Effectivity-optimized relativistic plasma harmonics for excessive fields

Experimental particulars

The experimental outcomes introduced on this paper have been obtained utilizing the Gemini laser system. A DPM system is used to enhance the laser distinction to I_max/I(t) > 10⁸ at instances greater than 1 ps earlier than the height of the heart beat, whereas the sub-ps distinction is mentioned within the textual content referring to Fig. 1. A complete throughput of fifty% was measured, which ends up in on-target pulse energies of 5 J within the 50 ± 5 fs period pulses with λ_L = 800 nm, that are targeted by an f/2 parabola onto a refined fused silica goal.

Pulses with energies of as much as 12 J (earlier than DPM system) in 50 ± 5 fs at a central wavelength of 800 nm have been used, which, when targeted to a FWHM spot measurement of two μm, attain peak intensities I > 10²¹ W cm⁻². As proven in Fig. 1a, these have been targeted onto optical-grade fused silica targets in p-polarization at an incidence angle of 45°, and the spectrum of utmost ultraviolet radiation emitted within the route of specular reflection was recorded.

The on-target depth was assorted by apodizing the beam, which each reduces the laser pulse vitality and will increase the focal spot measurement however maintains the identical near-field depth in order that the DPM response and distinction are unchanged. The mirrored harmonic beam was detected utilizing a cylindrically curved XUV flat-field spectrometer consisting of a 300 traces per mm grating imaging the supply within the spectral dimension. Aluminium filters with thicknesses starting from 0.2 μm to three μm have been used to attenuate optical emission. No focusing optic was used in order that the XUV sign is incident immediately onto the charge-coupled gadget (CCD). The harmonic spectra have been detected utilizing a back-thinned ANDOR CCD (Andor DV436) with a resolution-limited pixel measurement of 13.5 μm positioned 1.2 m from the interplay level.

The plasma density gradient was managed by a 25 mm diameter, 3 mm thick fused silica substrate with an anti-reflection coating on the entrance facet and excessive reflectivity on the rear. This pick-off mirror was inserted into the principle beam line in entrance of the final mirror earlier than the parabola. This launched a prepulse, which is concentrated by the identical parabola as the principle beam onto the goal however with a decrease depth due to the bigger focal spot measurement. Precise adjustment of the gap between the substrate and the full-beam mirror allowed for prepulse timing to be managed to inside 25 fs of the principle pulse. This fantastic timing management enabled correct tailoring of plasma growth earlier than the arrival of the driving pulse.

Laser distinction enhancement was achieved by way of measurements that decided that the anti-reflection coating on the primary plasma mirror (PM) of the DPM was breaking down too early within the rise time of the native (unaltered) Gemini pulse distinction. This resulted within the ‘slow rise time’ DPM configuration (t_HDR = 711 ± 25 fs; Fig. 1b, purple hint). By changing the primary PM with an uncoated substrate, we improved the DPM efficiency to have a t_HDR = 351 ± 25 fs (Fig. 1b, blue hint). The motivation for making this modification comes from a separate department of examine on ultrafast supplies science that focuses on the half performed by supplies which might be extremely structured on the nanoscale within the lifetime of excited electrons earlier than materials breakdown^45,46. Note that for various peak intensities, t_HDR will describe completely different absolute intensities, whereas the ratio stays the identical. This ought to, subsequently, be thought-about rigorously within the context of a given peak depth interplay.

Harmonic vitality deconvolution

To acquire the general effectivity, the spectral response as a perform of wavelength, λ, of all parts of the spectrometer have been accounted for individually as indicated by

the place Al is the aluminium filter transmission⁴⁷ (0.2–3 μm), QE is the quantum effectivity of the back-thinned Andor DV436 (ref. ⁴⁸), G is the calculated effectivity of the SHIMADZU-L0300-20-80 flat-field grating⁴⁹, Al₂O₃ is the contaminant aluminium oxide layer current on the filters (see aluminium oxide layer calibration under) and CH is the hydrocarbon contaminant layers⁴⁷.

The vitality per depend is calculated as

the place G = 2 is the variety of electrons per depend (e⁻ per depend), ε is the common vitality required to supply an electron–gap pair in silicon, 3.65 eV, q_e is the digital cost and ρ_frac is the fraction of the beam incident onto the CCD. This is restricted by the angle of acceptance of the flat subject grating of three.66 mm and the digital camera chip width of 27.6 mm. We assume a two-dimensional (2D) Gaussian beam propagating alongside the z-axis, with its transverse depth profile described by

the place (x₀, y₀) is the centre of the beam and σ_x and σ_y are the usual deviations of the beam alongside the x– and y-directions, respectively (measured values of beam divergence are used). We have to compute the fraction of the entire beam that falls inside an oblong aperture outlined by

This fraction is given by the ratio

The complete Gaussian beam over all area is (2{rm{pi }}{sigma }_{x}{sigma }_{y}) thus,

By measuring the beam divergence (Fig. 3), we then calculate the related fraction ρ_frac of the beam that was incident onto the CCD. The uncertainty within the vitality deconvolution is dominated by systematic uncertainty within the thickness of floor oxide (Al₂O₃), carbon contamination layers (CH) and the beam divergence measurement (ρ_frac). The correction is estimated by evaluating it at two excessive limits after which taking half the distinction between the ensuing corrected values as a symmetric uncertainty certain.

Aluminium oxide layer calibration

A 2.15 μm thick aluminium foil, which was used as a filter for the harmonic sign, was imaged on the Ewald Microscopy Centre, Queen’s University Belfast, to find out the thickness of the oxide containment layers. A cross-sectional lamellae was ready utilizing a Tescan targeted ion beam scanning electron microscope (SEM) Lyra3, which might be seen in Extended Data Fig. 3a. High-angle annular dark-field imaging was carried out, which photographs the entrance floor of the aluminium foil. Moreover, energy-dispersive X-ray (EDX) mappings have been carried out on this part of the lamellae, exhibiting a distinction within the layer construction. Elemental mapping was used throughout the foil to determine the oxygen area similar to the aluminium oxide shaped on the floor. The thickness of the oxide layers was noticed to differ between 7 nm and 16 nm as seen in Extended Data Fig. 4b. This is in keeping with the native oxide sometimes shaped on aluminium uncovered to ambient situations. However, it differs from different research that measured the oxide layer thickness as 8 nm in complete, not simply the entrance floor thickness as measured right here⁵⁰. This variation of oxide layers has been included within the experimental uncertainties within the introduced vitality calculations. The want for gold sputtering of the pattern floor and the low sign yield from gentle components make it tough to measure the hydrocarbon contaminant layer on the pattern; we confer with ref. ⁵⁰ for this worth.

Grating second and third orders

Using a diffraction grating to picture XUV harmonics, the looks of options at what appear to be ‘half-order’ or ‘third-order’ harmonic positions is attributed to the upper diffraction orders of the grating showing with the first-order sign. In Extended Data Fig. 5a, the lower-order harmonic area of the XUV spectrum between 47 nm and 45 nm, there are obvious ‘half-order’ harmonics noticed. However, when evaluating these options to the higher-order harmonics proven in Extended Data Fig. 5b between 20 nm and 22.5 nm, it may be seen within the magnified area of the plots that the half harmonic order that’s seen to be at 19.fifth order is definitely the Thirty ninth-order harmonic being imaged in second order due to the matching spatial and spectral form of those two photographs.

This highlights the significance of accounting for second- and even third-order diffraction contributions when analysing and calculating the entire vitality distribution within the lower-order spectral area. Using an aluminium filter permits transmission of sure XUV wavelength ranges whereas blocking others. By selecting a filter that transmits solely the specified harmonic vary and blocks shorter wavelengths, we are able to suppress undesirable second- or third-order contributions. However, utilizing a 300 l mm⁻¹ grating within the wavelength vary of 80–20 nm, indicators near the aluminium L-edge cut-off can seem within the second and third order with the lower-order harmonics.

In Extended Data Fig. 6a, the ratio of the primary and second diffraction orders of the measured SHIMADZU-LO300 flat-field grating is proven⁴⁹. This was measured utilizing the laser-driven excessive harmonics that have been incident onto a pinhole. It is seen in Extended Data Fig. 6a that the second-order effectivity approaches the first-order effectivity at larger frequencies. In Extended Data Fig. 6b, the ratio of the primary and third diffraction orders are proven, which has an analogous development in approaching the primary order when shifting in the direction of larger frequencies.

A logistic perform was fitted to the information

the place, for the second diffraction order, v = 0.92, n₀ = 40, ok = 0.37 and b = 0.12 and for the third diffraction order, v = 0.79, n₀ = 44, ok = 0.44 and b = 0.13. Failing to account for these results results in overestimating the mirrored harmonic vitality within the decrease orders, for which the aluminium filter attenuates the spectrum significantly. The development given by equation (6) was used to account for the second- and third-order diffraction contributions. Owing to this overlap being unavoidable, utilizing the grating response, we are able to deconvolve the overlapping orders from the spectrum.

Numerical simulations

The SHHG interplay was modelled utilizing high-resolution 2D simulations carried out with the PIC code Smilei⁵¹. A spatial decision of 512 cells per laser pulse wavelength and a temporal decision of 1,024 timesteps per laser pulse cycle have been used, enabling the decision of harmonics past the aluminium L-edge (17 nm) (ref. ²⁸). The p-polarized laser pulse is incident on a completely ionized SiO₂ goal of quantity density 6.62 × 10²³ cm⁻³ with an exponential preplasma ramp. Optimal preplasma scale lengths for XUV SHHG efficiencies have been decided from one-dimensional (1D) PIC simulations and ranged from 0.12λ_L to 0.16λ_L. XUV SHHG efficiencies are sometimes optimized for scale lengths within the vary of 0.1–0.2λ_L (refs. ^23,37,52). Particles are initialized with 100 macro-electrons per cell at a temperature of 115 eV and 4 macro-ions per species per cell at zero temperature. The laser pulse is modelled as a spatiotemporal Gaussian with a spatial FWHM of two μm and a temporal FWHM of 45 fs and 55 fs. As the heart beat period is anticipated to affect the attainable absolute efficiencies within the specularly mirrored harmonic cone, simulations have been carried out to certain the vary sampled within the experiment attributable to shot-to-shot variations (±5 fs), subsequently yielding a spread of optimized effectivity values (Fig. 1c, gray shaded space). Silver Müller boundary situations permit the free motion of particles and electromagnetic fields into and out of the simulation window⁵³. Smilei’s in-built Bouchard solver⁵⁴ is utilized to scale back the error from numerical dispersion in 2D inherent to conventional finite distinction solvers⁵⁵. It has been proven that modifications to the finite distinction method can sufficiently scale back this error⁵⁶.

Normalized vector potential

It is commonplace follow in PIC simulations to outline the normalized vector potential, a₀, as a substitute of depth. For a laser pulse of frequency, ω, and peak electrical subject amplitude, E, similar to an depth, (I=frac{1}{2}c{{epsilon }}_{0}{E}^{2}), the normalized vector potential is expressed as ({a}_{0}=frac{{eE}}{{m}_{{rm{e}}}comega }), the place c is the velocity of sunshine, ϵ₀ the vacuum permittivity and m_e the mass of an electron. In specific, the onset of relativistic results, together with SHHG, happens for a₀ ≥ 1. In Fig. 1c, an a₀ of 24 is used within the PIC simulation to mannequin the interplay. In Fig. 2, a₀ values starting from 2 to 25 are used to discover the parameter area and in Fig. 2b,c, the a₀ values are 13 and 21, respectively. An a₀ of 25 is utilized in Fig. 4b.

Analytical mannequin of XUV beam profiles

In actuality, the plasma floor isn’t anticipated to stay flat over the whole period of the Gemini laser pulse. The steep depth gradients lead to a ponderomotive drive, initially deforming the plasma floor right into a concave ‘dent’ that follows the profile of the central most of the focal spot. The dent depth Δz, outlined because the distinction between the place of the crucial density floor on the spatiotemporal peak of the driving laser pulse and its unique place, is calculated from an analytical mannequin²² detailed under for the experimental situations and in contrast with 2D PIC simulations with the identical parameters. An intensity-dependent section time period is included within the far-field harmonic beam profile mannequin to account for this denting.

The Fraunhofer diffraction equation describes the propagation of a monochromatic, λ, scalar subject amplitude to massive distances as a Fourier rework. It describes the propagation of a targeted laser pulse from its far-field profile, U₀(x′, y′), to its near-field profile, U(x, y, z), as

the place z is the gap from the focal spot. The far-field nth harmonic spatial profile is modelled by making use of the speculation of relativistic spikes¹⁰. Writing this adjustment as (S=S(n,{a}_{0}({x}^{{prime} },{y}^{{prime} }))), which incorporates a pointy roll-off at a harmonic order that scales as ({a}_{0}^{3}), the far-field amplitude of the nth harmonic is

Propagating the far-field harmonic beam profile to the near-field,

the place λ_n is the wavelength of the nth harmonic.

Radiation stress of the laser pulse on the goal floor causes a curvature of that floor, Δz = Δz_i + Δz_e, composed of the movement of the ion profile, Δz_i, and the tour of the electrons, Δz_e. This curvature is modelled with the addition of a section time period, ({phi }_{n}=2{ok}_{n}Delta zcos theta ), to the far-field profile. A earlier examine supplied a mannequin for Δz (ref. ²²). For the Gemini pulse period, the floor curvature varies slowly in time across the peak of the laser pulse, at t_p, when most harmonic emission happens. The ion floor dent produced by a laser pulse with normalized vector potential a_L(x′, y′, t′) incident at an angle θ on an exponential preplasma of scale size L is

the place ({varPi }_{0}={({RZ}{m}_{textual content{e}}cos theta /2A{M}_{textual content{p}})}^{1/2}); R is the reflectivity of the relativistic plasma mirror; Z and A are, respectively, the common cost state and mass variety of the ions; m_e is the electron mass and M_p is the proton mass. The electron tour is

Extended Data Fig. 7 compares the analytically calculated dent from ion movement to 2D PIC simulations with laser pulse intensities in keeping with the photographs of Fig. 3. Reflectivities are extracted from the simulations. Accounting for the corresponding section time period, the near-field depth profile of the nth harmonic is

We word that extra contributions to harmonic focusing might come up from intensity-dependent shifts of the obvious reflection level predicted by relativistic similarity concept^57,58. Overall, this modelling demonstrates that, in addition to elevated vitality within the harmonic beam, the experimental demonstration of the theoretically efficiency-optimized regime can be confirmed by (1) an intensity-dependent discount in spatial filtering (departure from Gaussian-like beam profile); (2) a fast, intensity-dependent progress in divergence; and (3) sturdy spectrospatial modulations within the emitted harmonic beam. The second and third factors are additionally indications of environment friendly emission from a concave-dented plasma.

Although a CHF was not measured immediately on this work, the commentary of fast will increase in beam divergence with depth, accompanied by the emergence of sturdy spectrospatial modulations supplies proof for the technology of CHF-suitable situations on this experiment. This is additional supported by 2D simulations, which predict the formation of a CHF through the experiment (Fig. 4). In the longer term, some extent of unbiased management of the plasma floor curvature is important to optimize each SHHG and the CHF efficiency concurrently. This management might be passive, that’s, a pre-shaped goal that compensates for induced curvature, or energetic—that’s, use of a considerably longer prepulse together with a single or few-cycle high-power driving laser pulse to exert absolute management over the form of the floor in the meanwhile of technology.

It stays unsure how the spectrospatial modulations noticed on this work will interaction with the spatiotemporal couplings that develop into more and more extra impactful at multi-PW ranges, however methods are actually out there for on-shot quantification of those couplings⁵⁹. Moreover, though spatiotemporal laser–plasma couplings can play a considerable half within the high quality of a CHF, within the effectivity restrict, simulations present that these are dominated, and so might be managed, by the driving subject of the incident laser pulse^3,60.

CHF 3D achieve convergence

The CHF achieve in 3D was estimated utilizing 2D PIC simulations, following the methodology in ref. ³⁹. The electromagnetic subject intensities of the mirrored harmonic beam are extracted from 2D PIC simulations on the time of technology of the brightest CHF, I_f, and on the time of technology of the attosecond pulse that produces this CHF, I₀. The snapshot of Fig. 4b is taken on the time of technology of the brightest CHF. The 3D increase can now be calculated. First, the achieve from temporal compression, Γ_D, is I₀/I_L, the place I_L is the depth of the incident laser pulse. The 2D achieve from spatial compression is Γ_2D = I_f/I₀. The 3D achieve that may be anticipated from spatial compression is ({varGamma }_{3text{D}}={varGamma }_{2text{D}}^{2}) and the total 3D achieve is Γ = Γ_DΓ_3D. The attosecond pulses of Fig. 4a are then scaled utilizing these equations and making the approximation that temporal achieve is fixed. As excessive harmonic orders contribute considerably to the depth on the CHF, the correct measurement of achieve requires an unfeasible simulation decision. To take a look at the standard of the decision used for the evaluation, a scan of achieve as a perform of simulation decision is plotted in Extended Data Fig. 8. A decrease certain on the achieve of 88 is extracted from the best accessible decision simulation, however larger good points appear possible.