2], fiber gratings [3-5], use of higher-order modes [6], or other grating/filter-based devices [7-10]. Electrically controlled PhCs have recently shown promise for use in Mach-Zehnder interferometers [11-12] as well as CDCs [13-14]. Our device’s operation primarily exploits two physical effects, the first being strong dispersion at the band edge near a PhC bandgap [15], and the second being our ability to adjust the position of this band by changing the material refractive index of the PhC host material electrically through the plasma dispersion effect [16].

To reduce the large losses commonly associated with PhC devices and avoid excessive mode field mismatch (for example, between an input fiber and a PhC slab), we present a structure whereby the PhC active region intercepts only a portion of the optical field in cross section, with the bulk of the field contained within a lowloss Si3N4 rectangular core. We demonstrate 3D simulation of this type of CDC and show that power consumption can be reduced by three orders-of-magnitude (114nW versus 113μW) relative to another PhC structure [11]. The device has been simulated in 3D both electrically and optically, showing a bandwidth of 40.5MHz at a wavelength of 1550nm.



The waveguide is shown schematically in Figure 1(a) and consists of a low-loss Si3N4 rectangular core bounded on one side by a 2D Si/SiO2 PhC layer over which a voltage may be applied, and on the other three sides by SiO2 cladding. Due to a thin layer (100nm) of Si/PhCs for electronic transport and a separate Si3N4 core for the bulk of the optical waveguiding, the electrical and optical properties can be individually tailored. This may enable low-cost and highly efficient CDCs.

The low-loss Si3N4 waveguide is 1μm wide and 400nm thick, and the PhC consists of a 2D triangular lattice of 242nm-diameter SiO2 pillars embedded in a host Si layer. The lattice pitch is 403nm in figure 1(b). The 100nm-thick Si layer serves as host medium for the PhC layer, and a 1μm-thick buried oxide layer provides a lower cladding [13]. The electrical transport characteristics of the device are significantly altered due to the presence of the PhCs. Since their periodicity is in both the x- and z-directions but the anode and cathode are separated along the y-axis, 3D electrical and optical simulations are necessary. A PhC band diagram of the composite structure, including the Si3N4 core, along the direction of propagation is shown in figure 1(c) along with the operating point (center wavelength) on the first band. This operating point shifts slightly with applied voltage, and though not visible in figure 1(a) and figure 1(b), the shift is significant enough to exploit.

Figure 1: (a) (Color online) Cross sectional schematic of the CDC. (b) Top-view schematic of the CDC with dimensions a~403nm, b~693nm, and d~242nm. (c) Photonic band diagram of the composite structure along the propagation direction. The light cone for the SiO2 cladding is located on the upper left portion; operating at 1.55μm as indicated by the dotted horizontal line, with a/λ = 0.2579 (OFF) or 0.2633 (ON). The dark solid lines represent even symmetric modes about the y-axis at the centre of the Si3N4 core. Dotted lines denote odd symmetric modes. (d) Top-view and cross-section optical mode profile for operation along the first photonic band. The optical mode, initially guided only in the Si3N4 layer (outside the CDC), is guided in the hybrid structure within the CDC as shown.


A CDC length of 1.35μm was used in electrical simulation, which is sufficient to account for carrier movements in the PhC lattice. In total, the CDC is physically 2mm long, and this value was used in all optical simulations. This length is necessary to obtain sufficient phase shift at this low power. The anode and cathode dopant windows (located in the 100nm-thick Si layer) are 2μm wide and doped to a concentration of 1020/cm3 with boron and arsenic respectively with an intrinsic spacing of 2.4μm in between, creating a p-i-n structure. The host silicon’s refractive index is controlled by the plasma dispersion effect (via the applied voltage) which then influences the effective guiding refractive index of the composite structure. This causes the operating point shift referred to above and enumerated in figure 1(c).

The optical field for operation along the first photonic band is illustrated in figure 1(d) where a frequency domain method [17] was employed in its calculation. The x-boundary of the Si3N4 layer was not considered in the calculation of the optical field distribution, since its effect on the vertical field distribution is minimal.



Due to the fact that 2D methods cannot be reliably used to simulate the electrical characteristics [14], a 3D device simulator from SILVACO was utilized to model the induced charge density as a function of the drive voltage, incorporating physical models to account for carrier statistics, doping and field dependent carrier mobilities, and doping dependent carrier recombination lifetimes.

Figure 2 shows the injected electron and hole concentrations and hence the change in refractive index, and subsequently the phase change which varies nonlinearly with applied voltage. This nonlinearity is induced by the sub-linear dependence of the change in free holes with the change in refractive index. As the CDC is driven harder, more free carriers are injected into the intrinsic region of the device. This results in a reduced lifetime in this region. Hence, the modulator must be driven harder to achieve an equivalent refractive index change than at lower drive powers.

Figure 2: (Color online) Predicted relationship between injected carriers and drive current along with the voltage characteristics of the CDC. Turn-on power, Pon is approximately 114nW. Left inset shows the entire dopant and current relationship with voltage from 0 to 10V. Right inset shows I-V characteristics that 2D simulation (the case without photonic crystals) would have overestimated the power usage, especially during fast switching when the anode is driven higher.


Figure 2 also shows the CDC I/V, and of special interest is the ON state with injected concentration of ~3 × 1017cm-3, which in turn translates to a refractive index change of ∆ n~10-3, and the OFF state corresponds to no carrier injection state, i.e., no carrier injection. At high concentration levels (>1017cm-3), the recombination dynamics of the photogenerated electron-hole pairs are dominated by nonradiative energy transfer to free carriers, which are transferred to excited states in the valence band (for holes) and conduction band (for electrons). This is the Auger effect which decreases the population of excited carriers. The current and voltage required to achieve the ON state are 124nA and 0.92V respectively; which results in a turn-on power of approximately 114nW. This is a power reduction of three orders-of-magnitude compared to a 2D Si-based PhC result of 113μW [11]. The turn-on power is indicated by an arrow highlighting the voltage point of 0.92V, with the left inset in figure 2 showing the full voltage sweep from 0 to 10V. The right inset of figure 2 shows the simulated I-V electrical characteristics of the CDC with and, hypothetically, without the PhCs. The large difference between the curves illustrates that 3D simulation methods were critical in this case. A CDC of this type has been fabricated, and full details will be published after testing is completed. Scanning electron micrographs of the fabricated device are available in references [13-14].

The refractive index change, and hence the phase modulation frequency, is determined by the rate of carrier density modulation in the photonic device which consequently can be determined by modeling the transient response to predict the device speed [18]. We employ a step-like drive voltage to the device and calculate the time dependence of the induced charge density in the electrically active Si region. Firstly, both anode and cathode were first zero biased for 10 ns, followed by a step increase to Von for 200ns and a subsequent step decrease to 0V. Von (~0.92V) is the voltage corresponding to an index change of 10-3 as reported previously. The rise time, tr is defined as the time required for the induced refractive index (RI) to change from 10% to 90% of the maximum value. Likewise, the fall time is defined as the time required for the induced refractive index change from 90% to 10% of the maximum value. The rise and fall times were determined to be 5.39ns and 8.65ns respectively. As the fall time is slower, it is used to quantify the modulation bandwidth, and is estimated to be ~40.5 MHz. Interestingly, the rise time is shorter than the fall time which is not typical of injection based devices. This could be due to the field induced carrier drift and the thickness of the modulation layer where injection effects are not expected to be as fast as other plasma dispersion mechanisms such as in depletion devices [19-20]. Although the device speed reported here is modest compared to high-speed optical modulators [21], such devices are still fast in application to dynamic compensation of chromatic dispersion. The small variation in refractive index causes a significant shift of spectral phase and chromatic dispersion.

Figure 3 demonstrates the effects of phase change and dispersion (Figure 3 inset) with respect to operating frequency for a biased and unbiased setting, with the biased setting corresponding to a refractive index change of 10-3, requiring a power of approximately 114nW. The coefficient of the quadratic dispersion term for the unbiased CDC is 3.1678ps2 from curve fitting, and leads to dispersion parameter D = -0.6297ps/nm/cm whereas for the biased CDC the quadratic-term is 3.3947ps2 and D = -0.675ps/nm/cm. The quadratic terms of the spectral phase has been increased by 7.16% and the dispersion parameter increased by 7.19%.

Figure 3: (Color online) Simulated relative phase and dispersion shift of the CDC as a function of operating frequency in biased and unbiased settings.




In conclusion, we have predicted a 3D device simulation of a highly efficient Si-based PhC CDC with power reduction of three orders-of-magnitude. We have also demonstrated that 3D simulation is critical, especially for active optical devices with structures that are nonhomogenous in the light propagation direction. The results presented also show that it is feasible to make a CDC compensator with both the optical and electrical behavior capably of being individually tailored.



The authors acknowledged funding by the A*Star Flagship Programme and Fujikura Ltd. This work was supported in part by the National University of Singapore/Ministry of Education Research Grant R263000414112/133.



  1. C. K. Madsen, G. Lenz, A. J. Bruce, M. A. Cappuzzo, L. T. Gomez, and R. E. Scotti, IEEE Photon. Tech. Lett., 11, 1623 (1999).
  2. L. M. Lunardi, D. J. Moss, S. Chandrasekhar, L. L. Buhl, M. Lamont, S. McLaughlin, G. Randall, P. Colbourne, S. Kiran, and C. A. Hulse, IEEE J. Lightwave Tech., 20, 2136 (2002).
  3. B. J. Eggleton, A. Ahuja, P. S. Westbrook, J. A. Rogers, P. Kuo, T. N. Nielsen, and B. Mikkelsen, IEEE J. Lightwave Tech., 18, 1418 (2000).
  4. K. O. Hill and G. Meltz, IEEE J. Lightwave Tech., 15, 1263 (1998).
  5. M. Onishi, Y. Koyano, M. Shigematsu, H. Kanamori, and M. Nishimura, Electron. Lett., 30, 161 (1994).
  6. C. D. Poole, K. T. Nelson, J. M. Wiesenfeld, and A. R. McCormick, Opt. Lett., 17, 985 (1992).
  7. K. Takiguchi, K. Okamoto, and K. Moriwaki, IEEE J. Lighwave Tech., 14, 2003 (1996).
  8. M. Shirasaki, IEEE Photon. Tech. Lett., 12, 1598 (1997).
  9. D. T. Neilson, R. Ryf, F. Pardo, V. A. Aksyuk, M. E. Simon, D. O. Lopez, D. M. Marom, and S. Chandresekhar, IEEE J. Lightwave Tech., 22, 101 (2004).
  10. C. R. Doerr, S. Chandrasekhar, and L. L. Buhl, IEEE Photon. Tech. Lett., 20, 560 (2008).
  11. Y. Jiang, W. Jiang, L. Gu, X. Chen, and R. T. Chen, Appl. Phys. Lett., 87, 221105 (2005).
  12. L. Gu, W. Jiang, X. Chen, L. Wang, and R. T. Chen, Appl. Phys. Lett., 90, 071105 (2007).
  13. K. Ogawa, K. Tomiyama, Y. T. Tan, M. T. Doan, Y. M. Bin, D. L. Kwong, S. Yamada, J. B. Cole, Y. Katayama, H. Mizuta, and S. Oda, Optical Fiber Conference, OThE4, 2006.
  14. C. E. Png, G. H. Park, S. T. Lim, E. P. Li, A. J. Danner, M. B. Yu, K. Ogawa, Y. T. Tan, Conference on Lasers and Electro-Optics, JThA98, 2008.
  15. M. Notomi, K. Yamada, A. Shinya, J. Takahashi, C. Takahashi, and I. Yokohama, Phys. Rev. Lett., 87, 253902 (2001).
  16. R. Soref and B. Bennett, IEEE J. of Quantum Elect., 23, 2159 (1987).
  17. S. G. Johnson and J. D. Joannopoulos, Optics Express 8, 173 (2001).
  18. A. Liu, D. Samara-Rubio, L. Liao, and M. Paniccia, IEEE J. Sel. Topics in Quantum Elect., 11, 367 (2005).
  19. G. T. Reed and C. E. Png, Mater. Today 8, 40 (2005).
  20. G. T. Reed, Silicon Photonics: the state of the art, Chapter 4, Optical Modulators in Silicon Photonic Circuits, Wiley ISBN: 978-0-470-02579-6 (2008).
  21. A. Liu, R. Jones, L. Liao, D. Samara-Rubio, D. Rubin, O. Cohen, R. Nicolaescu, and M. Paniccia, Nature (London), 615 (2004).


Download PDF version of this article

Copyright © 1984 - Silvaco, Inc. - Trademarks - Privacy Policy
facebook youtube twitter LinkedIn

facebook youtube twitter