The seed-source laser is a passively mode-locked fibre-based oscillator and pre-amplifier (OneFive), which seeds the Yb:doped thin-disk regenerative amplifier (Dira). Operated at a repetition price of 100 kHz, the amplifier produces an output energy of 220 W with a prepulse depth distinction of (10^-3). The collimated amplifier output has a near-Gaussian beam profile with (M^2=1.1) alongside each the most important and minor axes, and a (1/e^2) beam diameter of (sim ) 4 mm. A two-pass dielectric-grating compressor (Fig. 1) with an general effectivity of 90% yields a near-transform-limited pulse length of 860 fs. Following compression, the thermally-induced pointing drift of the beam is compensated (as much as (sim ) 1 mrad) by lively pointing stabilisation (MRC), monitored by quadrant photodiodes.
150 W of the 200-W output from the compressor is used to pump the ultimate optical parametric amplifier (OPA) stage—OPA3 (Fig. 1). The remaining (sim ) 50 W is frequency-doubled in a 1-mm thick antireflection-coated BBO crystal ((theta =23.4^circ , phi =0^circ )). The 5-W 515-nm output from the second-harmonic-generation (SHG) stage is used to pump the preliminary difference-frequency-generation (DFG) and OPA1 phases, whereas the residual (sim ) 45 W of the 1030-nm pump is used to drive the penultimate OPA2 stage. A small fraction of the compressor output is used for white-light-generation (WLG1) in a 15-mm thick YAG crystal, which is repeatedly rastered at a pace of (sim 1) mm/min to keep away from harm19,20. While seeding with WLG1 supplies passive CEP stability for the ultimate loafer output, another frequency-doubled WLG2 stage is utilised to keep up passive CEP stability for the ultimate sign output (additional particulars later).
The white-light-seeded and 515-nm-pumped collinear DFG stage contains an antireflection-coated 3.7-mm thick LBO crystal ((theta =90^circ , phi =13.8^circ )). The (sim ) 0.5-W near-infrared sign is rejected, whereas the (sim ) 0.2-W loafer (1.4–1.85 (upmu )m) is propagated to seed the next OPA phases, following pulse-shaping in a lithium-niobate (LN) acousto-optic programmable dispersive filter (Dazzler, Fastlite). The heartbeat-shaping introduces linear and higher-order dispersion to optimise the temporal overlap of the seed with the pump within the subsequent OPA phases and permits fine-tuning of the ultimate output compression. The (sim ) 10% transmission effectivity of the Dazzler supplies (sim ) 20 mW of formed seed to OPA1.
The collinear OPA1 stage contains an antireflection-coated 2.8-mm thick LBO crystal ((theta =90^circ , phi =13.8^circ )). It’s seeded by the pulse-shaped loafer from the DFG and pushed by the residual 515-nm pump from the DFG. OPA1 amplifies the seed to a (sim ) 0.2-W sign (at 1.4-1.85 (upmu )m), which then seeds OPA2, whereas the loafer is rejected. The non-collinear OPA2 stage is predicated on an antireflection-coated 5-mol% Mg-doped periodically-poled fan-out LN crystal with a quasi-phase-matching interval of 24-32 (upmu )m21. It’s pumped by the unconverted (sim ) 45 W of 1030-nm gentle from the SHG stage. The 5-W sign from OPA2 seeds the ultimate collinear OPA3 stage. OPA3 contains an antireflection-coated 1.5-mm thick low-defect MgO:LN crystal ((theta =45.4^circ , phi =30^circ )) at a temperature of 120(^circ )C and is pumped by 150 W of the 1030-nm output from the grating compressor22. The pump depth in OPA2 and OPA3 are maintained at (sim 70) GW/cm(^2).
Tunable infrared spectroscopy mode. (a) Self-normalised sign and loafer spectra. (b) Output energy and measured pulse durations. (c) Spectral stability of the sign at 1.55 (upmu )m. (d) Energy stability of the sign at 1.55 (upmu )m and the corresponding loafer at 3.1 (upmu )m, with rms fluctuations of 0.8% and 1.5%, respectively, over 10 h. (e) Shot-to-shot stability, with rms fluctuations of 0.5% and 1.4% for the sign and loafer, respectively, over 2 ms.
The sign from OPA3 is compressed by two totally different units of chirped mirrors to be able to cowl all the 1.4–1.85 (upmu )m spectral vary of the sign. The corresponding loafer is compressed by anti-reflection-coated bulk silicon home windows of various thicknesses and a pair of adjustable silicon wedges. The spectral tuning is carried out fully with the assistance of motorised controls.
The OPCPA generates as much as 23 W of < 100-fs sign ((sim ) 50 fs at 1.75 (upmu )m) and 13 W of < 200-fs loafer, spanning the spectral ranges 1.4-1.8 (upmu )m and a pair of.3-3.9(upmu )m, respectively (Fig. 2). The spectra (Fig. 2a) are measured concurrently with a 0.9–2.5 (upmu )m spectrometer (NIRQuest, Ocean Optics) and a 1–5 (upmu )m multi-octave infrared spectrum analyser (Mozza, Fastlite). The ability dip at (sim ) 3 (upmu )m (Fig. 2b) might be attributed to absorption bands in each lithium niobate and the ambient water-vapour outdoors the OPCPA unit, which is purged with dry air. The heartbeat durations (Fig. 2b) are measured by an SHG-based frequency-resolved optical gating (FROG)23 system (Frozzer, Fastlite), coupled to a spectrometer (Avantes or NIRQuest). The spectral stability of the sign can also be illustrated (Fig. 2c). The basis-mean-squared (rms) energy fluctuations, measured over a time interval of 10 h with thermopile powermeters, are 0.8% for the 1.55-(upmu )m sign and 1.5% for the corresponding 3.1-(upmu )m loafer (Fig. second). The OPCPA is supplied with an in-built fibre-coupled shot-to-shot power monitor (Massive Brozzer, Fastlite), which reveals rms fluctuations of 0.5% and 1.4% for the sign and loafer, respectively, measured over 0.2 ms (200 pictures). That is independently corroborated by measuring the shot-to-shot noise of the sign and loafer scatter with extended-InGaAs and MCT detectors (Fig. 2e). For comparability, the pump (Dira) reveals an rms fluctuation of (<0.1)%, measured at full energy (226 W) over 3 h, and a shot-to-shot rms fluctuation of 0.4% over (10^3), (10^4)and (10^5) pictures. The shot-to-shot stability is pivotal in discerning weak spectroscopic alerts amidst the background noise – as an example, in two-dimensional infrared (2DIR) spectroscopy24.
Tunable Raman spectroscopy mode. Self-normalised sign and loafer spectra with as much as 21 W and 10 W energy, respectively, and (sim ) 50 cm(^-1) spectral bandwidth.
The Dazzler-based pulse-shaping facilitates handy frequency filtering of the seed. Consequently, the broadband femtosecond pulses can readily be transformed to narrow-band ((sim ) 50 cm(^-1)) pulses with as much as 21 W and 10 W of energy for the sign and loafer, respectively (Fig. 3). The rise in spectral brightness is of profit in time-resolved Raman spectroscopy. While a better spectral decision (sometimes (<20) cm(^-1)) is preferable for condensed-phase Raman experiments, additional spectral filtering of the seed results in discount in energy stability.
Actively CEP-stabilised HHG modes. (a) Spectra of the 1.75-(upmu )m and the two.8-(upmu )m modes. (b) Close to-transform-limited pulse durations (FWHM) of 49 fs and 62 fs, respectively, measured with an SHG-FROG. (c) Lively CEP stabilisation utilizing an f–2f interferometer, able to single-shot detection (at 10 kHz), and the Dazzler to supply international CEP offsets with rms fluctuations of 86 mrad and 73 mrad over practically 10 h.
The Dazzler-based pulse-shaping might be utilised to generate 19 W of 49-fs sign at 1.75 (upmu )m and 5 W of 62-fs loafer pulses at 2.8 (upmu )m (Fig. 4a,b). The shorter pulse durations are extra amenable to HHG and are achieved utilizing chirped-mirror and bulk-silicon compressors. The long-term energy stability is measured to be 1.2% and a pair of.3% over > 8 h for the broadband 1.75-(upmu )m and the two.8-(upmu )m modes, respectively. Thermal distortions within the propagation of the two.8-(upmu )m beam on account of water-vapour absorption is minimised by purging the OPCPA with dry air. An (M^2) measurement (Fig. 5) for the 1.75-(upmu )m output reveals values of 1.3 and 1.4 alongside the most important and minor axes, respectively. 10% of the sign and loafer outputs are directed in direction of two totally different f–2f interferometers (Fringeezz, Fastlite)25that are able to single-shot detection (at 10 kHz) and supply lively suggestions to the Dazzler for CEP stabilisation26. The worldwide CEP offset for the 1.75-(upmu )m and the two.8-(upmu )m modes, measured over 10 h, present rms fluctuations of 86 mrad and 73 mrad, respectively (Fig. 4c). Notably, while the two.8-(upmu )m HHG mode makes use of the fundamental-driven WLG1 seed, the 1.75-(upmu )m HHG mode makes use of a frequency-doubled white-light stage (WLG2) to be able to keep passive CEP stability. It’s because the two.8-(upmu )m output originates in OPA3, which is pumped at 1030 nm, and is seeded by the amplified gentle from WLG1, additionally pumped at 1030 nm. Quite the opposite, the 1.75-(upmu )m output originates within the DFG stage, which is pumped at 515 nm. Consequently, for passive CEP stability, it’s seeded with WLG2, comprising frequency-doubled white gentle pushed at 1030 nm (since white gentle pushed by a frequency-doubled pump at 515 nm was discovered to be comparatively much less secure). The CEP stability will change into necessary following post-compression of the OPCPA output to few-cycle pulses27which is deliberate for the close to future.
Beam profile. M(^2) measurement of the HHG-mode sign at 1750 nm, with values of 1.3 and 1.4 alongside the most important and minor axes, respectively. (The plots are vertically offset for readability). The inset reveals the unfocussed beam profile.
