To date, dye and Ti:sapphire laser oscillators working in concert with chirped-pulse amplifiers, hollow fibers, and broadband optical parametric oscillators (OPOs) have produced very short pulses between 2.6 and 7.8 fs at a variety of wavelengths; however, each of these pulses contains fewer than 2.0 and more than 1.3 cycles of the electromagnetic field. Recognizing that a train of ultrashort pulses in the frequency domain (a frequency comb) is extremely important for metrology applications, while single ultrashort pulses in the time domain are critical for the field of attosecond science and spectroscopy, researchers at the University of Konstanz (Konstanz, Germany) have finally succeeded in creating a 4.3 fs single-cycle light pulse at the telecommunications wavelength of 1.5 µm.1, 2
To qualify as a single-cycle pulse, the electric field in the envelope of an ultrashort laser pulse must complete just one period or less within the full width at half maximum (FWHM) of the intensity. Researchers have tried to use coherent interference between the outputs of two modelocked lasers to synthesize single-cycle pulses, but timing jitter between the two sources has prevented success. To combat this drawback, the researchers used the ultralow-noise spectra of two erbium-doped fiber amplifiers (EDFAs) seeded by the same oscillator (OSC) as input to nonlinear optical fibers, and optimized the individual pulse trains to ensure a single-cycle result.
Optimizing coherent superposition
In order for the two pulses from the femtosecond EDFAs to interfere optimally, each output is fed into a silicon prism compressor (SiPC) to provide variable dispersion (see figure). A tailored supercontinuum pulse is then generated by sending each of these outputs into a length of standard telecommunications-grade fiber followed by two different highly nonlinear fibers in each path (HNF 1, HNF 2) with tailored group-velocity dispersion and dispersion slope values. The first fiber generates a dispersive wave with a large bandwidth centered at 1125 nm, while the second fiber produces a broadband soliton centered at 1770 nm. Two different pulse compressors (PCs) with F2 and SF10 Brewster prisms, as well as a low-pass filter (LPF) and variable delay lines (VDLs) in one of the paths block spectral components above 1450 nm and below 1600 nm, isolating the 1550 nm wavelength stream.
Because timing jitter between the two light paths is less than 50 attoseconds, coherent superposition of the two beams using a dichroic beam combiner (DBC) reveals a single-cycle pulse with FWHM pulse duration of 4.3 fs at 1.5 µm. The researchers said measuring the pulse shape correctly was difficult since no standard method exists that can handle 1 nJ pulses with a bandwidth significantly exceeding an optical octave. The Konstanz team carefully characterized each component in amplitude and phase via frequency-resolved optical gating (FROG). They then linked both spectra with a linear measurement of the relative amplitude and a nonlinear fringe-resolved autocorrelation of the total pulse, which yielded the missing phase offsets.
"We have generated the shortest pulse that might be used as a carrier of digital information in the telecom frequency band," said professor Alfred Leitenstorfer at the University of Konstanz. "Up until now, few-cycle pulse generation was dominated by Ti:sapphire systems. But the fiber technology comes with many benefits, including exceptional stability and compactness."
Leitenstorfer adds, "Concerning commercialization, we have already transferred other parts of our fiber-laser technologies to industrial companies. Toptica Photonics has set up a line of femtosecond systems based on our know-how, with one of their products serving as a laser source in a confocal microscope from Carl Zeiss, which is tunable over the visible spectrum. We here at Konstanz are working toward application of our single-cycle lasers in attosecond pulse generation, optical phase control, and time-domain quantum optics with single solid-state nanosystems."
REFERENCES
- G. Krauss et al., Nature Photonics 4, p. 33 (January 2010).
- U. Morgner, Nature Photonics 4, p. 14 (January 2010).
