Group velocities below the speed of light in vacuum c were known to be possible as far back as 1880, but could not be realized in a useful manner until 1991, when Stephen Harris and collaborators demonstrated electromagnetically induced transparency in trapped strontium atoms.
[3] In 1998, Danish physicist Lene Vestergaard Hau led a combined team from Harvard University and the Rowland Institute for Science which realized much lower group velocities of light.
[6][7] In 2005, IBM created a microchip that can slow light, fashioned out of fairly standard materials, potentially paving the way toward commercial adoption.
Understanding the behavior of light in a material is simplified by limiting the types of disturbances studied to sinusoidal functions of time.
The index of refraction is not a constant for a given material, but depends on temperature, pressure, and upon the frequency of the (sinusoidal) light wave.
Material dispersion mechanisms such as electromagnetically induced transparency (EIT), coherent population oscillation (CPO), and various four-wave mixing (FWM) schemes produce a rapid change in refractive index as a function of optical frequency, i.e., they modify the temporal component of a propagating wave.
Dispersion mechanisms such as photonic crystals at red and blue edges,[9] coupled resonator optical waveguides (CROW), and other micro-resonator structures[10] modify the spatial component (k-vector) of a propagating wave.
[13] A predominant figure of merit[clarification needed] of slow light schemes is the bandwidth-delay product (BDP).
Most slow light schemes can actually offer an arbitrarily long delay for a given device length (length/delay = signal velocity) at the expense of bandwidth.
Plasmon induced transparency – an analog of EIT – provides another approach based on the destructive interference between different resonance modes.