Brillouin-based phase shifter in a silicon waveguide
Luke Mckay, Moritz Merklein, Alvaro Casas Bedoya, Amol Choudhary, Micah Jenkins, Charles Middleton, Alex Cramer, Joseph Devenport, Anthony Klee, Richard DeSalvo, Benjamin J. Eggleton
BBrillouin-based phase shifter in a silicon waveguide L UKE M C K AY , M ORITZ M ERKLEIN , A
LVARO C ASAS B EDOYA , A MOL C HOUDHARY ,M ICAH J ENKINS , C HARLES M IDDLETON , A LEX C RAMER , J OSEPH D EVENPORT , A NTHONY K LEE , R ICHARD D E S ALVO , AND B ENJAMIN
J. E
GGLETON The University of Sydney Nano Institute (Sydney Nano), Institute of Photonics and Optical Science (IPOS), School of Physics, The University ofSydney, 2006, Sydney, NSW Australia Harris Corporation, 1025 W. NASA Boulevard, Melbourne, Florida 32919, United States + Current Address: Department of Electrical Engineering, Indian Institute of Technology, Delhi, 110016, India * [email protected] November 7, 2019 Integrated silicon microwave photonics offers great potential in microwave phase shifter ele-ments, and promises compact and scalable multi-element chips that are free from electromag-netic interference. Stimulated Brillouin scattering, which was recently demonstrated in silicon,is a particularly powerful approach to induce a phase shift due to its inherent flexibility, offeringan optically controllable and selective phase shift. However, to date, only moderate amounts ofBrillouin gain has been achieved and theoretically this would restrict the phase shift to a fewtens of degrees, significantly less than the required . Here, we overcome this limitation witha phase enhancement method using RF interference, showing a broadband phase shifterbased on Brillouin scattering in a suspended silicon waveguide. We achieve a full phase-shift over a bandwidth of 15 GHz using a phase enhancement factor of 25, thereby enablingpractical broadband Brillouin phase shifter for beam forming and other applications.
1. INTRODUCTION
Integrated microwave photonics (IMWP) holds great promisefor many applications including radio frequency (RF) phaseshifters and filters. IMWP benefits from established photonicphase manipulation techniques and has the inherent advantageof broad bandwidths and immunity to RF interference [1, 2]. RFphase shifting devices are particularly important for phased ar-ray antennas (PAAs) which can dynamically change their beamprofile by controlling the phase of the signal to each successiveantenna element, giving them a high degree of flexibility [3, 4].PAAs are quickly becoming integral components for many ap-plications such as satellite communication, RADAR, medicalimaging, 5G networks and sensing [4, 5].Silicon is the most actively researched material in the field ofintegrated photonics largely due to its compatibility with com-plementary metal oxide semiconductor (CMOS) processes, highdamage threshold and strong optical nonlinearities [6]. SiliconIMWP offers a massive reduction in size, weight and powerconsumption (SWaP) allowing a high degree of integration ofcomplex optical structures with low loss on a chip [7–10].The ideal IMWP phase shifter is able to induce a continuouslytunable 360° phase shift over a broad bandwidth with minimalinsertion loss, small amplitude fluctuations and is integrated on a chip with a small footprint and overall low power consump-tion. In an IMWP phase shifter, an input RF signal is modulatedon an optical carrier creating sidebands. An optical phaseshift is then applied to either the optical carrier or sideband,which leads to a phase shift of the RF output signal uponphotodetection (Fig. 1 b). There have been many approaches ofutilizing an optical phase shift to achieve an RF phase shifter inIMWP systems. For example, IMWP phase shifters have beenrealized based on the electro-optic effect utilizing free carrierdispersion and multiplexing/ de-multiplexing of the opticalcarrier and sidebands [11]. Another way of generating theoptical phase shift on a chip is with resonant structures, such asrings and gratings [12–14]. However, amplitude dependenceof the induced phase shift and in the case of rings, limitedtunability due to the free-spectral range are challenges forbroadband applications.On the other hand, Brillouin scattering offers a powerfulsolution as it can induce a tunable, gain-based phase shiftover a narrow bandwidth of around 30 MHz [15, 16]. Morerecently, forward Brillouin scattering has been demonstrated inan integrated silicon platform using a suspended waveguidestructure, which is capable of guiding both the photons andphonons [17–19]. Microwave photonic phase-shifting schemes1 a r X i v : . [ phy s i c s . op ti c s ] N ov aser PD AmplitudePhaseBrillouin responsea)b) Si SiO Si 220 nm Θ c) PhaseshiftSi chipRFin RFoutPhaseshiftMODMOD νν C Θ C SB ν C SB ν Fig. 1. a) Block diagram of a microwave photonic phase shifter.An input RF signal is modulated onto an optical carrier. Thephase of that microwave envelope is shifted in the optical do-main using Brillouin scattering in a silicon waveguide (inset)that leads to a phase shifted RF output upon photodetection.b) Time domain representation of an optical carrier modulatedby an RF signal and c) the according frequency domain rep-resentation. Shifting the phase of the optical sideband resultsin a phase shifted RF output signal upon beating the opticalcarrier and sideband at the photodetector. SB, sideband; C,carrier; MOD, modulator; PD, photodetector; ν , frequency; Si,silicon; SiO , silica.based on backward stimulated Brillouin scattering (SBS) havepreviously been implemented in optical fibre [20, 21]. Thisconcept has been further developed to implement the schemein an integrated chalcogenide platform, however, due to thelimited available on-chip Brillouin gain only a 240° phase shiftwas achieved [22]. To achieve a 360° phase shift, more Brillouingain would be required, demanding a very large amount ofpump power. Even record high Brillouin gain in optimizedchalcogenide waveguides of 52 dB could only achieve a ± ± Optical ν Electricala) ν ν Θ b) Θ Θ ν Θ Fig. 2. a) Narrowband RF tone applied to the optical carrier.The sideband beats with the optical carrier to generate an RFtone in the electrical domain (green). The phase of the RF toneexperiences a phase shift of θ when a phase shift of θ is ap-plied to the carrier. b) A broadband RF signal modulated onthe optical carrier. Similarly, the sideband will beat with thecarrier to produce a broadband output signal. Similarly, thephase of the broadband signal experiences a phase shift of θ when a phase shift of θ is applied to the carrier.
2. PRINCIPLE AND SETUP
The core principle of a broadband microwave phase shifter isshown in Fig. 1 a). An input RF signal (shown in green) ismodulated upon an optical carrier. The optical signal propagatesthrough the system where it experiences a phase shift beforedetection, shown in the time domain in Fig. 1 b). In the frequencydomain, the modulation process results in a carrier with opticalsidebands and the phase of the optical sideband is then shifted.This causes a shift in the phase of the RF output signal that isgenerated at the photodetector by beating the carrier with thesideband (see Fig. 1 c).In this work, the optical phase shift is induced using Brillouinscattering. Brillouin scattering involves two co-propagatingoptical waves interfering to generate an electrostrictive opticalforce, which in turn generates phonons which are guided in thesuspended waveguide structure. The moving phonons allowone optical signal to coherently amplify and induce a phase shiftin the other. Brillouin scattering is a narrowband χ nonlineareffect that resonantly couples two optical waves - known asStokes wave ω s and a higher frequency pump wave ω p - withan acoustic wave Ω . This process needs to fulfill energy con-servation given by Ω = ω p - ω s and momentum conservationwhich can be written as K( Ω ) = k( ω p ) - k( ω s ), where K( Ω ) is thewavevector of the acoustic wave and k( ω p / s ) the wavevector ofthe optical pump and Stokes wave, respectively. While Brillouin2 RF net a) RF Pump Ω B π sb P h a s e A m p li t u d e b) Optical π π π c)2C + SB SB + C νν νν sb RF Optical RF vectorTypical MWP phase shifter Enhanced MWP phase shifter π π ν RF Fig. 3. a) Basic scheme of the two optical carriers (C , C ) and sidebands (SB , SB ) used in the enhanced phase shifter scheme. Thelower sideband of C and the upper sideband of C are removed by a filter resulting in out-of-phase RF signals. A Brillouin pumpinduces an amplification and phase shift of C . b) Optical and RF phase space representation of a typical MWP phase shifter. Anoptical phase shift θ applied to one of the sidebands relates directly to a shift θ of the RF phase. c) Enhanced MWP phase shifterbased on interference and RF vector addition in phase space. As can be seen in the phase space diagram, a small optical phase shift θ can lead to a much larger RF phase shift θ of the resultant vector RF net (red arrow).scattering can occur in either the forward or backward directionin silicon it is typically studied in forward direction and hencethe acoustic wave is mainly a transverse wave and hence thewavevector difference between the pump and Stokes wave issmall.The optical signal ω p coherently amplifies an optical probe thatis located at the Stokes frequency ω s and in the process is in-ducing a phase shift as a consequence of the Kramers-Kronigrelations [25]. This is due to a requirement of causality, whichstates that a change in amplitude has a corresponding changein refractive index, and hence it experiences a phase shift [26].The induced optical phase shift due to Brillouin scattering scaleslinearly with Brillouin gain [21]. As we are operating in a lowgain regime the Brillouin gain itself is linearly proportional tothe effective length and the pump power. The width of theBrillouin resonance and the according phase shift is given bythe inverse of the lifetime τ of the acoustic wave and is in oursilicon waveguides around 30 MHz. To achieve a 360° phaseshift, however, more than 50 dB of Brillouin gain is required [21],which would demand an impractical amount of pump power.The maximum achievable gain in silicon is currently far belowthe threshold required to achieve a 360° phase shift [18].To achieve a broadband RF phase shift from the narrowbandBrillouin phase response, the phase shift is applied to the opticalcarrier instead of the sideband. A single frequency RF tone ismodulated upon the optical carrier and the resulting sideband isseparated by the frequency of the RF signal. If no phase shift isapplied to the carrier, the resultant RF signal has a phase definedas 0. However, when a phase θ is applied to the carrier, thephase of the resultant RF signal will shift to θ as shown in Fig.2 a). Similarly, this concept expands to the broadband regimewhere the RF input signal is modulated over a wide frequencyrange 2 b). If a phase shift θ is then applied to the optical carrier,the entire broadband RF signal will experience a uniform phaseshift θ . So by applying a narrowband phase shift to the carrier, abroadband RF phase shifter can be achieved.It has previously been shown that a phase shift can be enhancedusing interferometry [27–29]. However, so far these schemeshave only been applied over a relative narrow bandwidth [27–29]. Here, on the other hand, we propose and demonstrate aphase enhancement scheme that has a broad bandwidth of sev- eral GHz through acting on the carrier. Phase enhancement isachieved by destructively interfering an out-of-phase RF signalwith the original RF signal. The optical components which gen-erate this interference are shown in Fig. 3 a). The optical carrierfrom Laser 1 (yellow) and Laser 2 (blue) pass through the samephase modulator, producing sidebands. However, the lowersideband of carrier 2 and the upper sideband of carrier 1 arefiltered ensuring the remaining sidebands are out of phase. Thisassures that the generated RF signals are out-of-phase whenbeating on the photodetector. In the typical Brillouin phaseshifter scheme only laser 1 is connected, the entire phase shiftcomes from the induced Brillouin phase response as shown inFig. 3 b). When no gain is present, the resultant signal labeledRF has a phase of 0, but when the pump is on and a phaseis applied to the carrier, the resultant phase is θ . However, inthe phase enhanced scheme, the RF signal (RF ) from carrier 1destructively interferes with the RF signal from carrier 2 (RF )resulting in a smaller net vector RF net as shown in Fig. 3 c). TheBrillouin phase shift is applied to the carrier of RF which wasset to be slightly smaller than RF . Due to the Brillouin amplifi-cation and the induced phase shift the ratio between the two RFsignals changes and the resultant phase shift can be calculatedusing vector addition. It is worth noting that the phase shift ofthe resultant vector θ is in the opposite direction to the phaseshift experienced by RF , which is given by θ in Fig. 3 c). Largephase shifts from only a minimum amount of Brillouin ampli-fication can be achieved when the amplitudes of RF and RF prior to applying the Brillouin gain are closely matched. How-ever, due to the destructive interference inherent in the phaseenhancement scheme, there is an unavoidable trade-off betweenlink gain and the phase enhancement factor.3 MLaser 2
EDFA
BPF BPFDual BPFDPMZM MS Bias
OSA
VNA
PDOPMOPM
Laser 1 ISO Si Fig. 4.
Schematic of the experimental setup. Laser 1 is split in two arms; one is used as carrier for the input RF signal and the otherone acts as the Brillouin pump. Laser 2 is added to achieve the phase enhancement. Inset: scanning electron microscope (SEM)imagine of a suspended silicon waveguide. ISO: optical isolator; BPF: band-pass filter; DPMZM: dual-parallel Mach-Zehnder mod-ulator; EDFA: Erbium doped fibre amplifier; VNA: vector network analyzer; OPM: optical power meter; OSA: optical spectrumanalyzer; PD: photodetector; PM: phase modulator; VCO: voltage controlled oscillator; Si, Silicon chip. p h a s e s h i f t ( d e g r ee s ) Fig. 5.
Phase profile without phase enhancement (blue) andwith phase enhancement (orange).The experimental setup of our Brillouin RF phase shifter isshown in Fig. 4. A narrow linewidth distributed feedback (DFB)laser (laser 1) is split into two paths using an optical coupler, apump path and a signal path. The RF input is modulated onthe optical carrier in the signal path using a phase modulator,generating two sidebands. One of these sidebands is then fil-tered out. The carrier in the pump path is frequency-shiftedusing a dual-parallel Mach-Zehnder modulator (DPMZM) andis amplified afterwards with an Erbium doped fibre amplifier (EDFA). Both arms are then recombined before they are coupledinto the chip using grating couplers. The integrated waveguidein this demonstration is a suspended silicon rib structure withdimensions of 3050 nm wide, 150 nm thick membrane with acentral ridge protruding 70 nm with a width of 650 nm and alength of 2 cm. The suspended architecture enables the waveg-uide to guide photons and phonons without them dissipatinginto the silica [18].The waveguide was fabricated at IMEC through ePIXfab. Thenanowires were immersed in 10% diluted hydrofluoric acidwith an etching rate of 40 nm/min for 20 minutes to releasethem from the silica substrate. The insertion loss of the opticalwaveguide is 13 dB, consisting of 3.5 dB coupling loss per facetusing grating couplers and a transmission loss of 3 dB/cm. Thewaveguide used has a slightly lower gain coefficient than whatwas achieved in previous demonstrations [18]. Our waveguidestructure has a reduced footprint as it is arranged in a spiralbut also shows larger losses which limited the achievable netgain. Furthermore, inhomogeneous broadening might limit thegain [30], even though it has been shown to be less severe insuspended ridge waveguides compared to nanowires [18, 24].After the waveguide a highly selective band-pass filter isused to remove the optical pump. The optical carrier and side-band are detected at the photodetector. To achieve the phaseenhancement a second laser is coupled into the same phase mod-ulator as laser 1 (see dotted green box in Fig.4). After the phasemodulator, a dual band-pass filter is used to selectively removethe opposite sideband from each optical carrier, so the resultingRF tones are 180° out of phase.
3. EXPERIMENTAL RESULTS
We first measure the phase response of the forward Brillouinresonance in the silicon waveguide (see Fig. 5 shown in blue).We use a vector network analyser (VNA) to sweep an opticalprobe signal through the Brillouin resonance that is measured tobe 4.42 GHz. The linewidth of the resonance is around 30 MHz4 ) b) ±7.3° p h a s e s h i f t ( d e g r ee s ) frequency (GHz) frequency (GHz) p h a s e s h i f t ( d e g r ee s ) Fig. 6. a) Broadband phase shift without phase enhancement. b) Broadband phase shift with an enhancement factor of 25 (a low-pass filter was applied to the data).and we can see a ±
10° of phase shift for around 2 dB of Brillouingain. This moderate amount of phase shift can be enhanced byadding the second, out-of-phase RF signal via the second laser(Fig. 5 shown in orange). With the second laser and the vectorialaddition of the two RF components we achieve a full 360° phaseresponse with the same amount of pump power and Brillouingain. Hence a phase enhancement factor of 18 was achieved inthis demonstration.We now want to implement the Brillouin phase shift in a broad-band RF phase shifter configuration. Therefore we apply theBrillouin phase shift on the optical carrier signal. This allows usto sweep the sideband over a wide RF frequency range relativeto the carrier while all frequencies experience the same phaseshift. Using this technique a flat phase shift, i.e. a minimumdependence of the phase shift on RF frequency, can be achievedover tens of GHz that is only limited by the bandwidth of theelecto-optic components such as the modulator and the photode-tector. Figure 6 a) shows a broadband RF phase shift of ± ± and RF signals interfere.However, these challenges are not fundamental to the schemeand can be overcome by using, for example, a dual waveguidescheme that has the pump in a different waveguide than thecarrier signal with an acoustic mode coupling the two modes, asrecently demonstrated [31].Another important metric for microwave photonic phaseshifter is the amplitude variation with the phase shift. We mea-sured the amplitude variation over the full 360° phase shift inthe frequency window from 5 GHz to 20 GHz with a phase en-hancement factor 25 (see Fig. 7). The measured variation is ±
4. DISCUSSION
It is foreseeable that the system could be implemented entirelyin an integrated platform as all the required components havealready been demonstrated. Integrating the whole schemeon a chip would not only further reduce size, weight andpower (SWaP) consumption but also increase the stability of theinterference-based scheme due to smaller fluctuations in relativeoptical path length. Furthermore, avoiding coupling to and fromthe chip would decrease losses, which combined with improved5 a m p li t u d e v a r i a t i o n ( d B ) frequency (GHz)5 10 15 20-15150 Fig. 7.
Amplitude variation for a phase shift range of 360° us-ing a factor of 25 phase enhancement. The data is normalizedto the underlying RF link.Brillouin gain, would reduce the required enhancement factorand thus increase the overall link gain. In addition, improvingthe RF amplifier, reducing waveguide losses, increasing filterroll-off to remove the pump before photodetection, or awaveguide with a larger Brillouin frequency shift to ease thefilter requirements, would further improve the overall RF linkgain.There are many possible pathways to integrating the whole
DPMZM LaserPMSWGAWG RRPD
Fig. 8.
Artists impression of the phase shifting scheme entirelyintegrated on a chip. AWG, arrayed waveguide grating; SWG,suspended waveguide; PD, photodetector; PM, phase modu-lator; DPMZM, dual-parallel Mach-Zehnder modulator; RR,ring resonator.system on a chip, with the individual devices already showinga high level of development and performance. For example,integrated optical photodetectors have been demonstratedwith bandwidths over 100 GHz and high output power [32, 33].Also, on-chip modulators have been demonstrated basedon different effects, including plasmonic, P-N junctions, andelectro-optic materials [34–36]. Laser sources in silicon are still abig challenge due to the indirect bandgap of silicon, however,great progress has been made in III-V based lasers sources[37] that can be combined with silicon chips using differentbonding techniques, such as flip-chip bonding [38, 39]. Otherrequired components such as integrated optical filters anddemultiplexers such as arrayed waveguide gratings (AWGs) arealready mature technologies [40, 41].
5. CONCLUSION
Here, we have demonstrated a 360° Brillouin-based RF phaseshifter over a bandwidth of 15 GHz, with only minimal ampli-tude fluctuations in a CMOS-compatible silicon waveguide. Weintroduced a broadband RF interference scheme that enhancedthe phase shift by a factor of 25, which allowed us to achieve a360° phase shift from only 1.6 dB of Brillouin amplification. Thephase enhancement scheme does not only greatly reduce thepower requirement, but also enables the first demonstration ofan on-chip silicon Brillouin-based phase shifter that achieves afull 360° phase-shift. This is despite silicon waveguides havinga significantly reduced absolute Brillouin gain when comparedto chalcogenide rib waveguides.A Brillouin-based approach has many advantages over otherschemes such as the narrow bandwidth of the phase response,which can be selectively applied to only the carrier. This narrowbandwidth and high selectivity allow the phase shift to be ap-plied at low RF frequencies, where the sidebands are close to thecarrier without being effected. Furthermore, one can apply thephase shifter scheme to individual carriers in a wavelength mul-tiplexed scheme. For example, it has been shown that Brillouinscattering can be used to process individual spectral lines of afrequency comb in a single waveguide [42, 43]. This would notbe possible with inherently broadband phase shifting technolo-gies such as III-V media or thermo-optic phase shifters.The broadband phase enhancement concept introduced hereis not limited to Brillouin-based phase shifters and can find manyother practical applications. It can in general greatly reduce thepower requirements of phase shifters, delay lines and true-timedelay schemes but can also be utilized in sensing schemes thatrely on measuring a phase shift; in the latter, a potentially smallphase shift can be enhanced and in the process the sensitivitycan be increased.
6. ACKNOWLEDGEMENTS
Australian Research Council (ARC) Linkage grant (LP170100112)with Harris Corporation. U.S. Air Force (USAF) throughAFOSR/AOARD (FA2386-16-1-4036); U.S. Office of Naval Re-search Global (ONRG) (N62909-18-1-2013). We acknowledge thefabrication of the SOI nanowires, which were done in the frame-work of the ePIXnet and the University of New SouthWales(UNSW) node of the Australian National Fabrication Facility(ANFF) where the samples were etched.
REFERENCES
1. D. Marpaung, J. Yao, and J. Capmany, “Integrated mi-crowave photonics,” Nat. Photonics , 80–90 (2019).2. J. Capmany, B. Ortega, and D. Pastor, “A tutorial on mi-crowave photonic filters,” J. Light. Technol. , 201–229(2006).3. J. Volakis, Antenna Engineering Handbook, (McGraw-HillProfessional, 2007), fourth edition ed.4. R. C. Hansen,
Phased Array Antennas (John Wiley & Sons,Inc., 2009).5. A. Naqvi and S. Lim, “Review of recent phased arraysfor millimeter-wave wireless communication,” Sensors. ,3194 (2018).6. J. Leuthold, C. Koos, and W. Freude, “Nonlinear siliconphotonics,” Nat. Photonics , 535–544 (2010).7. B. Jalali and S. Fathpour, “Silicon photonics,” J. Light. Tech-nol. , 4600–4615 (2006).6. D. Thomson, A. Zilkie, J. E. Bowers, T. Komljenovic, G. T.Reed, L. Vivien, D. Marris-Morini, E. Cassan, L. Virot, J.-M.Fédéli, J.-M. Hartmann, J. H. Schmid, D.-X. Xu, F. Boeuf,P. O’Brien, G. Z. Mashanovich, and M. Nedeljkovic,“Roadmap on silicon photonics,” J. Opt. , 073003 (2016).9. Q. Lin, O. J. Painter, and G. P. Agrawal, “Nonlinear opticalphenomena in silicon waveguides: Modeling and applica-tions,” Opt. Express , 16604–16644 (2007).10. A. H. Atabaki, S. Moazeni, F. Pavanello, H. Gevorgyan,J. Notaros, L. Alloatti, M. T. Wade, C. Sun, S. A. Kruger,H. Meng, K. A. Qubaisi, I. Wang, B. Zhang, A. Khilo, C. V.Baiocco, M. A. Popovi´c, V. M. Stojanovi´c, and R. J. Ram,“Integrating photonics with silicon nanoelectronics for thenext generation of systems on a chip,” Nature , 349–354(2018).11. C. Porzi, G. Serafino, M. Sans, F. Falconi, V. Sorianello,S. Pinna, J. E. Mitchell, M. Romagnoli, A. Bogoni, andP. Ghelfi, “Photonic integrated microwave phase shifter upto the mm-wave band with fast response time in silicon-on-insulator technology,” J. Light. Technol. , 4494–4500(2018).12. M. Burla, L. R. Cortés, M. Li, X. Wang, L. Chrostowski,and J. A. na, “On-chip programmable ultra-wideband mi-crowave photonic phase shifter and true time delay unit,”Opt. Lett. , 6181–6184 (2014).13. A. Meijerink, C. G. H. Roeloffzen, R. Meijerink, L. Zhuang,D. A. I. Marpaung, M. J. Bentum, M. Burla, J. Verpoorte,P. Jorna, A. Hulzinga, and W. van Etten, “Novel ringresonator-based integrated photonic beamformer for broad-band phased array receive antennas - part i: Design andperformance analysis,” J. Light. Technol. , 3–18 (2010).14. L. Zhuang, C. G. H. Roeloffzen, A. Meijerink, M. Burla,D. A. I. Marpaung, A. Leinse, M. Hoekman, R. G. Hei-deman, and W. van Etten, “Novel ring resonator-basedintegrated photonic beamformer for broadband phased ar-ray receive antennas - part ii: Experimental prototype,” J.Light. Technol. , 19–31 (2010).15. M. Merklein, A. Casas-Bedoya, D. Marpaung, T. F. S. But-tner, M. Pagani, B. Morrison, I. V. Kabakova, and B. J. Eggle-ton, “Stimulated brillouin scattering in photonic integratedcircuits: Novel applications and devices,” IEEE J. Sel. Top.Quantum Electron. , 336–346 (2016).16. B. J. Eggleton, C. G. Poulton, and R. Pant, “Inducing andharnessing stimulated brillouin scattering in photonic inte-grated circuits,” Adv. Opt. Photonics , 536 (2013).17. H. Shin, W. Qiu, R. Jarecki, J. A. Cox, R. H. Olsson, A. Star-buck, Z. Wang, and P. T. Rakich, “Tailorable stimulatedBrillouin scattering in nanoscale silicon waveguides,” Nat.Commun. (2013).18. E. A. Kittlaus, H. Shin, and P. T. Rakich, “Large Brillouinamplification in silicon,” Nat. Photonics , 463–467 (2016).19. A. H. Safavi-Naeini, D. V. Thourhout, R. Baets, and R. V.Laer, “Controlling phonons and photons at the wavelengthscale: integrated photonics meets integrated phononics,”Optica , 213 (2019).20. A. Loayssa and F. J. Lahoz, “Broad-band RF photonic phaseshifter based on stimulated Brillouin scattering and single-sideband modulation,” IEEE Photonics Technol. Lett. ,208–210 (2006).21. M. Pagani, D. Marpaung, and B. J. Eggleton, “Ultra-wideband microwave photonic phase shifter with con-figurable amplitude response,” Opt. Lett. , 5854–5857(2014). 22. M. Pagani, D. Marpaung, D.-Y. Choi, S. J. Madden,B. Luther-Davies, and B. J. Eggleton, “Tunable widebandmicrowave photonic phase shifter using on-chip stimulatedBrillouin scattering,” Opt. Express , 28810–28818 (2014).23. I. Aryanfar, D. Marpaung, A. Choudhary, Y. Liu, K. Vu,D.-Y. Choi, P. Ma, S. Madden, and B. J. Eggleton, “Chip-based brillouin radio frequency photonic phase shifter andwideband time delay,” Opt. Lett. , 1313–1316 (2017).24. R. V. Laer, B. Kuyken, D. V. Thourhout, and R. Baets, “In-teraction between light and highly confined hypersoundin a silicon photonic nanowire,” Nat. Photonics , 199–203(2015).25. D. C. Hutchings, M. Sheik-Bahae, D. J. Hagan, and E. W.Van Stryland, “Kramers-krönig relations in nonlinear op-tics,” Opt. Quantum Electron. , 1–30 (1992).26. G. Agrawal, Nonlinear Fiber Optics (Optics and Photonics) (Academic Press, 2006).27. M. B. Ayun, A. Schwarzbaum, S. Rosenberg, M. Pinchas,and S. Sternklar, “Photonic radio frequency phase-shiftamplification by radio frequency interferometry,” Opt. Lett. , 4863–4866 (2015).28. Y. Liu, A. Choudhary, D. Marpaung, and B. J. Eggleton,“Gigahertz optical tuning of an on-chip radio frequencyphotonic delay line,” Optica. , 418–423 (2017).29. A. Choudhary, Y. Liu, B. Morrison, K. Vu, D.-Y. Choi,P. Ma, S. Madden, D. Marpaung, and B. J. Eggleton, “High-resolution, on-chip RF photonic signal processor using Bril-louin gain shaping and RF interference,” Sci. Reports (2017).30. C. Wolff, R. Van Laer, M. J. Steel, B. J. Eggleton, and C. G.Poulton, “Brillouin resonance broadening due to struc-tural variations in nanoscale waveguides,” New J. Phys. ,25006 (2016).31. E. A. Kittlaus, N. T. Otterstrom, P. Kharel, S. Gertler, and P. T.Rakich, “Non-reciprocal interband brillouin modulation,”Nat. Photonics , 613–619 (2018).32. A. Beling, X. Xie, and J. C. Campbell, “High-power, high-linearity photodiodes,” Optica , 328–338 (2016).33. Q. Li, K. Li, Y. Fu, X. Xie, Z. Yang, A. Beling, and J. C. Camp-bell, “High-power flip-chip bonded photodiode with 110GHz bandwidth,” J. Light. Technol. , 2139–2144 (2016).34. C. Wang, M. Zhang, X. Chen, M. Bertrand, A. Shams-Ansari, S. Chandrasekhar, P. Winzer, and M. Lonˇcar, “In-tegrated lithium niobate electro-optic modulators operat-ing at CMOS-compatible voltages,” Nature. , 101–104(2018).35. J. A. Dionne, K. Diest, L. A. Sweatlock, and H. A. Atwa-ter, “PlasMOStor: A metal-oxide-si field effect plasmonicmodulator,” Nano Lett. , 897–902 (2009).36. K. Liu, C. R. Ye, S. Khan, and V. J. Sorger, “Review andperspective on ultrafast wavelength-size electro-optic mod-ulators,” Laser & Photonics Rev. , 172–194 (2015).37. Z. Zhou, B. Yin, and J. Michel, “On-chip light sources forsilicon photonics,” Light. Sci. & Appl. , e358–e358 (2015).38. N. Fujioka, T. Chu, and M. Ishizaka, “Compact and lowpower consumption hybrid integrated wavelength tunablelaser module using silicon waveguide resonators,” J. Light.Technol. , 3115–3120 (2010).39. S. Tanaka, S.-H. Jeong, S. Sekiguchi, T. Kurahashi, Y. Tanaka,and K. Morito, “High-output-power, single-wavelengthsilicon hybrid laser using precise flip-chip bonding technol-ogy,” Opt. Express , 28057–28069 (2012).40. C. Dragone, “An n*n optical multiplexer using a planar7rrangement of two star couplers,” IEEE Photonics Technol.Lett. , 812–815 (1991).41. Y. Kokubun, “Waveguide filters and related technolo-gies:issues and solutions for practical use in transmissionsystems,” J. Light. Technol. , 6–18 (2018).42. M. Pelusi, A. Choudhary, T. Inoue, D. Marpaung, B. J. Eggle-ton, K. Solis-Trapala, H. N. Tan, and S. Namiki, “Low noisefrequency comb carriers for 64-qam via a brillouin combamplifier,” Opt. Express , 17847–17863 (2017).43. A. Choudhary, M. Pelusi, D. Marpaung, T. Inoue, K. Vu,P. Ma, D.-Y. Choi, S. Madden, S. Namiki, and B. J. Eggleton,“On-chip brillouin purification for frequency comb-basedcoherent optical communications,” Opt. Lett.42