Final Report Summary - OSIRIS (Optical signal regeneration in slow light silicon waveguides)
Using slow light to enhance nonlinear effects in planar waveguide structures has recently become a very promising concept in Photonics. By controlling the effective speed of propagation of optical waveguide modes, one can increase the energy density in the waveguide and thus enhance the light-matter interaction. The process of four-wave mixing (FWM) can be used in many functional integrated optical devices, for example in optical regenerators. Slow light adds an important control parameter to enhance this effect. Ideally, the FWM process scales with the fourth power of the slowdown factor (S4) if all four waves are slowed down equally. At the onset of the project, it was not known to which extent this promise could be realised in slow light photonic crystal waveguides.
Firstly, we experimentally investigated FWM in short (80 µm) dispersion-engineered slow light silicon photonic crystal (PhC) waveguides in collaboration with the Centre for Ultrahigh-bandwidth Devices for Optical Systems (CUDOS), University of Sydney. In the experiment, the pump, probe and idler signals were all positioned in a 14 nm wide low dispersion region with a near-constant group velocity of c/30. We measured an instantaneous conversion efficiency of up to 9dB between the idler and the continuous-wave probe, with 1W peak pump power and 6nm pump-probe detuning. This conversion efficiency was found to be considerably higher (> 10 ?) than that of a Si nanowire with a group velocity ten times faster. These results, supported by numerical simulations, emphasize the importance of engineering the dispersion of PhC waveguides to exploit the slow light enhancement of FWM efficiency, even for short device lengths.
Then we demonstrated the highest cw FWM conversion efficiency in slow light PhC waveguides for cw operation (-24 dB, 90 mW pump) reported to date, for waveguides as short as 396 µm and a group index of ng = 30. Flat band dispersion engineering allows for a pump-probe detuning bandwidth of 6 nm. In the waveguide, we introduced the input light via a silicon inverse taper and an SU8 polymer waveguide. The large cross-sectional area of the polymer waveguide avoids issues with facet damage that can occur for high continuous wave (cw) pump power and we measured an injection efficiency of 38 % injection for the coupler.
The result is in good agreement with a simple analytical model, which highlights the fact that the performance is limited by both the mode shape (mode shape increases 4-fold in an ng = 30 PhC waveguide compared to a nanowire) and the loss (loss limits the waveguide length to L < 1 mm). Mode shape and loss are the key constraints for obtaining the S4 enhancement that is expected in the ideal case, while we believe that dispersive effects and nonlinear losses are not an issue in these waveguides. As a result, we predict that by reducing the loss and increasing the group index to realistic values, a conversion efficiency approaching 10 dB can be achieved.
Based on these results, we have designed slow light PhC waveguides operating in a low loss and constant dispersion window of ?? = 5 nm around ? = 1565 nm with a group index of ng = 60. We experimentally demonstrated a relatively low propagation loss, of 130 dB/cm, for waveguides up to 800 microm in length. This result is particularly remarkable given that the waveguides were written on an electron-beam lithography tool with a writefield of 100 microm that exhibits stitching errors of typically 10-50 nm. We reduced the impact of these stitching errors by introducing "slow-fast-slow" mode conversion interfaces and showed that these interfaces reduce the loss from 320 dB/cm to 130 dB/cm at ng = 60. This significant improvement highlights the importance of the slow-fast-slow method and shows that high performance slow light waveguides can be realised with lengths much longer than the writing field of a given e-beam lithography tool.
Finally, we showed efficient FWM wavelength conversion in these slow light waveguides with a group index of ng = 60. Using only 15 mW cw coupled input power, we observed a conversion efficiency of-28 dB. This efficiency represents a 30 dB enhancement compared to a silicon nanowire of the same length. The efficiency did not improve for higher input power, however, as the strong enhancement of the energy density in the slow light regime causes heating, which leads to a sufficient thermal red shift to detune the waveguide. This thermal limitation can be overcome by using oxide-clad waveguides, which we demonstrated for group indices of ng = 30. Higher group indices are difficult (but possible) to achieve with oxide clad-waveguides, and we predicted conversion efficiencies approaching-10 dB, which is equivalent to that already achieved in silicon nanowires but for a 50x shorter length.
In addition, our collaborators at CUDOS, Sydney University, Australia, have used our low loss slow light waveguides in the context of ultrahigh bandwidth light (up to 160 Gb/s) for a number of novel applications of the FWM effect. These include:
(a) Ultracompact 160 Gbaud all-optical demultiplexing. We demonstrated all-optical demultiplexing of a high-bandwidth, time-division multiplexed 160 Gbit/s signal to 10 Gbit/s channels, exploiting slow light enhanced FWM in a dispersion engineered, 96 µm long planar photonic crystal waveguide. We reported error-free (bit error rate < 10-9) operation of all 16 demultiplexed channels, with a power penalty of 2: 22: 4 dB, highlighting the potential of these structures as a platform for ultracompact all-optical nonlinear processes.
(b) Ultracompact all-optical XOR logic gate. We demonstrated an ultracompact, chip-based, all-optical exclusive-OR (XOR) logic gate via slow-light enhanced FWM in a silicon PhC waveguide. We achieved error-free operation (< 10-9) for 40 Gbit/s differential phase shift keying (DPSK) signals with a 2. 8 dB power penalty. The XOR operation required = 41 mW, corresponding to a switching energy of 1 pJ/bit. We compared the slow-light PhC waveguide device performance with experimentally demonstrated XOR DPSK logic gates in other platforms and discussed scaling the device operation to higher bit-rates. The ultracompact structure suggests the potential for device integration.
(c) Slow-light enhanced correlated photon pair generation. We reported the generation of correlated photon pairs in the telecom C-band at room temperature from a dispersionengineered silicon photonic crystal waveguide. The spontaneous FWM process producing the photon pairs is enhanced by slow-light propagation enabling an active device length of less than 100 µm. With a coincidence to accidental ratio of 12. 8 at a pair generation rate of 0. 006 per pulse, this ultracompact photon pair source paves the way toward scalable quantum information processing realised on-chip.
In conclusion, the fellowship has given me an outstanding opportunity to experience world-leading research in one of the best Photonics groups in Europe, with the additional bonus of interacting with the prime Photonics consortium in Australia. I am grateful to the Marie Curie programme for having provided me with this opportunity and am sure that the established linkeages will continue after my return to China.
Firstly, we experimentally investigated FWM in short (80 µm) dispersion-engineered slow light silicon photonic crystal (PhC) waveguides in collaboration with the Centre for Ultrahigh-bandwidth Devices for Optical Systems (CUDOS), University of Sydney. In the experiment, the pump, probe and idler signals were all positioned in a 14 nm wide low dispersion region with a near-constant group velocity of c/30. We measured an instantaneous conversion efficiency of up to 9dB between the idler and the continuous-wave probe, with 1W peak pump power and 6nm pump-probe detuning. This conversion efficiency was found to be considerably higher (> 10 ?) than that of a Si nanowire with a group velocity ten times faster. These results, supported by numerical simulations, emphasize the importance of engineering the dispersion of PhC waveguides to exploit the slow light enhancement of FWM efficiency, even for short device lengths.
Then we demonstrated the highest cw FWM conversion efficiency in slow light PhC waveguides for cw operation (-24 dB, 90 mW pump) reported to date, for waveguides as short as 396 µm and a group index of ng = 30. Flat band dispersion engineering allows for a pump-probe detuning bandwidth of 6 nm. In the waveguide, we introduced the input light via a silicon inverse taper and an SU8 polymer waveguide. The large cross-sectional area of the polymer waveguide avoids issues with facet damage that can occur for high continuous wave (cw) pump power and we measured an injection efficiency of 38 % injection for the coupler.
The result is in good agreement with a simple analytical model, which highlights the fact that the performance is limited by both the mode shape (mode shape increases 4-fold in an ng = 30 PhC waveguide compared to a nanowire) and the loss (loss limits the waveguide length to L < 1 mm). Mode shape and loss are the key constraints for obtaining the S4 enhancement that is expected in the ideal case, while we believe that dispersive effects and nonlinear losses are not an issue in these waveguides. As a result, we predict that by reducing the loss and increasing the group index to realistic values, a conversion efficiency approaching 10 dB can be achieved.
Based on these results, we have designed slow light PhC waveguides operating in a low loss and constant dispersion window of ?? = 5 nm around ? = 1565 nm with a group index of ng = 60. We experimentally demonstrated a relatively low propagation loss, of 130 dB/cm, for waveguides up to 800 microm in length. This result is particularly remarkable given that the waveguides were written on an electron-beam lithography tool with a writefield of 100 microm that exhibits stitching errors of typically 10-50 nm. We reduced the impact of these stitching errors by introducing "slow-fast-slow" mode conversion interfaces and showed that these interfaces reduce the loss from 320 dB/cm to 130 dB/cm at ng = 60. This significant improvement highlights the importance of the slow-fast-slow method and shows that high performance slow light waveguides can be realised with lengths much longer than the writing field of a given e-beam lithography tool.
Finally, we showed efficient FWM wavelength conversion in these slow light waveguides with a group index of ng = 60. Using only 15 mW cw coupled input power, we observed a conversion efficiency of-28 dB. This efficiency represents a 30 dB enhancement compared to a silicon nanowire of the same length. The efficiency did not improve for higher input power, however, as the strong enhancement of the energy density in the slow light regime causes heating, which leads to a sufficient thermal red shift to detune the waveguide. This thermal limitation can be overcome by using oxide-clad waveguides, which we demonstrated for group indices of ng = 30. Higher group indices are difficult (but possible) to achieve with oxide clad-waveguides, and we predicted conversion efficiencies approaching-10 dB, which is equivalent to that already achieved in silicon nanowires but for a 50x shorter length.
In addition, our collaborators at CUDOS, Sydney University, Australia, have used our low loss slow light waveguides in the context of ultrahigh bandwidth light (up to 160 Gb/s) for a number of novel applications of the FWM effect. These include:
(a) Ultracompact 160 Gbaud all-optical demultiplexing. We demonstrated all-optical demultiplexing of a high-bandwidth, time-division multiplexed 160 Gbit/s signal to 10 Gbit/s channels, exploiting slow light enhanced FWM in a dispersion engineered, 96 µm long planar photonic crystal waveguide. We reported error-free (bit error rate < 10-9) operation of all 16 demultiplexed channels, with a power penalty of 2: 22: 4 dB, highlighting the potential of these structures as a platform for ultracompact all-optical nonlinear processes.
(b) Ultracompact all-optical XOR logic gate. We demonstrated an ultracompact, chip-based, all-optical exclusive-OR (XOR) logic gate via slow-light enhanced FWM in a silicon PhC waveguide. We achieved error-free operation (< 10-9) for 40 Gbit/s differential phase shift keying (DPSK) signals with a 2. 8 dB power penalty. The XOR operation required = 41 mW, corresponding to a switching energy of 1 pJ/bit. We compared the slow-light PhC waveguide device performance with experimentally demonstrated XOR DPSK logic gates in other platforms and discussed scaling the device operation to higher bit-rates. The ultracompact structure suggests the potential for device integration.
(c) Slow-light enhanced correlated photon pair generation. We reported the generation of correlated photon pairs in the telecom C-band at room temperature from a dispersionengineered silicon photonic crystal waveguide. The spontaneous FWM process producing the photon pairs is enhanced by slow-light propagation enabling an active device length of less than 100 µm. With a coincidence to accidental ratio of 12. 8 at a pair generation rate of 0. 006 per pulse, this ultracompact photon pair source paves the way toward scalable quantum information processing realised on-chip.
In conclusion, the fellowship has given me an outstanding opportunity to experience world-leading research in one of the best Photonics groups in Europe, with the additional bonus of interacting with the prime Photonics consortium in Australia. I am grateful to the Marie Curie programme for having provided me with this opportunity and am sure that the established linkeages will continue after my return to China.