Strong-field coherent control of isolated attosecond pulse generation

Attosecond pulses generated via high-order harmonic generation (HHG) have enabled the study of ultrafast electron dynamics in atoms, molecules, and solids on sub-femtosecond timescales. HHG-based attosecond sources are widely used due to their laboratory-scale implementation and intrinsic synchronization with the driving laser. However, current sources face limitations in photon flux and spectral tunability, which hinder measurements of low cross-section processes and selective excitation of specific transitions. Traditional approaches to isolate attosecond pulses rely on gating techniques or sub-cycle drivers but often sacrifice flux or tunability. Because HHG dynamics depend sensitively on the driving electric field, shaping the optical waveform offers a route to control ionization, electron excursion, and recombination, potentially enhancing efficiency and tunability. The research question addressed here is whether coherent, sub-cycle control of the driving waveform—specifically via the relative phase between spectral channels and the overall carrier-envelope phase—can deterministically tune the central photon energy, bandwidth/shape, and duration of isolated attosecond pulses, while maintaining high yield and avoiding additional gating or filtering.

The paper situates its work within three decades of HHG and attosecond science, citing foundational demonstrations of HHG and subsequent applications in time-resolved spectroscopy and electron dynamics. It contrasts HHG with free-electron lasers (FELs), noting HHG’s synchronization advantages and lower fluctuations. Traditional isolation of attosecond pulses has used gating techniques (e.g., double optical gating, ionization gating, attosecond lighthouse) or drivers with sub-cycle duration. Prior theoretical and experimental studies indicated that specific multicolor or synthesized waveforms can enhance HHG conversion efficiency and extend cutoff energies. However, experimental control of isolated attosecond pulse spectral shape/bandwidth and duration directly via sub-cycle waveform synthesis (controlling relative phase between spectral bands and CEP) had remained largely unexplored. Previous work tuned central energy via macroscopic parameters (gas pressure, target position, pulse energy) but not via coherent waveform control at the single-cycle level.

Experimental platform: A parametric waveform synthesizer (PWS) based on optical parametric amplifiers pumped by a cryogenic Ti:Sapphire amplifier generates and coherently combines two spectral channels: a near-infrared (NIR) channel (650–1000 nm, ~6 fs FWHM, up to 0.15 mJ) and an infrared (IR) channel (1200–2200 nm, ~8 fs FWHM, up to 0.6 mJ). The combined, 1.7-octave-spanning waveform is shaped via two parameters: the relative phase (RP, φ) between channels and the overall carrier-envelope phase (CEP, ψ). The RP can be locked/controlled over ~200 rad (~100 fs) with <0.1 rad rms residual noise; the CEP can be varied over ~10 rad with <0.25 rad rms noise. For φ≈0 (temporal overlap), the synthesized FWHM duration is <3 fs (sub-cycle). Multi-phase metrology and feedback stabilize (φ, ψ) over hours.

Attosecond beamline: The synthesized pulse is split; the higher-energy replica drives HHG, the other serves as the streaking field. At the HHG interaction point, channel energies are ~70 µJ (NIR) and ~170 µJ (IR). The IR beam is expanded to match focal waist with NIR. Focusing with f=375 mm spherical mirror into a ~2 mm gas cell yields peak intensity 2–3×10^15 W/cm² (for φ=0, ψ=0). The most intense half-cycle period varies between ~1.3–3.2 fs depending on (φ, ψ). HHG spectra are recorded with a McPherson 251MX grating spectrometer and Andor Newton 940 CCD. XUV is separated via metallic filters and refocused by an Au-coated toroidal mirror into a Kaesdorf ETF11 electron TOF spectrometer for streaking.

Waveform scan and HHG measurements: In argon, macroscopic conditions: ~300 mbar, gas cell a few mm after focus. RP is scanned from negative (IR leading) to positive (IR trailing) values in 0.8 rad steps; for each RP position, CEP is linearly ramped over ~10 rad with 4 s period, repeated twice. Data are binned to construct a 100×16 (RP×CEP) matrix with 3–10 spectra per bin, enabling visualization of cuts at fixed RP or CEP. A similar scan is performed in neon (same pressure, gas target position optimized closer to focus). Integration times: argon (not specified numerically here), neon 200 ms, with 300 nm Zr filter.

Attosecond streaking and retrieval: For selected (φ, ψ) settings producing representative continua, attosecond streaking spectrograms are recorded by scanning delay between XUV and streaking field. Retrieval uses the Volkov Transform Generalized Projection Algorithm (VTGPA), which does not rely on the central-momentum approximation and tolerates complex streaking fields and broadband XUV. Multiple acquisitions improve repeatability; total acquisition per set can take up to ~2 hours. Retrieved temporal/spectral intensities and phases are compared against independent XUV spectra for consistency.

Simulations: Single-atom HHG response using the Lewenstein model (short trajectories only) implemented via HHGMax. Input waveforms are constructed from experimentally characterized NIR and IR pulses measured by 2D spectral shearing interferometry (2DSI). RP is scanned between −11 and 11 rad at fixed CEP (π/2 in one example). Linear 1D propagation through a 2 mm, 300 mbar neutral gas column is included. Classical electron trajectory analysis correlates waveform half-cycle structure with ionization/recombination dynamics and resultant XUV continua (narrowband, broadband, intermediate bandwidth). Peak synthesized intensity in simulations normalized to 2.5×10^14 W/cm² for RP=0, CEP=0 (example condition).

• Coherent waveform control using RP (φ) and CEP (ψ) enables deterministic tuning of isolated attosecond pulse (IAP) spectra and durations without additional gating or spectral filtering.
• In argon, attosecond continua spanning 35–110 eV are generated across broad RP/CEP settings; central photon energy, bandwidth, and shape are tunable. Examples: φ≈−28 rad yields a narrowband continuum (~5 eV FWHM) centered at ~42 eV; φ≈−16 rad yields ~15 eV bandwidth centered at ~60 eV; φ≈0 rad yields a broadband continuum spanning ~36–101 eV (≈1.5 octaves) peaking at ~75 eV.
• Highest cutoff near φ≈0 does not coincide with highest yield (max yield observed around φ≈−5 rad), indicating waveform shaping can enhance conversion efficiency beyond simple pulse compression.
• Measured IAP energies: up to ~500 pJ for the highest-energy IAPs in argon; broadband IAPs around ~100 pJ.
• CEP strongly modifies the continua landscape at fixed RP: depending on waveform, CEP changes can reduce overall intensity, shift central photon energy, or broaden spectra into double-hump structures, evidencing coherent control of electron trajectories.
• In neon, tunability extends the continuum into the soft X-ray region up to ~200 eV, with clear CEP dependence confirming IAP generation at higher photon energies. Integration time per spectrum: 200 ms (with 300 nm Zr filter), indicating sufficiently strong HHG signal.
• Attosecond streaking reconstructions confirm continua correspond to isolated pulses and reveal tunable durations: in argon, IAPs of ~240 as and ~80 as (FTL durations ~240 as and ~77 as); in neon, durations decrease from ~140 as (FTL ~137 as) to ~80 as (FTL ~76 as) for different (φ, ψ).
• Synthesized waveforms feature dominant half-cycles with periods ~1.3–3.2 fs; sub-cycle FWHM durations <3 fs achievable at φ≈0. Peak intensities in the HHG gas cell are ~2–3×10^15 W/cm² (for φ=0, ψ=0).
• Dataset acquisition enabled a 4D map (photon energy, spectral intensity, RP, CEP) summarized as 100×16 RP/CEP responses with multiple spectra per bin, allowing robust identification of tunable IAP regimes.
• Simulations reproduce key qualitative features: asymmetric RP dependence, alternating bright continua and low-yield regions, broadband vs. narrowband continua at specific RP values, and minimal spectral fringes except at low photon energies for select RPs. Classical analysis links specific half-cycle ionization/acceleration sequences to narrowband (lower cutoff, longer IAP) and broadband (higher cutoff, shorter IAP) outcomes.

The study demonstrates that sub-cycle shaping of the HHG driving field via parametric waveform synthesis provides direct, coherent control over the ionization, acceleration, and recombination steps of the three-step HHG process. By adjusting the relative phase between NIR and IR channels and the global CEP, the temporal asymmetry and amplitude of successive half-cycles are tuned to favor or suppress specific electron trajectories. This enables isolation of single recollision events and deterministic control of the XUV continuum’s central energy, bandwidth, and temporal duration. The findings address the long-standing need for spectrally tunable, high-flux attosecond sources without resorting to lossy monochromators or gating schemes that constrain bandwidth. The observed decoupling of cutoff energy and yield underscores that waveform engineering can optimize both tunability and efficiency. Agreement between attosecond streaking reconstructions, independently measured XUV spectra, and qualitative simulations validates the physical picture. Extending tunable continua to ~200 eV in neon illustrates the approach’s versatility and potential for targeting core-level dynamics. Overall, coherent strong-field control through waveform synthesis expands the experimental parameter space for attosecond science, enabling selective excitation and timing of ultrafast electronic processes.

By synthesizing multi-octave waveforms and controlling their relative phase and CEP, the authors achieve unprecedented spectral and temporal tunability of isolated attosecond pulses. Under comparable macroscopic HHG conditions, they generate continua across 30–110 eV in argon and up to ~200 eV in neon, and tune IAP durations from ~240 as down to ~80 as, all without additional gating or filtering. Simulations and classical trajectory analyses connect specific waveform half-cycles to emission properties, confirming coherent control of free-electron trajectories. This capability promises higher selectivity and broader applicability of HHG-based attosecond sources for probing low cross-section and energy-specific processes. Future work could incorporate additional spectral channels for finer control, explore other gases and phase-matching conditions to extend into deeper soft X-ray regimes, develop feedback optimization of (φ, ψ) for target spectra/durations, and implement full macroscopic propagation models to optimize flux and beam quality.

• Characterization via attosecond streaking was performed for selected waveforms due to acquisition time (up to ~2 hours per set), limiting exhaustive temporal validation across the full RP/CEP space.
• Simulations focus on single-atom response with short trajectories and simplified 1D linear propagation through the gas target; full macroscopic effects (phase matching, absorption, plasma dynamics, spatiotemporal coupling) are not comprehensively modeled.
• Experiments were conducted in argon and neon under similar macroscopic conditions; generalization to other media, pressures, and geometries remains to be shown.
• Waveform synthesis employed two spectral channels; while highly effective, further flexibility (e.g., adding a visible channel under construction) may be needed to fully optimize efficiency and extend tunability.
• The study emphasizes tunability and isolation; a systematic, quantitative comparison of photon flux enhancements versus state-of-the-art gating methods was not the primary focus.