August 15, 2014 / Vol. 39, No. 16 / OPTICS LETTERS
4667
Improved configuration and reduction of phase noise in a narrow linewidth ultrawideband optical RF source David W. Grund, Jr.,* Shouyuan Shi, Garrett J. Schneider, Janusz Murakowski, and Dennis W. Prather Electrical and Computer Engineering, University of Delaware, Newark, Delaware 19716, USA *Corresponding author: [email protected] Received May 20, 2014; revised June 18, 2014; accepted June 19, 2014; posted June 25, 2014 (Doc. ID 212506); published August 5, 2014 In this Letter, we report on the improved configuration of a widely tunable optical RF generation system, particularly for the generation of low-frequency RF, as well as the reduction of phase noise in that same system. Using an amplitude modulator, a simplified system design was demonstrated with fewer components and improved phase noise performance, especially at RF frequencies below ∼36 GHz. Excess phase noise due to acoustic vibrations of the optical fibers was also successfully eliminated by mechanical isolation. A minimum phase noise of −124 dBc∕Hz at 10 kHz offset was demonstrated at 4 GHz. © 2014 Optical Society of America OCIS codes: (060.5625) Radio frequency photonics; (140.3520) Lasers, injection-locked; (060.2840) Heterodyne; (060.4080) Modulation. http://dx.doi.org/10.1364/OL.39.004667
Synthesis of radio frequency signals with a high spectral purity and ultrawide tuning range is of great interest for the realization of radio-over-fiber systems and instrumentation [1]. Optical RF generation inherently offers extreme bandwidth because small shifts in wavelength correspond to very large changes in frequency. For example, 200 GHz of RF bandwidth can be covered with a change in wavelength of just 1.6 nm when operating near 1550 nm. Optical RF is usually produced by heterodyning on a photodiode and can be implemented in a variety of ways. Systems have been demonstrated based on techniques such as direct mixing [2], direct mixing of modulation sidebands, multiwavelength lasers [3–6], optical phase locked loops (OPLL) [7–9], and various types of injection locking schemes [10–18]. Here we focus on the improved design of a widely tunable and spectrally pure optical RF generation system [19], particularly for the production of RF frequencies below ∼36 GHz, as well as the reduction of phase noise in that same system. This optical RF generation system is based on a modulation-sideband injection-locking scheme. Modulation-sideband injection-locking is advantageous because it offers continuous tunability and harmonic suppression without sacrificing power, while avoiding complex feedback mechanisms and the need for ultranarrow filters [20]. Phase noise has several potential sources in the optical RF generation system depicted in Fig. 1. In addition to the minimum phase noise of the scaled local RF oscillator, they include things such as spontaneous emission in the lasers, thermal and vibration effects on the optical fibers, path length mismatch, uncorrelated residual optical noise due to insufficient injection power, laser driving current fluctuations, and shot noise in the photodiode. The dominant two sources have been experimentally and theoretically determined to be due to the uncorrelated residual optical noise due to insufficient injection power and vibration effects in the optical fiber [19,21]. In the previous system [19], a phase modulator and two DWDM filters in series, as seen in the top callout of Fig. 1, were used to produce the sidebands for injection locking 0146-9592/14/164667-04$15.00/0
and reduce the strength of the carrier, which is typically stronger than the first order sidebands by 10 dB or more depending on the modulation depth. The filters had a total rolloff of approximately 1.33 dB∕GHz and an insertion loss of approximately 4 dB. These losses were the cause
Fig. 1. Layout and injection spectra of a radio frequency signal generation system based on injection-locked lasers. Modulation and carrier suppression were performed by a phase modulator and filters in Ref. [19] and by an amplitude modulator in this work. © 2014 Optical Society of America
4668
OPTICS LETTERS / Vol. 39, No. 16 / August 15, 2014
of insufficient injection power at frequencies below ∼36 GHz. It was also found that even at frequencies where there was adequate injection power, there was a large excess of phase noise above the scaled reference, for offset frequencies below 20 kHz, due to acoustic noise coupling into the loose optical fibers. In the bottom callout of Fig. 1 we can see the system layout under test in this work. We have replaced the phase modulator and filters with a single amplitude modulator. The amplitude modulator can be biased to the null point, performing carrier suppression, removing the need for the optical filter. This functionality makes the amplitude modulator configuration superior in several ways. First, we regain all the sideband power in the lower order harmonics that is lost in the filter from operating the master laser far enough down the filter edge to provide adequate suppression. Second, it eliminates the insertion loss associated with using the filter, providing a boost to the maximum injectable optical power. These two advantages are important because the noise suppression is governed by the injection power. Another advantage of this configuration is the ability to suppress the carrier at any optical frequency within the functional domain of the modulator, removing the limitation of working at the filter rejection band edge. This improves system flexibility. Finally, this configuration reduces the number of components and the amount of fiber through which mechanical vibrations can be coupled. The one possible drawback of using an amplitude modulator is that requiring a bias voltage could potentially lead to a long-term stability issue. Fortunately, amplitude modulators are extensively used in the telecommunications industry, and bias control feedback circuits are readily available to overcome this issue. In these experiments a single-arm-drive Z-cut lithium niobate modulator, seen in Fig. 2, was utilized. This modulator offers an extinction ratio greater than 20 dB for the carrier, which is sufficient to reduce it below the power of the first sidebands. The nulling capabilities could be improved by using a balanced-drive X-cut lithium niobate modulator with a dual-parallel Mach–Zehnder configuration, shown in Fig. 3. This allows for independent amplitude control in each arm in order to get optimal carrier extinction ratios at any working frequency. In addition, the balanced drive would allow for nulling the other even-order sidebands in addition to the carrier [22]. In the amplitude modulator system shown in Fig. 1, a tunable microwave signal generator (Agilent PSGE8267D) and a saturated broadband RF amplifier (Picosecond Pulse Labs 5882) that acts as a distortion element drive the amplitude modulator (EO Space AZ-AV1-40-PFU-PFU) that is biased simply by a voltage supply. The magnitude of the electric field of the harmonics produced by the saturated amplifier and passed
Fig. 2. Single-arm drive Z-cut lithium niobate Mach–Zender modulator.
Fig. 3. Dual-parallel balanced-drive X-cut lithium niobate Mach–Zender modulator.
through the modulator are given by J 1 x∕n, with a rolloff of 1∕nx∕2, while the harmonics produced by the nonlinearity of the modulator are given by J n x, with a rolloff of 1∕n!x∕2n , where x πV ∕V π . For the implemented system V ≤ 2.5 V, V π ∼ 5 V, and therefore x ≤ π∕2. This means that by the third harmonic, the modulator nonlinearity is contributing ≤1∕18 of the power and can be considered negligible. Therefore, the saturated amplifier is the primary source of harmonic power in the system. The harmonic sidebands are then amplified by a semiconductor optical amplifier (SOA) (Thorlabs S9FC1004P) and injected into the slave laser. An SOA is used instead of an EDFA because an EDFA has a minimum input power requirement, typically −3 dBm, which the sidebands produced in our system do not meet. The master and slave lasers (Mitsubishi FU-68PDF-5) are then mixed together on a photodiode (u2 t Photonics XPDV2150R) to produce the RF signal, which is amplified by a broadband RF amplifier (Picosecond Pulse Labs 5882) in order to prevent the phase noise measurement from being limited by the spectrum analyzer (Agilent PXA-N9030A) noise floor. In a frequency multiplication system such as this, the phase noise scales with the frequency multiplier N as 20log10 N [23]. It is noted that since the lasers in this system are not perfect filters, there is some residual content from the other injected harmonics. These are typically 25–40 dB weaker than the desired RF signal depending on the injected power. Such harmonic suppression is comparable to what is specified for typical commercial RF signal generators. In Fig. 4 we show that through mechanical isolation of the system, by mounting it on a stiff board, properly securing all fibers to the board with tape, and placing that board on foam, we were able to eliminate the fiber vibration noise, and closely match the scaled reference at all offset frequencies. Using the notation PNExcess PNSignal − PNScaled Reference ; Fig. 5 compares the difference between the optical RF and the scaled reference in this work and from the work of Ref. [19] and shows the dramatic improvement of up to 30 dB over the previously reported result at low offset frequencies. As discussed above, the minimum phase noise that can be achieved is limited to the phase noise of the scaled RF reference [21]. In order to see how well the noise had been suppressed, the optical RF generation from the 1st harmonic of 4 GHz was examined because it has a lower possible noise floor than the 36 GHz signal. In Fig. 6 we show the system’s best phase noise performance to date. This result has been made possible by the
August 15, 2014 / Vol. 39, No. 16 / OPTICS LETTERS
4669
Fig. 4. Comparison of the phase noise of the generated signal and reference at 36 GHz.
Fig. 6. Comparison of the phase noise of the generated signal and reference at 4 GHz.
elimination of approximately 30 dB of loss present in the phase modulator system due to the rolloff and insertion losses of the DWDM filters. It can be seen that there is a decent match at frequencies below 4 kHz at which point the noise levels out at −124 dBc∕Hz. Considering the shape of the curve, the difference between the reference and the optically generated signal is likely due to insufficient injection power to fully correlate the master and slave lasers. It is important to note that this minimum phase noise result is not just applicable at 4 GHz (N 1), but could be achieved at any frequency/multiplier, given a superior reference oscillator and enough injection power, because the minimum noise achievable is determined by the scaled reference oscillator, and the ability to conform the optically generated signal to the scaled reference oscillator is determined primarily by the injection strength [21]. In summary, we have demonstrated a system design using an amplitude modulator that offers improved system flexibility, operation with fewer components, and no sideband power loss, leading to improved phase noise performance at RF frequencies below ∼36 GHz. We also successfully eliminated the excess phase noise in our
system due to acoustic vibrations of the fibers by mechanically isolating the system. Finally we have demonstrated a new minimum phase noise of −124 dBc∕Hz at a 10 kHz offset from 4 GHz. In the future the system could be improved through the use of a dual-parallel Mach– Zehnder modulator and elimination of the fibers through complete system integration [20].
Fig. 5. Comparison of the difference between the phase noise of the generated signal and scaled reference at 36 GHz in this work and from the work of Ref. [19].
The authors would like to acknowledge the support of R. Nelson from the U.S. Air Force Research Laboratory and G. Pomrenke from the Air Force Office of Scientific Research, as well as other U.S. government agencies. References 1. S. Fukushima, C. Silva, Y. Muramoto, and A. Seeds, J. Lightwave Technol. 21, 3043 (2003). 2. R. K. Price, V. B. Verma, K. E. Tobin, V. C. Elarde, and J. J. Coleman, IEEE Photon. Technol. Lett. 19, 1610 (2007). 3. C. Cheng, Z. Ling-Juan, Q. Ji-Fiang, L. Yang, W. Wei, and L. Cai-Yun, Chin. Phys. B 21, 094208 (2012). 4. S. D. Roh, T. Yeoh, R. B. Swint, A. E. Huber, C. Y. Woo, J. S. Hughes, and J. Coleman, IEEE Photon. Technol. Lett. 12, 1307 (2000). 5. B. R. Koch, A. W. Fang, O. Cohen, and J. E. Bowers, Opt. Express 15, 11225 (2007). 6. A. R. Criado, C. de Dios, P. Acedo, G. Carpintero, and K. Yvind, J. Lightwave Technol. 30, 3133 (2012). 7. S. Ristic, A. Bhardwaj, M. J. Rodwell, L. A. Coldren, and L. A. Johansson, J. Lightwave Technol. 28, 526 (2010). 8. L. Ponnampalam, M. J. Fice, F. Pozzi, C. C. Renaud, D. C. Rogers, I. F. Lealman, D. G. Moodie, P. J. Cannard, C. Lynch, L. Johnston, M. J. Robertson, R. Cronin, L. Pavlovic, L. Naglic, M. Vidmar, and A. J. Seeds, J. Lightwave Technol. 29, 2229 (2011). 9. M. Lu, H. Park, E. Bloch, A. Sivananthan, A. Bhardwaj, Z. Griffith, L. A. Johansson, M. J. Rodwell, and L. A. Coldren, Opt. Express 20, 9736 (2012). 10. A. Criado, C. de Dios, E. Prior, G. Dohler, S. Preu, S. Malzer, H. Lu, A. Gossard, and P. Acedo, IEEE Trans. Terahertz Sci. Technol. 3, 461 (2013). 11. H. L. Stover and W. H. Steier, Appl. Phys. Lett. 8, 91 (1966). 12. L. Goldberg, H. F. Taylor, and J. F. Weller, Electron. Lett. 18, 1019 (1982). 13. L. Goldberg, H. F. Taylor, J. F. Weller, and D. M. Bloom, Electron. Lett. 19, 491 (1983).
4670
OPTICS LETTERS / Vol. 39, No. 16 / August 15, 2014
14. C. Laperle, M. Svilans, M. Porer, and M. Tetu, Opt. Express 17, 20727 (2009). 15. J. Huang, C. Sun, B. Xiong, and Y. Luo, Opt. Express 17, 20727 (2009). 16. C. Sun, J. Huang, B. Xiong, and Y. Luo, in OFC/NFOEC (2010), pp. 1–3. 17. L. Ponnampalam, C. Renaud, I. Lealman, L. Rivers, P. Cannard, M. Robertson, M. Moodie, F. van Dijk, A. Enard, F. Blache, M. Goix, F. Mallecot, and A. J. Seeds, “Injectionlocked integrated twin DBR lasers for mm-wave generation,” in European Workshop on Photonic Solutions for Wireless, Access, and In House Networks, Duisburg, Germany, 2009.
18. K. Kikuchi, C.-E. Zah, and T.-P. Lee, J. Lightwave Technol. 6, 1821 (1988). 19. G. J. Schneider, C. A. Schuetz, J. A. Murakowksi, S. Shi, and D. W. Prather, Nat. Photonics 7, 118 (2013). 20. D. W. Grund, G. A. Ejzak, G. J. Schneider, J. Murakowski, and D. W. Prather, J. Lightwave Technol. 32, 1363 (2014). 21. S. Shi, G. Schneider, and D. W. Prather, Proc. SPIE 8980, 898022 (2014). 22. C.-T. Lin, P.-T. Shih, J. Chen, W.-J. Jiang, S.-P. Dai, P.-C. Peng, Y.-L. Ho, and S. Chi, IEEE Trans. Microwave Theor. Tech. 57, 2084 (2009). 23. D. Scherer, RF and Microwave Measurement Symposium and Exhibition (Hewlett-Packard, 1981), p. 15.
4667
Improved configuration and reduction of phase noise in a narrow linewidth ultrawideband optical RF source David W. Grund, Jr.,* Shouyuan Shi, Garrett J. Schneider, Janusz Murakowski, and Dennis W. Prather Electrical and Computer Engineering, University of Delaware, Newark, Delaware 19716, USA *Corresponding author: [email protected] Received May 20, 2014; revised June 18, 2014; accepted June 19, 2014; posted June 25, 2014 (Doc. ID 212506); published August 5, 2014 In this Letter, we report on the improved configuration of a widely tunable optical RF generation system, particularly for the generation of low-frequency RF, as well as the reduction of phase noise in that same system. Using an amplitude modulator, a simplified system design was demonstrated with fewer components and improved phase noise performance, especially at RF frequencies below ∼36 GHz. Excess phase noise due to acoustic vibrations of the optical fibers was also successfully eliminated by mechanical isolation. A minimum phase noise of −124 dBc∕Hz at 10 kHz offset was demonstrated at 4 GHz. © 2014 Optical Society of America OCIS codes: (060.5625) Radio frequency photonics; (140.3520) Lasers, injection-locked; (060.2840) Heterodyne; (060.4080) Modulation. http://dx.doi.org/10.1364/OL.39.004667
Synthesis of radio frequency signals with a high spectral purity and ultrawide tuning range is of great interest for the realization of radio-over-fiber systems and instrumentation [1]. Optical RF generation inherently offers extreme bandwidth because small shifts in wavelength correspond to very large changes in frequency. For example, 200 GHz of RF bandwidth can be covered with a change in wavelength of just 1.6 nm when operating near 1550 nm. Optical RF is usually produced by heterodyning on a photodiode and can be implemented in a variety of ways. Systems have been demonstrated based on techniques such as direct mixing [2], direct mixing of modulation sidebands, multiwavelength lasers [3–6], optical phase locked loops (OPLL) [7–9], and various types of injection locking schemes [10–18]. Here we focus on the improved design of a widely tunable and spectrally pure optical RF generation system [19], particularly for the production of RF frequencies below ∼36 GHz, as well as the reduction of phase noise in that same system. This optical RF generation system is based on a modulation-sideband injection-locking scheme. Modulation-sideband injection-locking is advantageous because it offers continuous tunability and harmonic suppression without sacrificing power, while avoiding complex feedback mechanisms and the need for ultranarrow filters [20]. Phase noise has several potential sources in the optical RF generation system depicted in Fig. 1. In addition to the minimum phase noise of the scaled local RF oscillator, they include things such as spontaneous emission in the lasers, thermal and vibration effects on the optical fibers, path length mismatch, uncorrelated residual optical noise due to insufficient injection power, laser driving current fluctuations, and shot noise in the photodiode. The dominant two sources have been experimentally and theoretically determined to be due to the uncorrelated residual optical noise due to insufficient injection power and vibration effects in the optical fiber [19,21]. In the previous system [19], a phase modulator and two DWDM filters in series, as seen in the top callout of Fig. 1, were used to produce the sidebands for injection locking 0146-9592/14/164667-04$15.00/0
and reduce the strength of the carrier, which is typically stronger than the first order sidebands by 10 dB or more depending on the modulation depth. The filters had a total rolloff of approximately 1.33 dB∕GHz and an insertion loss of approximately 4 dB. These losses were the cause
Fig. 1. Layout and injection spectra of a radio frequency signal generation system based on injection-locked lasers. Modulation and carrier suppression were performed by a phase modulator and filters in Ref. [19] and by an amplitude modulator in this work. © 2014 Optical Society of America
4668
OPTICS LETTERS / Vol. 39, No. 16 / August 15, 2014
of insufficient injection power at frequencies below ∼36 GHz. It was also found that even at frequencies where there was adequate injection power, there was a large excess of phase noise above the scaled reference, for offset frequencies below 20 kHz, due to acoustic noise coupling into the loose optical fibers. In the bottom callout of Fig. 1 we can see the system layout under test in this work. We have replaced the phase modulator and filters with a single amplitude modulator. The amplitude modulator can be biased to the null point, performing carrier suppression, removing the need for the optical filter. This functionality makes the amplitude modulator configuration superior in several ways. First, we regain all the sideband power in the lower order harmonics that is lost in the filter from operating the master laser far enough down the filter edge to provide adequate suppression. Second, it eliminates the insertion loss associated with using the filter, providing a boost to the maximum injectable optical power. These two advantages are important because the noise suppression is governed by the injection power. Another advantage of this configuration is the ability to suppress the carrier at any optical frequency within the functional domain of the modulator, removing the limitation of working at the filter rejection band edge. This improves system flexibility. Finally, this configuration reduces the number of components and the amount of fiber through which mechanical vibrations can be coupled. The one possible drawback of using an amplitude modulator is that requiring a bias voltage could potentially lead to a long-term stability issue. Fortunately, amplitude modulators are extensively used in the telecommunications industry, and bias control feedback circuits are readily available to overcome this issue. In these experiments a single-arm-drive Z-cut lithium niobate modulator, seen in Fig. 2, was utilized. This modulator offers an extinction ratio greater than 20 dB for the carrier, which is sufficient to reduce it below the power of the first sidebands. The nulling capabilities could be improved by using a balanced-drive X-cut lithium niobate modulator with a dual-parallel Mach–Zehnder configuration, shown in Fig. 3. This allows for independent amplitude control in each arm in order to get optimal carrier extinction ratios at any working frequency. In addition, the balanced drive would allow for nulling the other even-order sidebands in addition to the carrier [22]. In the amplitude modulator system shown in Fig. 1, a tunable microwave signal generator (Agilent PSGE8267D) and a saturated broadband RF amplifier (Picosecond Pulse Labs 5882) that acts as a distortion element drive the amplitude modulator (EO Space AZ-AV1-40-PFU-PFU) that is biased simply by a voltage supply. The magnitude of the electric field of the harmonics produced by the saturated amplifier and passed
Fig. 2. Single-arm drive Z-cut lithium niobate Mach–Zender modulator.
Fig. 3. Dual-parallel balanced-drive X-cut lithium niobate Mach–Zender modulator.
through the modulator are given by J 1 x∕n, with a rolloff of 1∕nx∕2, while the harmonics produced by the nonlinearity of the modulator are given by J n x, with a rolloff of 1∕n!x∕2n , where x πV ∕V π . For the implemented system V ≤ 2.5 V, V π ∼ 5 V, and therefore x ≤ π∕2. This means that by the third harmonic, the modulator nonlinearity is contributing ≤1∕18 of the power and can be considered negligible. Therefore, the saturated amplifier is the primary source of harmonic power in the system. The harmonic sidebands are then amplified by a semiconductor optical amplifier (SOA) (Thorlabs S9FC1004P) and injected into the slave laser. An SOA is used instead of an EDFA because an EDFA has a minimum input power requirement, typically −3 dBm, which the sidebands produced in our system do not meet. The master and slave lasers (Mitsubishi FU-68PDF-5) are then mixed together on a photodiode (u2 t Photonics XPDV2150R) to produce the RF signal, which is amplified by a broadband RF amplifier (Picosecond Pulse Labs 5882) in order to prevent the phase noise measurement from being limited by the spectrum analyzer (Agilent PXA-N9030A) noise floor. In a frequency multiplication system such as this, the phase noise scales with the frequency multiplier N as 20log10 N [23]. It is noted that since the lasers in this system are not perfect filters, there is some residual content from the other injected harmonics. These are typically 25–40 dB weaker than the desired RF signal depending on the injected power. Such harmonic suppression is comparable to what is specified for typical commercial RF signal generators. In Fig. 4 we show that through mechanical isolation of the system, by mounting it on a stiff board, properly securing all fibers to the board with tape, and placing that board on foam, we were able to eliminate the fiber vibration noise, and closely match the scaled reference at all offset frequencies. Using the notation PNExcess PNSignal − PNScaled Reference ; Fig. 5 compares the difference between the optical RF and the scaled reference in this work and from the work of Ref. [19] and shows the dramatic improvement of up to 30 dB over the previously reported result at low offset frequencies. As discussed above, the minimum phase noise that can be achieved is limited to the phase noise of the scaled RF reference [21]. In order to see how well the noise had been suppressed, the optical RF generation from the 1st harmonic of 4 GHz was examined because it has a lower possible noise floor than the 36 GHz signal. In Fig. 6 we show the system’s best phase noise performance to date. This result has been made possible by the
August 15, 2014 / Vol. 39, No. 16 / OPTICS LETTERS
4669
Fig. 4. Comparison of the phase noise of the generated signal and reference at 36 GHz.
Fig. 6. Comparison of the phase noise of the generated signal and reference at 4 GHz.
elimination of approximately 30 dB of loss present in the phase modulator system due to the rolloff and insertion losses of the DWDM filters. It can be seen that there is a decent match at frequencies below 4 kHz at which point the noise levels out at −124 dBc∕Hz. Considering the shape of the curve, the difference between the reference and the optically generated signal is likely due to insufficient injection power to fully correlate the master and slave lasers. It is important to note that this minimum phase noise result is not just applicable at 4 GHz (N 1), but could be achieved at any frequency/multiplier, given a superior reference oscillator and enough injection power, because the minimum noise achievable is determined by the scaled reference oscillator, and the ability to conform the optically generated signal to the scaled reference oscillator is determined primarily by the injection strength [21]. In summary, we have demonstrated a system design using an amplitude modulator that offers improved system flexibility, operation with fewer components, and no sideband power loss, leading to improved phase noise performance at RF frequencies below ∼36 GHz. We also successfully eliminated the excess phase noise in our
system due to acoustic vibrations of the fibers by mechanically isolating the system. Finally we have demonstrated a new minimum phase noise of −124 dBc∕Hz at a 10 kHz offset from 4 GHz. In the future the system could be improved through the use of a dual-parallel Mach– Zehnder modulator and elimination of the fibers through complete system integration [20].
Fig. 5. Comparison of the difference between the phase noise of the generated signal and scaled reference at 36 GHz in this work and from the work of Ref. [19].
The authors would like to acknowledge the support of R. Nelson from the U.S. Air Force Research Laboratory and G. Pomrenke from the Air Force Office of Scientific Research, as well as other U.S. government agencies. References 1. S. Fukushima, C. Silva, Y. Muramoto, and A. Seeds, J. Lightwave Technol. 21, 3043 (2003). 2. R. K. Price, V. B. Verma, K. E. Tobin, V. C. Elarde, and J. J. Coleman, IEEE Photon. Technol. Lett. 19, 1610 (2007). 3. C. Cheng, Z. Ling-Juan, Q. Ji-Fiang, L. Yang, W. Wei, and L. Cai-Yun, Chin. Phys. B 21, 094208 (2012). 4. S. D. Roh, T. Yeoh, R. B. Swint, A. E. Huber, C. Y. Woo, J. S. Hughes, and J. Coleman, IEEE Photon. Technol. Lett. 12, 1307 (2000). 5. B. R. Koch, A. W. Fang, O. Cohen, and J. E. Bowers, Opt. Express 15, 11225 (2007). 6. A. R. Criado, C. de Dios, P. Acedo, G. Carpintero, and K. Yvind, J. Lightwave Technol. 30, 3133 (2012). 7. S. Ristic, A. Bhardwaj, M. J. Rodwell, L. A. Coldren, and L. A. Johansson, J. Lightwave Technol. 28, 526 (2010). 8. L. Ponnampalam, M. J. Fice, F. Pozzi, C. C. Renaud, D. C. Rogers, I. F. Lealman, D. G. Moodie, P. J. Cannard, C. Lynch, L. Johnston, M. J. Robertson, R. Cronin, L. Pavlovic, L. Naglic, M. Vidmar, and A. J. Seeds, J. Lightwave Technol. 29, 2229 (2011). 9. M. Lu, H. Park, E. Bloch, A. Sivananthan, A. Bhardwaj, Z. Griffith, L. A. Johansson, M. J. Rodwell, and L. A. Coldren, Opt. Express 20, 9736 (2012). 10. A. Criado, C. de Dios, E. Prior, G. Dohler, S. Preu, S. Malzer, H. Lu, A. Gossard, and P. Acedo, IEEE Trans. Terahertz Sci. Technol. 3, 461 (2013). 11. H. L. Stover and W. H. Steier, Appl. Phys. Lett. 8, 91 (1966). 12. L. Goldberg, H. F. Taylor, and J. F. Weller, Electron. Lett. 18, 1019 (1982). 13. L. Goldberg, H. F. Taylor, J. F. Weller, and D. M. Bloom, Electron. Lett. 19, 491 (1983).
4670
OPTICS LETTERS / Vol. 39, No. 16 / August 15, 2014
14. C. Laperle, M. Svilans, M. Porer, and M. Tetu, Opt. Express 17, 20727 (2009). 15. J. Huang, C. Sun, B. Xiong, and Y. Luo, Opt. Express 17, 20727 (2009). 16. C. Sun, J. Huang, B. Xiong, and Y. Luo, in OFC/NFOEC (2010), pp. 1–3. 17. L. Ponnampalam, C. Renaud, I. Lealman, L. Rivers, P. Cannard, M. Robertson, M. Moodie, F. van Dijk, A. Enard, F. Blache, M. Goix, F. Mallecot, and A. J. Seeds, “Injectionlocked integrated twin DBR lasers for mm-wave generation,” in European Workshop on Photonic Solutions for Wireless, Access, and In House Networks, Duisburg, Germany, 2009.
18. K. Kikuchi, C.-E. Zah, and T.-P. Lee, J. Lightwave Technol. 6, 1821 (1988). 19. G. J. Schneider, C. A. Schuetz, J. A. Murakowksi, S. Shi, and D. W. Prather, Nat. Photonics 7, 118 (2013). 20. D. W. Grund, G. A. Ejzak, G. J. Schneider, J. Murakowski, and D. W. Prather, J. Lightwave Technol. 32, 1363 (2014). 21. S. Shi, G. Schneider, and D. W. Prather, Proc. SPIE 8980, 898022 (2014). 22. C.-T. Lin, P.-T. Shih, J. Chen, W.-J. Jiang, S.-P. Dai, P.-C. Peng, Y.-L. Ho, and S. Chi, IEEE Trans. Microwave Theor. Tech. 57, 2084 (2009). 23. D. Scherer, RF and Microwave Measurement Symposium and Exhibition (Hewlett-Packard, 1981), p. 15.
