Physicists routinely exploit femtosecond laser pulses with sub-10 fs (1 fs=10-15s) pulse duration in the near-infrared spectral range, pushing limit to just one optical cycle which is only 2.7 fs at 800 nm central wavelength. Even shorter bursts of X-rays in the attosecond (1 as=10-18s) timescale are most challenges for study and control electrons motion inside atoms. In the Bohr Model, the electron orbit is in the attosecond timescale. So can time-resolved spectroscopy be extended to attosecond timescale to capture the motion of electrons in atoms? The answer of the question is followed by the generation of isolated attosecond pulses since 2001, and opens the door to attosecond spectroscopy. Attosecond is not a dream now!
The basic principle of single attosecond pulse generation is laser-driven high harmonic generation (HHG) using few-cycle pulses with carrier-envelope phase (CEP) stabilization. The interaction of intense linearly polarized ultrashort laser pulse with atoms, atom clusters, and molecules results in the generation of HHG radiation in the extreme ultraviolet (XUV) and soft- X- ray spectral range. In this picture the electron tunnels the Coulomb barrier under the laser electric field influence, and then accelerates. Subsequently it has probability to return back at the core when the electric field reverses its direction and recombine in the ground state emitting harmonic photons. The maximum energy radiated by this process is Emax=Ip+3.17Up where Ip is the ionization potential and Up the ponteromotive energy of the electron oscillating classically in the laser electric field. Few mutually coherent harmonics are enough to produce either a train of attosecond pulses or even a single attosecond pulse depending on the absolute phase of the driving laser field.
We have developed a Ti:sapphire laser amplifier which can deliver 800 uJ, 25 fs pulses at 1 kHz. By introducing the pulses into a gas filled hollow-core capillary super-continuum covering the entire visible and near-infrared was obtained. Subsequently dispersion compensation was implemented by a set of broadband chirped mirrors. At last 4.6 fs pulses with energy up to 460 uJ were obtained.
We use phase lock loop (XPS 800, MenloSystems GmbH) for fast locking the CEP of the seed oscillator at 20 MHz. CEP of the amplified pulses was detected through spectral interferometry. A small part of the amplified pulse energy was injected into a 1 mm thick sapphire plate where octave spanning spectral broadening occurs. By frequency doubling the infrared and then superposition with the fundamental blue spectrum, spectral interferometry is achieved. Fast Fourier Transform (FFT) finds the slow drift of CEP, then we feedback to the phase lock loop through LabVIEW software and PCI circuit. The root-mean-square（RMS） value of the CEP drift is better than 53 rmad.
Currently, we have built a vacuum chamber and the static pressure could achieve 10-6 mbar scale. The laser- atom interaction volume is formed by a thin- wall nickel tube, in which holes are made by the pump laser beam. This tube can be squeezed to yield an effective interaction length as short as <0.2 mm and it is continuously backed with some noble gas. After the nickel tube, filter and aperture are used to suppress the driving laser beam and the low-order harmonics. With intense laser and excellent vacuum chamber, the HHG will be generated before long.