We describe a procedure for increasing the fractional delay (or delay-bandwidth product) of a slow-light system. A broadband input signal is sliced into several frequency bands. The light in each band propagates through a separate channel which possesses a highly reduced group velocity over a narrow frequency band. The output of each channel is then combined to form a single output field. For certain discretely distributed values of the group velocity of each channel, the output can replicate the input waveform without the need to adjust the output phase of each channel. Because this scheme makes use of many parallel channels, it can overcome a fundamental limit [D. A. B. Miller, Phys. Rev. Lett. 99, 203903 (2007)] to the delay-bandwidth product for single-channel devices. A practical design is proposed using spectral slicers and stimulated Brillouin scattering as the narrow-band slow-light process. Numerical simulation shows that such a channelized slow-light element can have nearly uniform gain and group index over very large bandwidths.
We describe a procedure for increasing the fractional delay (or delay-bandwidth product) of a slow-light system. A broadband input signal is sliced into several frequency bands. The light in each band propagates through a separate channel which possesses a highly reduced group velocity over a narrow frequency band. The output of each channel is then combined to form a single output field. For certain discretely distributed values of the group velocity of each channel, the output can replicate the input waveform without the need to adjust the output phase of each channel. Because this scheme makes use of many parallel channels, it can overcome a fundamental limit [D. A. B. Miller, Phys. Rev. Lett. 99, 203903 (2007)] to the delay-bandwidth product for single-channel devices. A practical design is proposed using spectral slicers and stimulated Brillouin scattering as the narrow-band slow-light process. Numerical simulation shows that such a channelized slow-light element can have nearly uniform gain and group index over very large bandwidths.
Over the past few years, researchers have directed a significant amount of effort towards realizing tunable all-optical devices using nonlinear optical methods. It is now possible to exercise dynamic control of the group velocity of light traveling through a wide variety of material systems. The slow and fast light refer to situations in which the group velocity íg of an optical pulse through a dispersive material can be made to be smaller and larger, respectively, than the phase velocity vp = c/n. This ability could overcome the remaining challenge in current optical networks of storing and manipulating an optical signal directly in optical domain so as to avoid a bottleneck due to optical-to-electrical (O/E) and electrical-to-optical (E/O) conversions. The overall purpose of the dissertation is to study novel slow-light systems that provide controlled generation of large pulse delays relative to the pulse width with minimal pulse shape distortion by optimally design resonance profiles of such systems. The system design studies utilize several measures of performance such as the fractional delay, power throughput, and signal distortion under the limited system resource constraints. To this end, powerful data fidelity metrics are required to quantify the performance of tunable delay devices. Here, a new framework for measuring an information velocity and throughput is described and implemented using Shannon mutual information concepts. This new technique is used to investigate trends, trade-offs, and limits in slow light devices, which are physically sensible and in good agreement with analyses obtained using a conventional eye-opening(EO) metric. Using these information-theoretic and/or conventional metrics, we present the quantifying performance of gain-based stimulated Brillouin scattering (SBS) system in optical fibers as well as optical passive devices such as Fabry-Perot, fiber Bragg gratings, and ring resonators. It is shown that combining the SBS gain medium with these passive devices can compensate their respective disadvantages and thus increase delay performance without using additional resource of SBS pump power. The results show the possibility of achieving a fractional delay up to 10 at a signal bandwidth up to tens of GHz.
The manipulation of the group velocity of optical signals (slow and fast light) has many possible applications in telecommunications, sensing, nonlinear optics, and so on. The exploitation of the effect of stimulated Brillouin scattering (SBS) is of special interest for the alteration of the group velocity in optical fibers, since it just requires standard telecom equipment and low-to-moderate optical power. However, each delay in one single, low-gain SBS-based slow-light system is accompanied by a pulse broadening. This is a result of the inherent Kramers–Kronig relations among the gain, the phase change, and the accompanied group velocity. For an ideal flat gain, the phase change is nonideal, and for an ideal phase change the gain curve leads to a broadening. Furthermore, if the gain bandwidth is adapted to the signal by broadening, the delay will be reduced. Thus the signal broadening can be reduced by several methods for one single low-gain slow-light system, but it cannot be zero. Here, we will show how a zero-broadening SBS-based slow light can be achieved with several methods. The basic idea is a reshaping of the original pulse in a second stage by saturation of the SBS system, or by the superposition of two gains. As will be shown, all of these systems show an almost ideal overall gain and phase function over the bandwidth of the pulses. Thus, SBS-based slow light with a delay bandwidth product of 1 bit and zero distortion is possible.
Slow light is light that travels at unusual, extreme group velocities—sometimes as slow as walking speed or slower. Light waves can be described by many velocities, but for a narrow-band pulse of light (a carrier frequency modulated by an envelope), the group velocity is the speed of the pulse envelope. The term slow light also encompasses other exotic group velocities due to similar techniques, including fast light, stopped light, and backwards light. The science of slow light has been established over the past several years, and research attention is now turning to potential applications of slow light. One application of slow light is as an all-optical true-time delay. Two slow light methods in optical fibers, stimulated Brillouin scattering (SBS) and dispersive delay, are used to provide a controllable pulse delay in a prototype slow-light phased-array laser radar, called SLIDAR. These slow light methods compensate the group delay mismatch of pulses of 6 ns duration while the phased array is steered in two dimensions. A phase control system is described that maintains phase lock among three signal channels and a reference channel, each containing 2.2 km of optical fiber, while accommodating the demands of the slow light techniques. Residual phase error is kept below π /5 radians (1/10 wave) RMS. Slow light can also enhance the spectral sensitivity of spectrometers and interferometers. A design for a slow-light-enhanced nanophotonic spectrometer is presented. One important use of spectrometers is to detect specific chemicals, and I describe an approach to multivariate optical computation, which can be used for automatic chemical spectrum recognition. This technique, which could one day be implemented with a slow-light-enhanced spectrometer as a single-chip chemical detection platform, is explored experimentally in the visible spectrum using a spatial light modulator.
In this paper, we investigate slow light via stimulated Brillouin scattering (SBS) in a room temperature optical fiber that is pumped by a spectrally broadened laser. Broadening the spectrum of the pump field increases the linewidth Deltaomegapof the Stokes amplifying resonance, thereby increasing the slow-light bandwidth. One physical bandwidth limitation occurs when the linewidth becomes several times larger than the Brillouin frequency shift OmegaBso that the anti-Stokes absorbing resonance substantially cancels out the Stokes amplifying resonance and, hence, the slow-light effect. We find that partial overlap of the Stokes and anti-Stokes resonances can actually lead to an enhancement of the slow-light delay-bandwidth product when Deltaomegapsime1.3OmegaB. Using this general approach, we increase the Brillouin slow-light bandwidth to over 12 GHz from its nominal linewidth of ~30 MHz obtained for monochromatic pumping. We controllably delay 75-ps-long pulses by up to 47 ps and study the data-pattern dependence of the broadband SBS slow-light system
To change the group velocity of optical signals has a lot of possible applications in telecommunications, sensing, nonlinear optics and so on. Especially the exploitation of the effect of stimulated Brillouin scattering (SBS) in optical fibers is of special interest since it just requires standard telecom equipment and low to moderate optical power. However, each delay in one single, low-gain SBS based slow-light system is accompanied by pulse broadening. This is a result of the inherent Kramers-Kronig relations between the gain, the phase-change and the accompanied group velocity. For an ideal flat gain the phase-change is non-ideal, and for an ideal phase-change the gain curve leads to a broadening. Furthermore, if the gain bandwidth is broadened in order to adapt it to the signal, the delay will be reduced. Thus, for one single low-gain slow-light system the broadening can be reduced by several methods but it cannot be zero. Here we will show how a zero-broadening SBS based slow-light system can be achieved by two different methods. The basic idea is a reshaping of the original pulse by an adapted gain in a second stage. This adaptation is achieved by the superposition of two Gaussian gain profiles or by a single saturated gain. As will be shown, these systems show an almost ideal over-all gain and phase function over the bandwidth of the pulses. Thus, SBS based slow-light with a delaybandwidth product of more than 1 bit and zero distortion is possible.