5176
OPTICS LETTERS / Vol. 38, No. 23 / December 1, 2013
Bidirectional multiplexing of heralded single photons from a silicon chip C. Xiong,1,* T. D. Vo,1,2 M. J. Collins,1 J. Li,3,4 T. F. Krauss,3,5 M. J. Steel,6 A. S. Clark,1 and B. J. Eggleton1 1
Centre for Ultrahigh bandwidth Devices for Optical Systems (CUDOS), Institute of Photonics and Optical Science (IPOS), School of Physics, University of Sydney, NSW 2006, Australia 2
3 4
Maritime Division, Defence Science and Technology Organisation (DSTO), Department of Defence, P.O. Box 44, Pyrmont, NSW 2009, Australia
SUPA, School of Physics and Astronomy, University of St. Andrews, St. Andrews, Fife KY16 9SS, UK
Currently at State Key Laboratory of Optoelectronic Materials & Technologies, Sun Yat-sen University, Guangzhou 510275, China 5 Department of Physics, University of York, York YO10 5DD, UK 6
CUDOS, MQ Photonics Research Centre, Department of Physics and Astronomy, Macquarie University, NSW 2109, Australia *Corresponding author: [email protected] Received September 16, 2013; revised November 1, 2013; accepted November 1, 2013; posted November 4, 2013 (Doc. ID 197792); published November 27, 2013 We demonstrate integrated spatial multiplexing of heralded single photons generated from a single 96 μm long silicon photonic crystal waveguide in a bidirectional pump configuration. By using a low-loss fiber-coupled opto-ceramic switch, the multiplexing technique enhances the brightness of the single photon source by 51.2 4.0% while maintaining the coincidence-to-accidental ratio. Compared with the demonstration of multiplexing two individual sources, the bidirectional pump scheme represents a twofold reduction in the footprint of nonlinear devices for future large-scale integration of on-chip single photon sources. The 51.2 4.0% gain will make any quantum operation requiring n photons 1.5n times faster. © 2013 Optical Society of America OCIS codes: (130.5296) Photonic crystal waveguides; (270.5585) Quantum information and processing. http://dx.doi.org/10.1364/OL.38.005176
Single photon sources on a photonic chip are a key building block of integrated quantum photonic technologies. One of the popular strategies to create single photons from a photonic chip is to use spontaneous four-wave mixing (SFWM) [1–5] or spontaneous parametric downconversion (SPDC) [6,7] in nonlinear photonic waveguides to generate correlated photon pairs. The detection of one photon from a pair heralds the existence of the other, forming a heralded single photon source. The major challenge of such photon pair sources is the inability to produce heralded single photons on demand because single pairs cannot be produced with high probability, while simultaneously suppressing the probability of yielding two or more pairs. To decouple these probabilities, a scheme was proposed to actively multiplex many such sources to enhance the probability of single photon output while maintaining a constant multiphoton noise level [8,9]. This was experimentally implemented using free-space SPDC sources and optical components [10], which are intrinsically unstable and not scalable. In this Letter, we demonstrate the multiplexing of heralded single photon sources from a single ultra-compact silicon photonic crystal waveguide (PhCW) pumped from both directions. We achieve 51.2 4.0% enhancement to the source brightness without reducing the coincidenceto-accidental ratio (CAR). Recently we demonstrated the spatial multiplexing of two individual waveguide sources [11]; however, this was challenging to implement due to fabrication constraints in matching the properties of the nonlinear devices. Here, as the multiplexed heralded photons are from the same waveguide, they are automatically indistinguishable. The brightness achieved here is an order of magnitude greater than our previous demonstration with higher CAR values throughout. For future large-scale integration, this scheme potentially 0146-9592/13/235176-04$15.00/0
represents a twofold reduction in the footprint of the nonlinear device. Our 51.2 4.0% gain will make any quantum operation requiring n photons 1.5n times faster, giving an exponential speed-up to quantum information experiments and providing a clear route to scalability. The bidirectional multiplexing of single photons from a single PhCW is schematically illustrated in Fig. 1(a). The 96 μm long silicon slow-light PhCW used in this study was fabricated on a silicon-on-insulator wafer comprising a 220 nm silicon layer on 2 μm of silica using electron beam lithography and reactive ion etching [12,13]. The PhCW was created from a triangular lattice of air holes etched in a suspended silicon membrane with a row of holes missing along the ΓK direction. The two rows adjacent to the waveguide center were laterally shifted to engineer the waveguide dispersion such that it exhibits a group index of ∼30 with low dispersion (β2 ≈ 1 × 10−21 s2 ∕m) and total loss of 10 dB across a 15 nm window [Fig. 1(b)] for efficient SFWM [4,5]. The PhCW effective nonlinear coefficient γ eff ng ∕n0 γ was approximately 4000 W−1 m−1 , where the slow-down factor, the ratio of the group index ng to the native refractive index of the material n0 , has been included. The PhCW region had a linear propagation loss of 0.6 dB. Silicon access waveguides, including inverse tapers terminated by wide polymer waveguides, were added to the input and output of the PhCW region to improve coupling efficiency to optical fibers. The waveguide was pumped from the left (L) and right (R) simultaneously by 10 ps pulses at 1555.7 nm from a mode-locked fiber laser (Pritel) with a repetition rate of 50 MHz. Correlated photon pairs at 1550.9 and 1560.6 nm were generated through slow-light enhanced SFWM [4,5] with equalized probability (up to 0.05 pairs per pulse) for pumping from both sides. Photons © 2013 Optical Society of America
December 1, 2013 / Vol. 38, No. 23 / OPTICS LETTERS
Fig. 1. (a) Schematic of the multiplexing setup. See main text for description. (b) Group index (black) and transmission (gray) of the PhCW. Green, blue, and red shaded regions indicate the pump (p), signal (s), and idler (i) wavelengths, respectively.
emerging from either end were collected by a circulator (CIR), separated by an arrayed waveguide grating (AWG), further filtered by 1 nm bandpass filters (BPFs), and conditioned by a polarization controller (PC) before being sent to superconducting single photon detectors (SSPDs, Single Quantum, ∼10% detection efficiency with 100 Hz dark count, polarization sensitive). BPF and PC are not shown in the diagram of Fig. 1(a). The AWG had 40 channels with 100 GHz channel spacing and 50 GHz channel bandwidth. The total insertion loss of the circulator, AWG, BPF, and PC was approximately 8 dB. After taking into account all losses and the detector efficiency, the total photon collection efficiency of the signal/idler channel was 0.5%. The signal photon (higher frequency, shown in blue), when detected by a SSPD, heralded the existence of its partner (idler photon, lower frequency, shown in red) triggering an RF logic gate and driving a fast switch to form a multiplexed stream of indistinguishable single photons at the common output of the switch. To match the electronic delay imposed on the heralding signals by the detectors and switch circuits, a 200 m long single-mode
5177
fiber was used to optically delay the idler photons from each source for approximately 1 μs before they were sent to the switch. Each individual heralded single photon source, pumped from either the left or the right, has been proven to be able to operate in the single photon regime by measuring the second-order correlation function g2 0 [11]. In the less probable case that both sources generate a photon simultaneously (up to 0.0025 per pulse in our experiment), one is discarded and the other is routed to the output. To achieve a net gain for the heralded single photon rate, the switch insertion loss must be below the 3 dB threshold. In our experiment, a fiber-coupled opto-ceramic electro-optic switch, made from ultra-low-loss lead lanthanum zirconium titanate (PLZT) [14], provided a switching speed of 1 MHz. The transmission of the two switch channels was measured to be 71.0% and 85.5%, respectively. Compared with a single photon source without switching, this gives a maximum potential multiplexing gain of 56.5% if the two sources to be multiplexed are matched in CAR and brightness. To demonstrate the advantage of bidirectional multiplexing, as a reference we first measured the CAR and heralded single photon rate with the switch removed from the setup for single (either left or right) pumping as a function of pump power. To clearly show the enhancement of multiplexing later, in this reference measurement we have matched the two sources pumped from either side. The results are plotted in Fig. 2(a), and they agree with the fit using the model in [2]. As expected, the CAR decreased when the heralded single photon rate increased due to multipair generation [1–5]. The fit for the heralded single photon rate shows a quadratic dependency on power, which is the signature of SFWM. We did not observe a saturation effect in the power range of operation, which means we were operating in the regime below the two-photon absorption threshold [4,5]. We then added the switch to enable the multiplexing, and measured the CAR and heralded single photon rate at the switch common output for the same pump conditions as the reference measurement. The measurement integration time for all data points was between 2 and 5 min. For comparison, we plot the CAR as a function of heralded single photon rate in Fig. 2(b) for left and right pumping without the switch, and bidirectional pumping after multiplexing. It can be seen from Fig. 2(b) that after multiplexing, at the same CAR level, the heralded single photon rate, or the source brightness, was enhanced. On the other hand, at the same coincidence rate, the CAR was improved by ∼50%. This broke the limit on maximum CAR set by the average number of pairs per pulse for each individual pair source [2]. To quantify the enhancement of multiplexing, we define the enhancement factor as C M ∕C L − 1 or C M ∕C R − 1, where C L and C R denote the single photon rate for the pump from the left and right without switch, and C M denotes the single photon rate after multiplexing at the same pump conditions to the single pumping sources. If we use the model developed in [11] to fit the plot in Fig. 2(b), we can numerically extract a set of C M , C L , and C R (C L is equal to C R for the two matched sources)
5178
OPTICS LETTERS / Vol. 38, No. 23 / December 1, 2013
coupling losses from the optical chip and any component losses will also increase the useful photons heralded at the output of the device, but we point out that this multiplexing scheme can be used for any currently available heralded photon source and will enhance the photon rate for a given signal-to-noise level. If the photon collection efficiency can be increased close to 100%, then the count rate of the heralding photons could be more than 1 MHz, and one would be required to either match the repetition rate of the laser to the bandwidth of the switch (currently 1 MHz) or increase the switch speed. It should be noted that such a reduction would still enable a 1 MHz deterministic photon source, which is many orders of magnitude brighter than current devices and would enable massive parallel coincidence measurements to be taken. In conclusion, we have demonstrated a 51.2 4.0% enhancement to the brightness of an ultra-compact silicon chip-based single photon source without increasing the noise, enabled through bidirectional multiplexing with integrated and fiber-coupled devices. The decoupling of the heralded single photon rate from the noise allows the useful single photon rate to be enhanced. This spatial multiplexing can be scaled to include many nonlinear heralded single photon sources to create a single photon source with deterministic operation, creating scalable and efficient quantum photonic technologies.
Fig. 2. (a) CAR and heralded single photon rate measured at different pump power without switch. (b) Measured CAR as a function of detected heralded single photon rate before and after multiplexing. Black squares are for the left pump, red circles for the right pump, and blue triangles for the multiplexed output. The solid and dashed lines are theoretical fits. Poissonian error bars are used in the plot.
at each CAR level, which yields a constant enhancement factor of 51.2% 4.0%. It should be noted that this bidirectional multiplexing shows an order of magnitude higher brightness and more than twice higher CAR throughout than multiplexing two individual sources [11]. This is due to the fact that the two sources in the bidirectional pump scheme are easily matched and both can operate at optimum conditions, whereas to match two individual sources turns out to be challenging and it is hard to make each individual source work at optimum status. Although it will be unavoidable to match different sources in future experiments and applications involving more than two sources, the bidirectional multiplexing still represents a twofold reduction in the footprint of the photonic circuits. In experiments involving quantum interference of indistinguishable photons, the coincidence rate is proportional to the square of the single photon rate and the visibility depends on CAR. If we use two multiplexed single photon sources demonstrated here in future quantum experiments, the measurement of interference will be approximately 1.52 times faster while maintaining the visibility. In a quantum system requiring n photons, the speed of operation will be 1.5n times faster, which represents an exponential speed-up. Improvements in
This work was supported by the Centre of Excellence (CUDOS, project no. CE110001018), the Laureate Fellowship (L120100029) and Discovery Early Career Researcher Award programs (DE130101148 and DE120100226) of the Australian Research Council (ARC). The silicon waveguide chip was fabricated under the Engineering and Physical Sciences Research Council, U.K. Silicon Photonics Consortium, and was supported by the European Union Seventh Framework Programme Marie Curie project “OSIRIS.” References 1. J. E. Sharping, K. F. Lee, M. A. Foster, A. C. Turner, B. S. Schmidt, M. Lipson, A. L. Gaeta, and P. Kumar, Opt. Express 14, 12388 (2006). 2. K. Harada, H. Takesue, H. Fukuda, T. Tsuchizawa, T. Watanabe, K. Yamada, Y. Tokura, and S. Itabashi, IEEE J. Sel. Top. Quantum Electron. 16, 325 (2010). 3. S. Clemmen, A. Perret, S. K. Selvaraja, W. Bogaerts, D. van Thourhout, R. Baets, P. Emplit, and S. Massar, Opt. Lett. 35, 3483 (2010). 4. C. Xiong, C. Monat, A. S. Clark, C. Grillet, G. D. Marshall, M. J. Steel, J. Li, L. O’Faolain, T. F. Krauss, J. G. Rarity, and B. J. Eggleton, Opt. Lett. 36, 3413 (2011). 5. C. Xiong, C. Monat, M. J. Collins, L. Tranchant, D. Petiteau, A. S. Clark, C. Grillet, G. D. Marshall, M. J. Steel, J. Li, L. O’Faolain, T. F. Krauss, and B. J. Eggleton, IEEE J. Sel. Top. Quantum Electron. 18, 1676 (2012). 6. M. Hunault, H. Takesue, O. Tadanaga, Y. Nishida, and M. Asobe, Opt. Lett. 35, 1239 (2010). 7. M. Lobino, G. D. Marshall, C. Xiong, A. S. Clark, D. Bonneau, C. M. Natarajan, M. G. Tanner, R. H. Hadfield, S. N. Dorenbos, T. Zijlstra, V. Zwiller, M. Marangoni, R. Ramponi, M. G. Thompson, B. J. Eggleton, and J. L. O’Brien, Appl. Phys. Lett. 99, 081110 (2011). 8. A. L. Migdall, D. Branning, and S. Castelletto, Phys. Rev. A 66, 053805 (2002).
December 1, 2013 / Vol. 38, No. 23 / OPTICS LETTERS 9. J. H. Shapiro and F. N. C. Wong, Opt. Lett. 32, 2698 (2007). 10. X. Ma, S. Zotter, J. Kofler, T. Jennewein, and A. Zeilinger, Phys. Rev. A 83, 043814 (2011). 11. M. J. Collins, C. Xiong, I. H. Rey, T. D. Vo, J. He, S. Shahnia, C. Reardon, M. J. Steel, T. F. Krauss, A. S. Clark, and B. J. Eggleton, Nat. Commun. 4, 2582 (2013).
5179
12. J. Li, T. P. White, L. O’Faolain, A. Gomez-Iglesias, and T. F. Krauss, Opt. Express 16, 6227 (2008). 13. J. Li, L. O’Faolain, I. H. Rey, and T. F. Krauss, Opt. Express 19, 4458 (2011). 14. H. Jiang, Y. K. Zou, Q. Chen, K. K. Li, R. Zhang, Y. Wang, H. Ming, and Z. Zheng, Proc. SPIE 5644, 380 (2005).
OPTICS LETTERS / Vol. 38, No. 23 / December 1, 2013
Bidirectional multiplexing of heralded single photons from a silicon chip C. Xiong,1,* T. D. Vo,1,2 M. J. Collins,1 J. Li,3,4 T. F. Krauss,3,5 M. J. Steel,6 A. S. Clark,1 and B. J. Eggleton1 1
Centre for Ultrahigh bandwidth Devices for Optical Systems (CUDOS), Institute of Photonics and Optical Science (IPOS), School of Physics, University of Sydney, NSW 2006, Australia 2
3 4
Maritime Division, Defence Science and Technology Organisation (DSTO), Department of Defence, P.O. Box 44, Pyrmont, NSW 2009, Australia
SUPA, School of Physics and Astronomy, University of St. Andrews, St. Andrews, Fife KY16 9SS, UK
Currently at State Key Laboratory of Optoelectronic Materials & Technologies, Sun Yat-sen University, Guangzhou 510275, China 5 Department of Physics, University of York, York YO10 5DD, UK 6
CUDOS, MQ Photonics Research Centre, Department of Physics and Astronomy, Macquarie University, NSW 2109, Australia *Corresponding author: [email protected] Received September 16, 2013; revised November 1, 2013; accepted November 1, 2013; posted November 4, 2013 (Doc. ID 197792); published November 27, 2013 We demonstrate integrated spatial multiplexing of heralded single photons generated from a single 96 μm long silicon photonic crystal waveguide in a bidirectional pump configuration. By using a low-loss fiber-coupled opto-ceramic switch, the multiplexing technique enhances the brightness of the single photon source by 51.2 4.0% while maintaining the coincidence-to-accidental ratio. Compared with the demonstration of multiplexing two individual sources, the bidirectional pump scheme represents a twofold reduction in the footprint of nonlinear devices for future large-scale integration of on-chip single photon sources. The 51.2 4.0% gain will make any quantum operation requiring n photons 1.5n times faster. © 2013 Optical Society of America OCIS codes: (130.5296) Photonic crystal waveguides; (270.5585) Quantum information and processing. http://dx.doi.org/10.1364/OL.38.005176
Single photon sources on a photonic chip are a key building block of integrated quantum photonic technologies. One of the popular strategies to create single photons from a photonic chip is to use spontaneous four-wave mixing (SFWM) [1–5] or spontaneous parametric downconversion (SPDC) [6,7] in nonlinear photonic waveguides to generate correlated photon pairs. The detection of one photon from a pair heralds the existence of the other, forming a heralded single photon source. The major challenge of such photon pair sources is the inability to produce heralded single photons on demand because single pairs cannot be produced with high probability, while simultaneously suppressing the probability of yielding two or more pairs. To decouple these probabilities, a scheme was proposed to actively multiplex many such sources to enhance the probability of single photon output while maintaining a constant multiphoton noise level [8,9]. This was experimentally implemented using free-space SPDC sources and optical components [10], which are intrinsically unstable and not scalable. In this Letter, we demonstrate the multiplexing of heralded single photon sources from a single ultra-compact silicon photonic crystal waveguide (PhCW) pumped from both directions. We achieve 51.2 4.0% enhancement to the source brightness without reducing the coincidenceto-accidental ratio (CAR). Recently we demonstrated the spatial multiplexing of two individual waveguide sources [11]; however, this was challenging to implement due to fabrication constraints in matching the properties of the nonlinear devices. Here, as the multiplexed heralded photons are from the same waveguide, they are automatically indistinguishable. The brightness achieved here is an order of magnitude greater than our previous demonstration with higher CAR values throughout. For future large-scale integration, this scheme potentially 0146-9592/13/235176-04$15.00/0
represents a twofold reduction in the footprint of the nonlinear device. Our 51.2 4.0% gain will make any quantum operation requiring n photons 1.5n times faster, giving an exponential speed-up to quantum information experiments and providing a clear route to scalability. The bidirectional multiplexing of single photons from a single PhCW is schematically illustrated in Fig. 1(a). The 96 μm long silicon slow-light PhCW used in this study was fabricated on a silicon-on-insulator wafer comprising a 220 nm silicon layer on 2 μm of silica using electron beam lithography and reactive ion etching [12,13]. The PhCW was created from a triangular lattice of air holes etched in a suspended silicon membrane with a row of holes missing along the ΓK direction. The two rows adjacent to the waveguide center were laterally shifted to engineer the waveguide dispersion such that it exhibits a group index of ∼30 with low dispersion (β2 ≈ 1 × 10−21 s2 ∕m) and total loss of 10 dB across a 15 nm window [Fig. 1(b)] for efficient SFWM [4,5]. The PhCW effective nonlinear coefficient γ eff ng ∕n0 γ was approximately 4000 W−1 m−1 , where the slow-down factor, the ratio of the group index ng to the native refractive index of the material n0 , has been included. The PhCW region had a linear propagation loss of 0.6 dB. Silicon access waveguides, including inverse tapers terminated by wide polymer waveguides, were added to the input and output of the PhCW region to improve coupling efficiency to optical fibers. The waveguide was pumped from the left (L) and right (R) simultaneously by 10 ps pulses at 1555.7 nm from a mode-locked fiber laser (Pritel) with a repetition rate of 50 MHz. Correlated photon pairs at 1550.9 and 1560.6 nm were generated through slow-light enhanced SFWM [4,5] with equalized probability (up to 0.05 pairs per pulse) for pumping from both sides. Photons © 2013 Optical Society of America
December 1, 2013 / Vol. 38, No. 23 / OPTICS LETTERS
Fig. 1. (a) Schematic of the multiplexing setup. See main text for description. (b) Group index (black) and transmission (gray) of the PhCW. Green, blue, and red shaded regions indicate the pump (p), signal (s), and idler (i) wavelengths, respectively.
emerging from either end were collected by a circulator (CIR), separated by an arrayed waveguide grating (AWG), further filtered by 1 nm bandpass filters (BPFs), and conditioned by a polarization controller (PC) before being sent to superconducting single photon detectors (SSPDs, Single Quantum, ∼10% detection efficiency with 100 Hz dark count, polarization sensitive). BPF and PC are not shown in the diagram of Fig. 1(a). The AWG had 40 channels with 100 GHz channel spacing and 50 GHz channel bandwidth. The total insertion loss of the circulator, AWG, BPF, and PC was approximately 8 dB. After taking into account all losses and the detector efficiency, the total photon collection efficiency of the signal/idler channel was 0.5%. The signal photon (higher frequency, shown in blue), when detected by a SSPD, heralded the existence of its partner (idler photon, lower frequency, shown in red) triggering an RF logic gate and driving a fast switch to form a multiplexed stream of indistinguishable single photons at the common output of the switch. To match the electronic delay imposed on the heralding signals by the detectors and switch circuits, a 200 m long single-mode
5177
fiber was used to optically delay the idler photons from each source for approximately 1 μs before they were sent to the switch. Each individual heralded single photon source, pumped from either the left or the right, has been proven to be able to operate in the single photon regime by measuring the second-order correlation function g2 0 [11]. In the less probable case that both sources generate a photon simultaneously (up to 0.0025 per pulse in our experiment), one is discarded and the other is routed to the output. To achieve a net gain for the heralded single photon rate, the switch insertion loss must be below the 3 dB threshold. In our experiment, a fiber-coupled opto-ceramic electro-optic switch, made from ultra-low-loss lead lanthanum zirconium titanate (PLZT) [14], provided a switching speed of 1 MHz. The transmission of the two switch channels was measured to be 71.0% and 85.5%, respectively. Compared with a single photon source without switching, this gives a maximum potential multiplexing gain of 56.5% if the two sources to be multiplexed are matched in CAR and brightness. To demonstrate the advantage of bidirectional multiplexing, as a reference we first measured the CAR and heralded single photon rate with the switch removed from the setup for single (either left or right) pumping as a function of pump power. To clearly show the enhancement of multiplexing later, in this reference measurement we have matched the two sources pumped from either side. The results are plotted in Fig. 2(a), and they agree with the fit using the model in [2]. As expected, the CAR decreased when the heralded single photon rate increased due to multipair generation [1–5]. The fit for the heralded single photon rate shows a quadratic dependency on power, which is the signature of SFWM. We did not observe a saturation effect in the power range of operation, which means we were operating in the regime below the two-photon absorption threshold [4,5]. We then added the switch to enable the multiplexing, and measured the CAR and heralded single photon rate at the switch common output for the same pump conditions as the reference measurement. The measurement integration time for all data points was between 2 and 5 min. For comparison, we plot the CAR as a function of heralded single photon rate in Fig. 2(b) for left and right pumping without the switch, and bidirectional pumping after multiplexing. It can be seen from Fig. 2(b) that after multiplexing, at the same CAR level, the heralded single photon rate, or the source brightness, was enhanced. On the other hand, at the same coincidence rate, the CAR was improved by ∼50%. This broke the limit on maximum CAR set by the average number of pairs per pulse for each individual pair source [2]. To quantify the enhancement of multiplexing, we define the enhancement factor as C M ∕C L − 1 or C M ∕C R − 1, where C L and C R denote the single photon rate for the pump from the left and right without switch, and C M denotes the single photon rate after multiplexing at the same pump conditions to the single pumping sources. If we use the model developed in [11] to fit the plot in Fig. 2(b), we can numerically extract a set of C M , C L , and C R (C L is equal to C R for the two matched sources)
5178
OPTICS LETTERS / Vol. 38, No. 23 / December 1, 2013
coupling losses from the optical chip and any component losses will also increase the useful photons heralded at the output of the device, but we point out that this multiplexing scheme can be used for any currently available heralded photon source and will enhance the photon rate for a given signal-to-noise level. If the photon collection efficiency can be increased close to 100%, then the count rate of the heralding photons could be more than 1 MHz, and one would be required to either match the repetition rate of the laser to the bandwidth of the switch (currently 1 MHz) or increase the switch speed. It should be noted that such a reduction would still enable a 1 MHz deterministic photon source, which is many orders of magnitude brighter than current devices and would enable massive parallel coincidence measurements to be taken. In conclusion, we have demonstrated a 51.2 4.0% enhancement to the brightness of an ultra-compact silicon chip-based single photon source without increasing the noise, enabled through bidirectional multiplexing with integrated and fiber-coupled devices. The decoupling of the heralded single photon rate from the noise allows the useful single photon rate to be enhanced. This spatial multiplexing can be scaled to include many nonlinear heralded single photon sources to create a single photon source with deterministic operation, creating scalable and efficient quantum photonic technologies.
Fig. 2. (a) CAR and heralded single photon rate measured at different pump power without switch. (b) Measured CAR as a function of detected heralded single photon rate before and after multiplexing. Black squares are for the left pump, red circles for the right pump, and blue triangles for the multiplexed output. The solid and dashed lines are theoretical fits. Poissonian error bars are used in the plot.
at each CAR level, which yields a constant enhancement factor of 51.2% 4.0%. It should be noted that this bidirectional multiplexing shows an order of magnitude higher brightness and more than twice higher CAR throughout than multiplexing two individual sources [11]. This is due to the fact that the two sources in the bidirectional pump scheme are easily matched and both can operate at optimum conditions, whereas to match two individual sources turns out to be challenging and it is hard to make each individual source work at optimum status. Although it will be unavoidable to match different sources in future experiments and applications involving more than two sources, the bidirectional multiplexing still represents a twofold reduction in the footprint of the photonic circuits. In experiments involving quantum interference of indistinguishable photons, the coincidence rate is proportional to the square of the single photon rate and the visibility depends on CAR. If we use two multiplexed single photon sources demonstrated here in future quantum experiments, the measurement of interference will be approximately 1.52 times faster while maintaining the visibility. In a quantum system requiring n photons, the speed of operation will be 1.5n times faster, which represents an exponential speed-up. Improvements in
This work was supported by the Centre of Excellence (CUDOS, project no. CE110001018), the Laureate Fellowship (L120100029) and Discovery Early Career Researcher Award programs (DE130101148 and DE120100226) of the Australian Research Council (ARC). The silicon waveguide chip was fabricated under the Engineering and Physical Sciences Research Council, U.K. Silicon Photonics Consortium, and was supported by the European Union Seventh Framework Programme Marie Curie project “OSIRIS.” References 1. J. E. Sharping, K. F. Lee, M. A. Foster, A. C. Turner, B. S. Schmidt, M. Lipson, A. L. Gaeta, and P. Kumar, Opt. Express 14, 12388 (2006). 2. K. Harada, H. Takesue, H. Fukuda, T. Tsuchizawa, T. Watanabe, K. Yamada, Y. Tokura, and S. Itabashi, IEEE J. Sel. Top. Quantum Electron. 16, 325 (2010). 3. S. Clemmen, A. Perret, S. K. Selvaraja, W. Bogaerts, D. van Thourhout, R. Baets, P. Emplit, and S. Massar, Opt. Lett. 35, 3483 (2010). 4. C. Xiong, C. Monat, A. S. Clark, C. Grillet, G. D. Marshall, M. J. Steel, J. Li, L. O’Faolain, T. F. Krauss, J. G. Rarity, and B. J. Eggleton, Opt. Lett. 36, 3413 (2011). 5. C. Xiong, C. Monat, M. J. Collins, L. Tranchant, D. Petiteau, A. S. Clark, C. Grillet, G. D. Marshall, M. J. Steel, J. Li, L. O’Faolain, T. F. Krauss, and B. J. Eggleton, IEEE J. Sel. Top. Quantum Electron. 18, 1676 (2012). 6. M. Hunault, H. Takesue, O. Tadanaga, Y. Nishida, and M. Asobe, Opt. Lett. 35, 1239 (2010). 7. M. Lobino, G. D. Marshall, C. Xiong, A. S. Clark, D. Bonneau, C. M. Natarajan, M. G. Tanner, R. H. Hadfield, S. N. Dorenbos, T. Zijlstra, V. Zwiller, M. Marangoni, R. Ramponi, M. G. Thompson, B. J. Eggleton, and J. L. O’Brien, Appl. Phys. Lett. 99, 081110 (2011). 8. A. L. Migdall, D. Branning, and S. Castelletto, Phys. Rev. A 66, 053805 (2002).
December 1, 2013 / Vol. 38, No. 23 / OPTICS LETTERS 9. J. H. Shapiro and F. N. C. Wong, Opt. Lett. 32, 2698 (2007). 10. X. Ma, S. Zotter, J. Kofler, T. Jennewein, and A. Zeilinger, Phys. Rev. A 83, 043814 (2011). 11. M. J. Collins, C. Xiong, I. H. Rey, T. D. Vo, J. He, S. Shahnia, C. Reardon, M. J. Steel, T. F. Krauss, A. S. Clark, and B. J. Eggleton, Nat. Commun. 4, 2582 (2013).
5179
12. J. Li, T. P. White, L. O’Faolain, A. Gomez-Iglesias, and T. F. Krauss, Opt. Express 16, 6227 (2008). 13. J. Li, L. O’Faolain, I. H. Rey, and T. F. Krauss, Opt. Express 19, 4458 (2011). 14. H. Jiang, Y. K. Zou, Q. Chen, K. K. Li, R. Zhang, Y. Wang, H. Ming, and Z. Zheng, Proc. SPIE 5644, 380 (2005).
