Examples / Nonlinear Optics / Near infrared supercontinuum in YAG

Near infrared supercontinuum in YAG

Femtosecond pulses in a bulk crystal broaden into a continuum; the red side of a 1035 nm pump in YAG is a ready near-infrared band.

Open in the canvas →

A live trace of the setup above, shown as a picture. Open it in the canvas to take it apart, retune it, and save or export it as your own.

Background

Focus intense femtosecond pulses into a transparent crystal and, above the critical power for self-focusing, the beam collapses into a filament whose self-phase modulation broadens the spectrum across hundreds of nanometres on both sides of the pump. Bulk crystals do this with no fiber to couple into and little alignment. Where the spectrum ends depends on the medium and the pump: intensity clamping ties the blue cut-off mostly to the bandgap and the material's dispersion, and the red side grows with the pump wavelength, a looser focus and a longer medium^[1].

YAG is a good match for Yb lasers. Its large nonlinear index starts a continuum at sub-µJ energies, and pumped in the near infrared it gives more infrared light than sapphire^[1]. That red side is useful on its own: Vernuccio and co-workers pumped a 10 mm YAG plate with 1035 nm, 270 fs pulses at 2 MHz, kept the red lobe with a longpass filter, and used 1050–1300 nm as the broadband Stokes of a fingerprint multiplex CARS microscope, alongside an etalon-narrowed pump, a prism compressor and a delay line^[2].

What this setup demonstrates

A 1035 nm, 270 fs, 2 MHz laser is focused by a 100 mm lens into a crystal in Supercontinuum mode with YAG as the medium, and a second lens recollimates the output. The crystal's band is estimated from the pump that arrives: no YAG reference is at 1035 nm, so each edge is interpolated between its own neighbours — the blue edge between the 800 nm and 1.1 µm references, the red edge between 800 nm and 2 µm — giving about 506–1776 nm. The 2 µm red edge was limited by the detector, which is why the crystal's Continuum readout notes that the spectrum can reach further.

A longpass dichroic at 1050 nm reflects the residual pump and the visible side of the continuum to one spectrometer and transmits the red side. A bandpass then selects 1050–1300 nm, and the second spectrometer and its screen show that band. The upper edge is an authored selection: a longpass alone would pass everything out to 1776 nm. Change the laser's wavelength or the crystal's medium and the band follows; pump fused silica at 1035 nm, outside the reference data this estimate includes, and the crystal draws no continuum and asks for a manual range.

What you won't see

The continuum is a flat band between estimated edges, not a model of filamentation or self-phase modulation. The edges are interpolated between reference spectra from single experiments and summaries with different focusing, energies, durations and crystal lengths, so they illustrate rather than predict. The 10 mm plate length is drawn for context and does not enter the estimate, and neither do the focus, pulse energy or duration; whether the pump exceeds the critical power, and the damage threshold, are not checked. The converted fraction is authored, and the spectral shape, the continuum's chirp and its pulse duration are not calculated.

The rest of the multiplex CARS bench — the pump etalon, the prism compressor for the continuum, the delay, the sample and the CARS detection — is not drawn. The specimen model draws CARS as a single line and does not evaluate a broadband Stokes spectrum against Raman resonances, so it cannot produce a multiplex CARS spectrum.

Near infrared supercontinuum in YAG

Background

What this setup demonstrates

Related components

References