Prof Morandotti’s group performed an experiment where a pulse from a fiber laser successively passes through a wave shaper and an amplifier. A fiber-based polarizer, a 10-90° coupler, and a power meter were employed to make sure that the polarization and the average power of the pulse are always the same at the input of a 625 m long optic fiber.

With this setup they were able to generate secondary Raman solitons through interaction with tail trailing Airy beams.

Experimental observations of the dependence of the Raman soliton self-frequency shift upon the linear prechirp coefficient C , obtained by using (a) a tail-leading and (b),(c) a tail-trailing Airy pulse, respectively, for different input powers. In (d), a Gaussian-like pulse is presented for comparison. Each spectral pattern is obtained by imposing different linear chirps at the input. The white circles in (b) and (c) trace out the peak of the primary SSFS of tail-leading Airy pulses for 18- and 22-mW input powers, respectively.

Additional and somehow surprising experimental evidence shows that the primary Raman soliton can also be larger if proper (i.e. longer) Airy tails are employed, as a result of a larger energy transfer via the nonlinear interaction, that is it is possible to tune the soliton’s energy.

These results are certainly interesting from a fundamental point of view, but the most important consequences come in terms of applications, such as the improvement of the performance of fiber Raman amplifiers and the implementation of novel spectroscopy sources. Up to date, Gaussian-like input pulses are typically used to induce frequency conversion via the SSFS by employing a linear chirp.

We envisage the possibility to generate and control multicolor Raman solitons, which could lead to a useful light source for fluorescence microscopy, on top of possible telecom applications. For example, two Raman solitons can be simultaneously obtained by using the tail-trailing Airy pulse.