Low-distortion slow light using two absorption resonances

Abstract

We consider group delay and broadening using two strongly absorbing and widely spaced resonances. We derive relations which show that very large pulse bandwidths coupled with large group delays and small broadening can be achieved. Unlike single resonance systems the dispersive broadening dominates the absorptive broadening which leads to a dramatic increase in the possible group delay. We show that the double resonance systems are excellent candidates for realizing all-optical delay lines. We report on an experiment which achieved up to 50 pulse delays with 40% broadening.A variety of applications in telecommunications and quantum information have been driving recent interest in slow group velocities of light pulses. Among these applications are continuously tunable delay lines, all-optical buffers [1], optical pattern correlation, ultra-strong crossphase modulation [2], low light level nonlinear optics [3,4,5], and numerous others. The means for obtaining ultra-slow group velocities have usually involved a Lorentzian transparency or gain resonance: electromagnetically induced transparency (EIT) [6,7,8,9,10], coherent population oscillations (CPO) [11,12,13], stimulated Brillouin scattering (SBS) [14,15,16], stimulated Raman scattering (SRS) [17,18] etc..In this paper we discuss delaying pulses whose center frequency lies between two strongly absorbing resonances. Many researchers have considered using gain doublets in the context of pulse advancement [19,20,21,22,23], and Macke and Segard [22] have discussed pulse advancements for absorptive doublets. Grischkowsky [24] measured the delay of a pulse between two Zeemanshifted absorbing resonances, and Tanaka et al. [25] performed initial measurements of the delay of a pulse between two atomic hyperfine resonances. This work considers both delay and broadening with an emphasis on the suitability of the delay and broadening characteristics for practical applications.In the context of optical delay lines, several criteria must be satisfied for slow light to be useful. First, the slowed light pulses must meet system bandwidth specifications. Second, the delay-bandwidth product must be much larger than unity. Third, the delay re-configuration rate should be faster than the inverse pulse propagation time through the medium. Fourth, pulse absorption must be kept to a minimum. Fifth, the pulse broadening should also be minimal. The exact requirements for practical optical buffers are application dependant. A typical system operating at 10 Gb/sec with return-tozero coding and a 50 % duty cycle might require 7.5 GHz signal bandwidth, a delay bandwidth product of 1000, a re-configuration rate in excess of 7.5 MHz with less than 90% absorption and pulse broadening of less than 2.Despite widespread interest in large pulse delays, simultaneously satisfying all five criteria for most applications has proven difficult. In this paper we show that double Lorentzian systems manages four of these criteria well: large bandwidth, large delay bandwidth product, minimal absorption and ...

Ultraslow Propagation of Matched Pulses by Four-Wave Mixing in an Atomic Vapor

“…They rely either on a reduction of the absorption, such as electromagnetically induced transparency (EIT) [1], coherent population oscillations [2], and dual absorption lines [3], or on a gain resonance, like stimulated Brillouin scattering [4] and stimulated Raman scattering [5]. To be useful in the context of all-optical signal processing, an optical delay line should be able to produce a fractional delay (defined as the ratio of the delay to the duration of the pulse) larger than unity with only modest absorption and pulse broadening.…”

Raman-induced slow-light delay of THz-bandwidth pulses

“…For example, cascaded Bragg gratings [9], Moiré gratings [10,11], ring resonators [12], and photonic crystal structures [13][14][15][16][17] have all been used to demonstrate slow light. However, such structures offer limited tunability and lack continuous optical control when compared with optical absorption and gain resonance techniques.Slow light has been achieved for pulses tuned between absorbing resonances in an atomic vapour [18][19][20][21][22]. In one demonstration, pulses as short as 275 ps were delayed by up to 6.8 ps [19].…”

“…Slow light has been achieved for pulses tuned between absorbing resonances in an atomic vapour [18][19][20][21][22]. In one demonstration, pulses as short as 275 ps were delayed by up to 6.8 ps [19].…”

“…Dipole transitions between the three lower states and the upper manifold of states |m are permitted, resulting in Raman coupling of levels |1 → |2 and |1 → |3 when the control and signal pulses are applied. The gradient in linear dispersion between two absorption lines can be used to reduce or increase the group velocity of light [18,20]. Here, this is achieved for the signal field by setting (ω sig − ω c ) = ω 2 − ω 1 + ∆/2 = ω 3 − ω 1 − ∆/2, where ∆ is the energy splitting of levels |2 and |3 , and ω k is the energy of level |k .…”

Slow light in flight imaging

“…Slow light has been proposed and studied as a resource for building quantum networks; carefully controlled transmission of quantum states of light [1,2], tuneable delay lines [3,4] with fast reconfiguration rates [5] and spectral resolution [6], and tuneable delay of entangled images [7] are just a few examples. More recently slow light has been exploited to enhance measurement, either through improving the sensitivity of interferometers [8], or by enhancing effects that would normally be too small to observe, such as rotary photon drag [9] and Fresnel light dragging [10,11], or the proposed observation of gravitational deflection of classical light [12].…”