Multi-octave spectral comb from a gas-filled hollow fiber
Attoscience is the production of light pulses lasting on the order of a millionth of a billionth of a second. These pulses make it possible to measure and control physical and chemical processes extremely rapidly. The pulses are conventionally produced from an initial optical excitation using high-harmonic generation (HHG) in a gas-phase (gaseous) medium. The result is an ultrabroad, comblike optical spectrum, with mutually coherent ‘teeth’ spanning spectral regions from the visible to XUV and soft x-ray. Up to now, HHG has demanded complex, cumbersome, pulsed lasers to drive the gas-phase medium. These pump lasers emit ultrashort pulses (<10fs) with extraordinarily high instantaneous power (≥1MW). The stringent requirements on the driving lasers limited their use to a few specialized laboratories.
Recently, our team from the University of Bath collaborated with the Technical University of Denmark and the University of Oregon to demonstrate a much simpler system.1 We generated a coherent spectral comb, covering more than three octaves of frequency, using a single pump laser that emitted moderately powerful, quasi-continuous infrared pulses. The novel generation medium was a compact photonic structure consisting of an ∼1m length of a newly designed, hollow-core photonic-crystal fiber (HC-PCF), shown in Figure 1. Although prior work had found unprecedented nonlinearity at low optical powers in HC-PCF, the spectral width was limited by the modest bandgap achievable in the fibers used.
The results1 represent two advances: a new theory of mode guiding in these fibers, and harmonic generation with unprecedented low powers and long pulses. The first advance theoretically relates photonic guidance in HC-PCF to the ‘Von Neumann–Wigner’ bound and quasi-bound states within a continuum. In contrast to common intuition, the HC-PCF concurrently accommodates both discrete core-guided modes and continuum cladding modes, even if they have the same symmetry. This situation arises from the strong transverse phase mismatch between the fields in the fiber core and cladding, as shown in Figure 2. The fiber we developed has transmission bandwidths several times wider than state-of-the-art photonic-bandgap HC-PCF, with relatively low loss.
The unique properties of this fiber allowed the second development, in ultrafast optics. A coherent comblike spectrum, spanning three octaves, was generated using transient stimulated Raman scattering (SRS) in a hydrogen-filled HC-PCF (see Figure 3). Relative to previous techniques, the comb is generated with power levels six orders of magnitude lower and with pulses five orders of magnitude longer in duration. The extension of transient SRS to much longer pulses than the usual transform-limited, subpicosecond pulses was made possible by the combination of tight transverse confinement and long interaction length.2 The results were obtained in a new excitation regime for coherent SRS. Since it is seeded by quantum noise, the preparation is not adiabatic, but it is also not impulsive because the pulses are longer than the period of the Raman molecular excitation.
Both of these developments are supported by rigorous theory and modeling (see supplemental online material1). The implications are wide-ranging. The photonic guidance theory not only creates a design direction for next-generation broadband HC-PCF but can also be exploited by the whole photonic-crystal community. The similarity between this guidance and the Von Neumann–Wigner states constitutes a novel conceptual bridge between photonics and quantum mechanics. Furthermore, multiline coherent stimulated Raman scattering is a long-standing basic problem in nonlinear optical physics, and our new experiments and theory, stimulated by the availability of the new hollow-core fiber, are a large step forward in this area.
In addition, the generation of ultrabroad, coherent, comblike spectra using simple and moderately powerful lasers suggests that synthesizing attosecond pulses could be achieved more readily than they are now. The wavelengths required are also more ‘tamed’ than those currently used, which lie in the XUV and soft x-ray regions. Our techniques would then enable low-cost compact systems for subfemtosecond optical-pulse generation, with a potentially major impact in many research areas, including laser science, materials science, and biology.