Fiber optical parametric amplifier (FOPA), which is based on the four-wave mixing (FWM) effect in optical fibers, is an important amplifier in fiber-based communication systems. To date, FOPAs have extensively studied in variety of single mode fibers. Recently, few-mode fiber (FMF) has attracted much attention because of its potential for providing further increase in per-fiber transmission capacity via mode-division multiplexing (MDM) technology. To amplify the signal of MDM system, few-mode FOPA (FM-FOPA) with high gain and large bandwidth are required. So far, a lot of efforts have been made on proposing the structure and design of FMFs for simultaneously amplifying the telecom band signals in different spatial modes via FWM in FMFs, however, the experimental demonstration has not been carried out yet. In this work, using 90-m-long homemade few-mode dispersion-shifted fiber, we demonstrate the first experimental realization of FM-FOPA and study its gain dependence on polarization and spatial mode. The gain spectra of the intramodal FWMs in LP01 and LP11 modes are in the telecom C and S bands, respectively. When the average powers of pulsed pump in LP01 and LP11 modes are 7 mW and 10 mW, the measured gains are about 24.5 dB and 7 dB, respectively. Moreover, we show that the gain equalized amplification can be realized for 1535 nm seed injection in LP01 and LP11 mode, respectively. Our investigation has potential application in developing low noise amplifier for MDM communication systems.
Spatial-mode-selective frequency conversion is potentially useful for both classical and quantum communication applications. By a judicious choice of the quasi-phase-matching period in a Kai(2) multimode waveguide, such conversion can be achieved with high efficiency (close to 100%) and with low crosstalk (< -20 dB). For space-division multiplexing application with classical signals, where each spatial mode represents a separate signal channel, the selective conversion of a spatial mode without disrupting other signal modes can be used for reconfigurable spatial-mode de-multiplexing. This classical de-multiplexing capability can be also extended to the quantum regime, where the quantum state of the signal is preserved during frequency conversion, owing to the unitary nature of the sum-frequency generation (SFG) process.
Building upon our previous experimental demonstration of the classical spatial-mode-selective frequency up-conversion in a two-mode PPLN waveguide, here we report the extension of this work into the single-photon-level regime. The signal (1540 nm) in either a single mode (TM00 or TM01) or a superimposition mode (TM00+TM01, TM00+iTM01) of the waveguide is selectively up-converted into TM01 SFG mode, by interacting with an appropriate pump mode (1560 nm). An accurate measurement of the single-photon-level SFG signals requires thorough filtering of the unwanted photons contributed by the second harmonic of the pump, residual pump noise extending to the signal band, and the Raman noise generated in the waveguide. We have investigated these unwanted photon sources and suppressed them by a combination of thin-film-interference and volume-Bragg-grating filters. Resulting single-photon-counting measurements show >70% internal conversion efficiency, better than -12dB crosstalk, and >100 ratio of the signal to background photon counts for all selected modes and mode superpositions.
We investigate spatial-mode-selective frequency up-converters of quantum states from infrared to visible region, which could be useful not only for interfacing the optical fiber links with quantum memories and for increasing the photon detection efficiency, but also for classical demultiplexing of spatial modes that are otherwise difficult to discriminate in both spatial and spatial-frequency domains. We consider two approaches: first, based on sum-frequency generation (SFG) in 2D free space, and second, based on SFG in a multimode waveguide with 2D confinement. For the latter approach, we find that under proper quasi-phase-matching arrangement, several different pairs of signal and pump modes are converted to the same SFG mode. By adjusting the relative phases and magnitudes of the pump modes, any superposition of the corresponding signal modes can be selected for up-conversion without affecting other modes.
We study sum-frequency generation (SFG) in a multimode PPKTP waveguide. We show that under proper quasi-phasematching, it can support one of the two scenarios. In the first, a single pump mode up-converts several different signal modes to different SFG modes. In the second, several different pairs of signal and pump modes are converted to the same SFG mode. By adjusting the relative phases and magnitudes of the pump modes, any superposition of the corresponding signal modes can be selected for up-conversion without affecting other modes, which can be used for spatial-mode de-multiplexing in both classical and quantum communications.
We analyze sum-frequency generation (SFG) in a χ(2) slab waveguide with the goal of achieving a single spatial-mode operation. We first develop Green’s function formalism for the SFG equations and then perform singular-value decomposition (SVD) of the Green’s function. By adjusting the spatial profile of the pump, we manipulate the SVD spectrum to maximize the up-conversion of one signal mode while minimizing the up-conversion of all others, which opens a possibility of realizing a spatial-mode-selective quantum frequency converter for future optical communications.
We propose an all-optical regeneration scheme for 16-QAM signal. An incoming signal is 50/50 split into I and Q arms. Each arm contains three phase sensitive amplifiers (PSAs); between the PSAs the signal propagates through highly nonlinear fiber (HNLF) and acquires nonlinear phase shift due to self-phase modulation (SPM). At the end, another 50/50 coupler combines the regenerated I and Q signals.
In each arm, the first PSA is used to amplify one quadrature of incoming signal and deamplify the other quadrature in order to squeeze the phase noise. Since data encoded on the deamplified quadrature is also erased by the first PSA, signal in each arm retains only the data on amplified quadrature. After the first PSA, only amplitude noises remain on the two power levels of each signal.
The amplitude noise is regenerated by the SPM in the HNLF, followed by the PSA. The SPM converts the amplitude noise to the phase noise and thus causes rotation of the uncertainty (noise) ellipse around the signal phasor in the complex plane. For a proper nonlinear phase shift, the long axis of the noise ellipse can be significantly compressed by the PSA. To suppress the noise on both signal levels with 1:9 power ratio, there are two HNLF+PSA combinations in each arm.
Our modeling shows better than 4 dB noise reduction for all constellation points. Moreover, the higher signal levels experience even greater noise suppression, which is beneficial for their nonlinear propagation through the transmission fiber.
We investigate the performance of phase-sensitive versus phase-insensitive pre-amplification in optical resolution
enhancement with a binary hypothesis test. Phase-sensitive pre-amplification is shown to outperform phaseinsensitive
pre-amplification by more than 2 dB.
We demonstrate a balanced-homodyne LADAR receiver employing a phase-sensitive amplifier (PSA) to raise the
effective photon detection efficiency (PDE) to nearly 100%. Since typical LADAR receivers suffer from losses in the
receive optical train that routinely limit overall PDE to less than 50% thus degrading SNR, PSA can provide significant
improvement through amplification with noise figure near 0 dB. Receiver inefficiencies arise from sub-unity quantum
efficiency, array fill factors, signal-local oscillator mixing efficiency (in coherent receivers), etc. The quantum-enhanced
LADAR receiver described herein is employed in target discrimination scenarios as well as in imaging applications. We
present results showing the improvement in detection performance achieved with a PSA, and discuss the performance
advantage when compared to the use of a phase-insensitive amplifier, which cannot amplify noiselessly.
Phase-sensitive amplification (PSA) can enhance the signal-to-noise ratio (SNR) of an optical measurement suffering
from detection inefficiency. Previously, we showed that this increased SNR improves LADAR-imaging
spatial resolution when infinite spatial-bandwidth PSA is employed. Here, we evaluate the resolution enhancement
for realistic, finite spatial-bandwidth amplification. PSA spatial bandwidth is characterized by numerically
calculating the input and output spatial modes and their associated phase-sensitive gains under focused-beam
pumping. We then compare the spatial resolution of a baseline homodyne-detection LADAR system with homodyne
LADAR systems that have been augmented by pre-detection PSA with infinite or finite spatial bandwidth.
The spatial resolution of each system is quantified by its ability to distinguish between the presence of 1 point
target versus 2 closely-spaced point targets when minimum error-probability decisions are made from quantum
limited measurements. At low (5-10 dB) SNR, we find that a PSA system with a 2.5kWatts pump focused to
25μm × 400μm achieves the same spatial resolution as a baseline system having 5.5 dB higher SNR. This SNR
gain is very close to the 6 dB SNR improvement possible with ideal (infinite bandwidth, infinite gain) PSA at
our simulated system detection efficiency (0.25). At higher SNRs, we have identified a novel regime in which
finite spatial-bandwidth PSA outperforms its infinite spatial-bandwidth counterpart. We show that this performance
crossover is due to the focused pump system's input-to-output spatial-mode transformation converting
the LADAR measurement statistics from homodyne to heterodyne performance.
Fully reconfigurable broadcast and select OADMs are compared to conventional designs for ULH networks. The feasibility of an 80 x10.7 Gb/s broadcast and select OADM chain with an unregenerated reach exceeding 4160 km is demonstrated. Key engineering issues for widespread commercial deployment of all-optical ULH networks include ASE noise accumulation, filter concatenation effects, dispersion, fiber non-linearity, and crosstalk impairments.
By use of novel technologies for broadcast and select (B&S) node architecture and a dispersion-managed fiber system, we demonstrate an ultra-long-haul (ULH) network system consisting of 13 optical add/drop multiplexers (OADM) in an optically transparent 4160-km network chain. All 80 channels at a bit rate of 10.7-Gbps, spaced with a 50-GHz wavelength grid, perform >13.6 dBQ, which offer a 2-dBQ optical margin when forward error correction is applied. A dynamic spectral equalizer with 40 dB extinction ratio is used for the B&S OADM, which offers simultaneous blocking and leveling functions for channel drop and power equalization. The cross-talk penalty requirement of an OADM in the ULH system is studied.
Advanced optical fibers enable high-capacity transmission for long reach systems and help to retain margin for networking. To do this they must support broadband operation, reduce nonlinear impairments, enable distributed gain and simplify networking. Here we describe fibers that incorporate dispersion management to extend span length and system reach over a broad bandwidth by compensating dispersion slope, reducing nonlinear impairments, and optimizing noise figure.