Optical amplification in optical fibers by means of degenerate four-wave mixing (DFWM) is studied by classical equations. A balanced interferometer is used to mix and separate the signal and the pump. Exact solutions are obtained even with pump depletion. If the cross-phase modulation is small, the signal power gain varies as the square of the pump self-phase shift. This gain can be sizable for moderate fiber lengths and laser powers, before the onset of stimulated Raman scattering. We have experimentally verified these calculations with a Sagnac interferometer, comprising a 300 m long polarization-preserving fiber, emitting pulses with 100-200 W peak power and 100-200 nsec duration. We have measured a signal power gain of 22 (13.3 dB); this value correlated well with a pump self-phase shift of 3π/2 measured in a separate two-wavelength experiment. This system can be used for squeezing fluctuations of the signal quadrature which is not amplified. We have performed a classical derivation of squeezing, based upon a geometrical study of the transformation of the signal phasor. The results are in agreement with those of more elaborate quantum-mechanical derivations.
Optical amplification in optical fibers by means of degenerate four-wave mixing (DFWM) is studied by classical equations. A balanced interferometer is used to mix and separate the signal and the pump. Exact solutions are obtained even with pump depletion. If the cross-phase modulation is small, the signal power gain varies as the square of the pump self-phase shift. This gain can be sizable for moderate fiber lengths and laser powers, before the onset of stimulated Raman scattering. We have experimentally verified these calculations with a Sagnac interferometer, comprising a 300 m long polarization-preserving fiber, emitting pulses with 100-200 W peak power and 100-200 nsec duration. We have measured a signal power gain of 22 (13.3 dB); this value correlated well with a pump self-phase shift of 3π/2 measured in a separate two-wavelength experiment. This system can be used for squeezing fluctuations of the signal quadrature which is not amplified. We have performed a classical derivation of squeezing, based upon a geometrical study of the transformation of the signal phasor. The results are in agreement with those of more elaborate quantum-mechanical derivations.
Degenerate four-wave mixing (DFWM) in optical fibres can be used to obtain gain and to generate squeezed states, by utilizing a balanced Mach-Zehnder fibre interferometer (MZFI) with two non-linear arms, which behaves as a degenerate parametric amplifier. In a classical derivation for the optical gain; pump depletion and propagation loss are considered. Maximum achievable gain is calculated in the presence of loss. A numerical study compares the signal gain under different input pump-to-signal field ratios and its limit (undepleted pump), which is the assumption for the derivation of squeezing. Pump depletion and fibre loss hinder the validity of using large non-linear phase shifts to generate squeezed states. A semiclassical derivation provides the correct expression for the degree of squeezing in the absence of loss and pump depletion. Quantum-mechanical solutions provide the same result. The squeezing limitation due to loss is studied by a lumped-parameter model.
A nonlinear Mach-Zehnder fiber-optic interferometer as shown in Fig. 1 can be used for light squeezing1 and amplification.2 The mechanism can be viewed either as the Kerr effect,1 or degenerate four-wave mixing (DFWM).2 Using the classical DFWM approach, one finds that if fields of the same frequency with phasors E s and E p (′signal′ and ′pump′) are fed into the inputs, then one output phasor is given by where the nonlinear phase shift is given by δφnl= QL |E S E p | cosδ, Q is the fiber nonlinear coefficient, L is the fiber length, and δ is the phase difference between the inputs.
The stagnating increase in data transmission capacity in optical communication systems combined with the ever growing demand of transmission bandwidth is leading to an impending capacity crunch, referring to the point in time after which the available bandwidth of the individual user starts to decrease. To postpone this point in time, existing technologies in terms of data transmission through optical fibers must be optimized and new degrees of freedom must be introduced to continue the exponential increase in available bandwidth; space-division multiplexing is believed to be the strongest candidate for another degree of freedom in transmission fibers. This thesis is two-fold: firstly, starting at Maxwell’s equations and basic principles of quantum mechanics, a semi-classical model of the noise properties of fiber optical parametric amplifiers and frequency converters is presented. The model accounts for multiple effects present in nonlinear fibers such as four-wave mixing, Raman scattering, distributed loss, and dispersion, and it is valid in the depleted pump regime. After validating the model against well-known results of quantum models, the model is used to predict the impacts of Raman noise, loss, and pump depletion on the noise properties of parametric frequency conversion and phase-insensitive and phase-sensitive parametric amplification. An important part of realizing space-division multiplexing is the ability of optical signal processing so the second part of this thesis addresses few-mode Raman fiber amplifiers and parametric amplifiers and frequency converters. A model of weak random linear mode coupling in the pump of a two-mode distributed Raman fiber amplifier is presented and it is shown that an amplification noise figure induced by mode coupling increases with the degree of mode coupling and that this tendency increases as the pump depletes. Also, a very low mode-dependent gain of 0.25 dB per 10 dB gain is experimentally demonstrated in a two-mode distributed Raman fiber amplifier by exciting the pump in a combination of two modes. A comprehensive model of four-wave mixing in two-mode fibers acvi counting for six simultaneous processes is derived, and the conversion efficiency from signal to idler in the four-wave mixing processes of phase conjugation and Bragg scattering in two two-mode fibers with different phase matching properties are experimentally investigated. A conversion efficiency of > −2.70 dB is demonstrated for Bragg scattering in the conversion of a signal in the LP01-mode to the idler in the LP11-mode; the signal-to-idler separation is ∼ 25 nm. Good qualitative agreement between experiments and theory is found for both processes in both fibers.
Summary form only given. When a signal propagates through a transmission line, it is deteriorated by amplitude and phase noise, which reduces its signal-to-noise ratio (SNR) [1]. According to Shannon's theory, the SNR determines the maximum spectral efficiency. It is therefore necessary to limit the noise [3]. For this purpose, one can use phase sensitive amplifiers (PSA) instead of phase insensitive amplifiers (PIA) like Erbium-doped fiber amplifiers. PSAs amplify only one phase quadrature of the signal while deamplifying the other, thanks to a required phase relationship between the pump and the signal. This unique property, known as phase-squeezing, allows to obtain noise figures (NF) as low as 0 dB whereas PIAs, based on stimulated emission, cannot be go below the 3 dB quantum limit [2]. Moreover, the phase-squeezing property of PSAs can be used for quantum communication or even in metrology [1]. PSA can be realized with χ(2)material through degenerate parametric amplification [2]. The large spectral separation between the pump and the signal makes it very easy to filter out the pump after amplification. A ± 11 dB amplification with a sub-quantum NF and a regeneration of a DPSK signal was experimentally performed in LiNbO3 [3]. A fiber based approach would however be preferable for telecommunication purposes. PSA's based on four wave-mixing have been demonstrated, but the long length and the difficulty to separate pump from the signal [1] hinders its practical implementation. Here we present the first demonstration of a PSA based on periodically poled silica fiber (PPSF). This design has the advantage of all-fiber devices as well as simple pump filtering. The experimental set-up is depicted in Figure 1. A 20 mW CW fiber laser at 1550 nm is sent through a 99:1 coupler to be divided into a signal and a pump. The pump is modulated into 1 ns wide pulses with a repetition rate of 50 MHz. It is then amplified and frequency doubled in a PPLN crystal. The peak power of the 775 nm pump pulses reaches 10W. The pump and CW signal are then combined in a WDM and sent through a 30 cm long PPSF [4] where the signal undergoes parametric amplification. A second WDM is used to separate the pump from the amplified signal. The wave plates and the polarization controller in the pump path, allow to simultaneously control the polarization of the pump and the power in the fundamental mode of the PPSF (at 775 nm the fibre is slightly multimode). The relative phase between the pump and signal is controlled by fine tuning the frequency of the laser via a built-in piezo.
The four-wave-mixing-based phase-sensitive amplification of pulsed signals is analytically investigated, taking into account the pump depletion and the self- and cross-phase modulation (SPM and XPM) effects of the pumps and signals. Formulas for output phase and power are derived. It is found that, for a single-pulse signal input, the SPM/XPM of the signal will cause frequency chirping to the output; for a pulse-sequence, the amplitude-fluctuation of the input signal will destroy the phase regeneration through the SPM/XPM. In addition, the amplified signal pulse will initially be compressed and distorted, and then it will get broadened due to pump-depletion.
We present a quantum theory for generating two-photon states by means of four-wave mixing in optical fiber. We start with an interaction Hamiltonian that can correctly describe all nonlinear interactions among the four waves present in the fiber, namely, the frequency non-degenerate pumps, signal, and idler, including the terms responsible for self-phase modulation (SPM), cross-phase modulation (XPM), and four-photon scattering (FPS). The exact form of this Hamiltonian is obtained through comparison between the classical and quantum equations of motion. The two-photon state is then calculated by means of first-order perturbation theory. It turns out that only the FPM and the pump SPM terms contribute to the formation of the two-photon state. The entangled nature of this state is verified in a coincidence counting experiment. The results of the theoretical calculation agree well with experimental data.