Schematic of the commercially available CPA system. Part of the oscillator is sent to the f-to-2f interferometer, the other part is stretched in a dispersive stretcher. It is then amplified in a 9-pass amplifier pumped by a frequency-doubled Nd:YLF laser. The amplified pulse is compressed using prisms. The other part of the oscillator is sent to an f-to-2f interferometer for CEP stabilization. The CEP is stabilized by changing the pump power with an AOM.
Left: Optical layout of the f-to-2f interferometer used to detect the CEP drift of the oscillator pulses. λ/2: half-wave plate, PCF: photonic crystal fiber, DM: dichroic mirror, PPLN: SHG crystal, IF: infrared filter, PBS: polarizing beam splitter, and G: grating. Right: spectrum measured before (red) and after (black) the PCF.
Optical layout of the collinear f-to-2f interferometer used to detect the slow CEP drift of the amplified pulses. ND: neutral density filter; A: aperture; L1,2,3: lens; Sapph: sapphire plate; BBO: SHG crystal; λ/2: half-wave plate; P: polarizing beam splitter; Ω: spectrometer; WLG: white light generation; and INT: interferometer.
Overview of the complete CEP stabilization system. Shown are the oscillator and amplifier with their active control elements: AOM in the oscillator pump beam, the intracavity wedges and the rear prism pair in the amplifier's compressor. The f-to-2f interferometer creates a reference voltage Uref which is used as a signal for the AOM. A PC reads out an averaged reference signal ⟨Uref⟩ and moves the wedges if needed.
A typical measurement of the CEP after the compressor over a period of almost 2 hours, which shows the retrieved CEP value (a) showing a mean value of 0.06 rad with a standard deviation of 60 mrad. Every measurement was averaged over 9 shots, the out-of-loop CEP fluctuations are therefore on the order of 180 mrad. The histogram of the CEP values magnified by a factor ten with respect to the graph in (a) is shown in (b). (c) The measured interferograms over time and the averaged spectrum (c).
Schematic of the differentially pumped hollow fiber setup. The CPA output is coupled into the hollow fiber and then re-compressed with chirped mirror pairs afterwards. The fiber entrance is under vacuum and the gas is fed into the fiber exit. A CCD camera measures the position of the focus and controls the focusing mirror to compensate for slow term drifts.
Pulse broadening in a 1 m long 250 μm inner diameter neon filled hollow fiber for the differentially pumped (a) and the statically filled case (b) over a range of pressures.
The pulse to pulse power fluctuations before (blue, top) and after (red, bottom) the hollow fiber pulse compression setup.
The focal spot after the hollow fiber.
SEA-F-SPIDER setup: BS1/2: beamsplitter, λ/2: halfwave-plate, A: aperture, FA, B: color filters, FM1/2: focusing mirror, and BBO: nonlinear crystal. Bottom right: Transmission spectra of the narrowband filters FA, B.
(a) Example SEA-F-SPIDER trace of a pulse after propagating through a 1 m differentially pumped hollow-core fiber filled with 1.0 bar of argon for a shear of Ω = 24 mrad/fs on a 40 dB color scale. (b) Spatial lineout at 404 nm showing the spatial fringes. Adapted from Ref. 84.
Spatially resolved reconstruction of a pulse out of a hollow-core fiber compressed with chirped mirrors. (a) Temporal intensity of the beam center and (b) spatially reolved reconstructed temporal intensity. Line-out on the right: spatial profile in y-direction (here vertical). (c) Temporal line-outs at −50, 0, and +50 μm. Adapted from Ref. 84.
Mach-Zehnder interferometer for the generation of two few-cycle pulse replica with adjustable time-delay, spatial shear, and spatial tilt. BS1,2: beam splitter and M1–M4: low-dispersive Ag mirrors.
An interferometer producing an annular beam and a time-delayed inner beam using mirrors that are cored at 45°. The inner beam is delayed relative to the annular beam using a delay stage. The beams are recombined using a second cored mirror.
A schematic of the attosecond beam line. A detailed description can be found in the text.
Vibration measurements on our beamline. The vibrations are measured on the chambers and on the vibrationally isolated breadboard. See text for more detail.
Schematic of the different gas targets used to produce high harmonics. (a) Continuous flow tube and (b) pulsed gas valve.
Plot of the transmission (right axis) of the Al and Zr foils (each 200 nm thick) used to spectrally select the high harmonic radiation. Also shown is the reflectivity (left axis) of the Mo:Si multi-layer mirror used for isolated attosecond pulse experiments.
Schematic of the XUV flat-field spectrometer. The variable line spaced grating spectrally disperses and focuses the harmonics which are detected using an imaging microchannel plate detector. The detector is mounted on a movable flange to enable a larger range of harmonics to be imaged (5–50 nm) while maintaining operating vacuum (see also inset of Figure 15). The grating can be translated out of the beam under vacuum to allow the spectrometer to be bypassed in the beamline.
High harmonics produced from neon with and without a 200 nm Al filter inserted. The aluminium L-edge of the filter creates a sharp absorption feature (see the Al filter transmission curve103) that can be used to calibrate the XUV spectrometer.
High harmonic spectrum for a cosine-pulse. Top: The spatially and spectrally resolved high harmonics. Bottom: Spectral lineout and MoSi mirror reflectivity. The reflectivity is scaled to the maximum reflectivity (Rmax = 31%).
Left: CEP scan of high harmonics generated in Ne with a 4 fs pulse and an intensity of 4 × 1014 W/cm2. The CEP is changed by moving a thin glass wedge. Right: Spectral lineouts for a sine and cosine pulse indicated with the arrows on the left.
TOF length calibration. Top: ATI spectrum from a 30 fs pulse centered around λ = 798 nm. Bottom: Linear fit to the energy separation of the ATI peaks. The energy spacing is ΔE = 1.55 ± 0.02 eV corresponding to central wavelength of the driving field, λ = 798 nm.
(a) Variation of the interference amplitude of the spatially overlapped focal spot center as the inner and outer pulses are delayed with respect to one another. Example interference patterns at a fixed delay are shown in (b) for constructive (delay at red dot) and (c) for destructive (delay at green dot) interference fringes.
Photoelectron energy spectra from the interaction of XUV radiation with a neon gas target where the XUV has been generated by HHG in neon at a fixed CEP of the driving field.
Above-threshold ionization photoelectron energy spectra generated by two time-delayed 30 fs IR pulses interacting with Ar gas. Constructive and deconstructive interference lead to the periodic structure of the observed photoelectron spectra as the time-delay is scanned. The peaks in the energy are a result of absorbing a multiple integer number of IR photons.
(a) Illustration of the formation of HCOs. Each peak of the laser field can induce an electron trajectory (labeled 1–4) leading to a different HCO, the energy of which depends on the field strength. (b) Spatially and energy resolved high harmonics generated in neon gas. The HCOs are visible as the spectral bands with low divergence angle due to phase matching. Adapted from Haworth et al., Nat. Phys.3, 52 (2007)10.1038/nphys463.
(a) Harmonic spectra (solid lines) generated in neon using 8.5 fs pulses with fixed CEP. The broad humps are a signature of the HCOs and are clearly seen in the averaged spectra (dashed lines). The HCOs shift in energy with CEP change as demonstrated in (b), where the solid lines are theoretical results. Plot (c) shows that the HCOs are still visible using 13 fs pulses and can even be resolved in single shot measurements with no CEP stabilization (d). Adapted from Haworth et al., Nat. Phys.3, 52 (2007)10.1038/nphys463.
Generating single attosecond pulses by spatio-spectral filtering of harmonics generated in a HCO phase-matching regime. A single attosecond pulse with a duration of 290 as (a) after spectral and spatial filtering. This is done using the a 1 mrad half-angle spatial filter on-axis as indicated in (b) and spectral filtering to filter the lower order harmonics. Using only spectral filtering shows a train of attosecond pulses (c). Adapted from Chipperfield et al., Laser Photonics Rev.4, 697 (2010)10.1002/lpor.200900028.
Top and side (inset) views of the setup. The symbols are explained in the text. Adapted from Austin et al., Opt. Lett.36, 1746 (2011)10.1364/OL.36.001746.
Processing steps. (a) Raw data, linear gray scale. The 15th and 25th harmonics are labeled for reference. (b) Horizontal DFT of raw data, logarithmic gray scale with 10−6 amplitude dynamic range, ν x is spatial frequency in x. The red lines indicate the filter passband. (c) Interferogram phase, [−π, π] rad linear gray scale. (d) Phase difference between two interferograms, [−5,5] rad linear gray scale. In (c) and (d), regions of low signal to noise ratio are set to a constant value. Adapted from Austin et al., Opt. Lett.36, 1746 (2011)10.1364/OL.36.001746.
Far-field intensity (left column) and phase (right column) of the 13th, 19th, and 25th harmonics (indicated at top left), showing experiment (blue, thinner line, uncertainty represented by shaded region) and simulation (red, thicker line). Adapted from Austin et al., Opt. Lett.36, 1746 (2011)10.1364/OL.36.001746.
Illustration of the attosecond streaking technique. An XUV and linearly polarized IR pulse (electric field E(t)) are focused on an atomic target. The XUV pulse releases a photoelectron that, depending on the IR vector potential A(t) at the given temporal delay, will be streaked up or down in energy along the electron time-of-flight direction.
Illustration of the effect that the CEP, ϕCEP, has on the production of XUV pulses. For a “cosine” pulse (ϕCEP=0) as shown in (a), a single recollision event occurs (within the selected spectral bandpass) leading to production of a single attosecond pulse (top panel). Correspondingly this pulse is streaked up in energy by the IR vector potential (lower panel). The reverse occurs for a “−cosine” pulse as shown in (c). In contrast, in panel (b) a “sine” pulse (ϕCEP=π/2) leads to a double recollision event hence a double attosecond pulse. One pulse is streaked up and the other down in energy.
(a) Streaked photoelectron energy spectra from a Ne target at fixed XUV–IR delay as the CEP of the sub 4 fs pulse generating the XUV harmonics is scanned. The dotted line indicates the energy separation used for the pulse contrast measurement, shown in (b). Figure taken from Ref. 133.
Streaked photoelectron spectrogram from a Ne target as the delay between the XUV (93 eV) and IR pulses (3.5 fs, 1×1012 W/cm2) is varied. Figure taken from Ref. 133.
(a) Background corrected attosecond streaking measurement, (b) FROG-CRAB reconstruction, and (c) the resulting temporal amplitude and phase of the attosecond pulse.
Article metrics loading...
Full text loading...