Journal: Nature

Article Title: Super-resolution stimulated X-ray Raman spectroscopy

doi: 10.1038/s41586-025-09214-5

Figure Lengend Snippet: a , A SASE XFEL beam is focused onto a 4.5-mm gas cell filled with neon. The incident X-ray intensity was controlled by a gas attenuator, and the transmitted X-ray beam was attenuated by Al filters with thicknesses 3.5 μm, 5 μm and 10 μm. A slit constrained the size of the transmitted X-ray beam before the spectrum was measured by dispersing the beam on a 2D detector. b , Level scheme for stimulated X-ray Raman transitions. The incident SASE X-ray (in blue) produces core-excited states (1 s −1 3 p , 1 s −1 4 p ), which transition to final states (2 p −1 3 p , 2 p −1 4 p ) by emission of Raman photons. Each spectral spike in the SASE pulse introduces a corresponding Raman transition, and covariance analysis of a collection of single-shot Raman scattering spectra ( ω 2 ) with the transmitted incident spectra ( ω 1 ) generates dispersive lines indicative of the constant energy loss of Raman transitions through different channels.

Article Snippet: The layout of the s-SXRS experimental setup is shown in Fig. . An incident SASE XFEL pulse (7.5 eV full width at half-maximum (FWHM), 40 fs) from the European XFEL was focused into a gas cell containing 1–4 bar neon with the central photon energy resonant with the 1 s → 3 p transition in neon.

Techniques:

Journal: Nature

Article Title: Super-resolution stimulated X-ray Raman spectroscopy

doi: 10.1038/s41586-025-09214-5

Figure Lengend Snippet: The incident SASE intensity is changed by the gas attenuator (GATT) to produce transmissions of 25%, 50%, 70%, 80% and 100%. The results of 1,000 shots (blue dots) for each transmission value fluctuate because of the randomness of the single-shot SASE energy spectrum—SXRS gain occurs when a SASE spike matches a transition energy as shown in Fig. . SASE pulses at the lowest peak intensity generate emission spectra with signals that barely emerge from the detector noise, setting the limit for discrimination of XRL and XRS. The average emission pulse energy of the top 10% shots (red) increases exponentially before saturation. The stimulated emission has two components: XRL (magenta) and XRS (green). The inset shows the decomposition of the XRL and XRS yield for the highest intensity data obtained by averaging the individual shots. Simulations of isolated attosecond pulse propagation using a 0.25-fs Gaussian pulse with 7.5 eV bandwidth (black) as a function of peak intensity parallel the experimentally observed gain with SASE pulses (red).

Article Snippet: The layout of the s-SXRS experimental setup is shown in Fig. . An incident SASE XFEL pulse (7.5 eV full width at half-maximum (FWHM), 40 fs) from the European XFEL was focused into a gas cell containing 1–4 bar neon with the central photon energy resonant with the 1 s → 3 p transition in neon.

Techniques: Transmission Assay, Isolation

Journal: Nature

Article Title: Super-resolution stimulated X-ray Raman spectroscopy

doi: 10.1038/s41586-025-09214-5

Figure Lengend Snippet: Spike analysis of measured (first row) and simulated (lower rows) SASE spectra. (A) a randomly selected single-shot spectrum normalized to the peak intensity of the averaged spectra. The peaks are identified with adjacent spacing larger than 0.1 eV and height above 0.1. The statistics of the spacing between adjacent spikes (B) and the number of peaks per shot (C) are obtained by analyzing 1000 shots. The maximum of distribution of spacing between adjacent spikes is around 0.5 eV, and most of the shots have approximately 23 spikes. The corresponding analyses of the simulated SASE are shown in the lower panels. The distribution curves and spectra look similar, but there are more spikes (90) in the simulated SASE, and the distribution of the spacing between the spikes peaks at around 0.1 eV. The difference is due to the 0.2 eV spectrometer broadening of the measured SASE pulse. The simulated spectra and statistics are close to the experimental results after convolving with a 0.2 eV Gaussian function (G). A representative simulated single-shot temporal profile is shown in (J). The statistics of the temporal profile of 1000 shots show a most likely adjacent spike spacing of 0.6 fs with 93 temporal spikes in a pulse.

Article Snippet: The layout of the s-SXRS experimental setup is shown in Fig. . An incident SASE XFEL pulse (7.5 eV full width at half-maximum (FWHM), 40 fs) from the European XFEL was focused into a gas cell containing 1–4 bar neon with the central photon energy resonant with the 1 s → 3 p transition in neon.

Techniques:

Journal: Nature

Article Title: Super-resolution stimulated X-ray Raman spectroscopy

doi: 10.1038/s41586-025-09214-5

Figure Lengend Snippet: (A) Single-shot SASE spectra were measured as the gas cell was empty. The spikes in the pulse exhibit random fluctuations from shot to shot. The average spectrum displays a smooth Gaussian profile with a 7.5 eV bandwidth (FWHM). The standard deviation (std) of the spectral intensity fluctuation at a fixed photon energy is calculated, and the resulting percentage fluctuation (std/mean) remains steady at approximately 40% within the main spectrum. (B) The average simulated spectrum also exhibits a Gaussian profile with a 7.5 eV bandwidth (FWHM). However, it contains more spikes that fluctuate fully randomly from shot to shot, resulting in a steady 100% fluctuation of the intensity across the entire spectrum.

Article Snippet: The layout of the s-SXRS experimental setup is shown in Fig. . An incident SASE XFEL pulse (7.5 eV full width at half-maximum (FWHM), 40 fs) from the European XFEL was focused into a gas cell containing 1–4 bar neon with the central photon energy resonant with the 1 s → 3 p transition in neon.

Techniques: Standard Deviation

Journal: Nature

Article Title: Super-resolution stimulated X-ray Raman spectroscopy

doi: 10.1038/s41586-025-09214-5

Figure Lengend Snippet: (A) Calibration of the detection system using spectra integral to pulse energy. The SASE pulses with a center photon energy of 850 eV pass through an empty gas cell, and the GATT and filters are adjusted to optimize the signals on the detector. Three filters with thicknesses of 3.5, 5, and 10 μ m are employed, and their attenuation is divided to obtain the spectra. (B) The incident pulse energy is obtained by multiplying the energy measured by an upstream gas monitor detector with the beamline transmission factor of approximately 0.5. The linear relationship between integrated spectra intensity and incident pulse energy provides a calibration that can be used to estimate the pulse energy emitted for subsequent runs with the gas in the cell.

Article Snippet: The layout of the s-SXRS experimental setup is shown in Fig. . An incident SASE XFEL pulse (7.5 eV full width at half-maximum (FWHM), 40 fs) from the European XFEL was focused into a gas cell containing 1–4 bar neon with the central photon energy resonant with the 1 s → 3 p transition in neon.

Techniques: Transmission Assay

Journal: Nature

Article Title: Super-resolution stimulated X-ray Raman spectroscopy

doi: 10.1038/s41586-025-09214-5

Figure Lengend Snippet: The XFEL pulses are focused on a gas cell by K-B mirrors. The transmitted and emitted x rays pass through an Al filter and a slit before being dispersed on the pixelated Andor detector by a spherical grating with 5-m radius and 1200 lines/mm .

Article Snippet: The layout of the s-SXRS experimental setup is shown in Fig. . An incident SASE XFEL pulse (7.5 eV full width at half-maximum (FWHM), 40 fs) from the European XFEL was focused into a gas cell containing 1–4 bar neon with the central photon energy resonant with the 1 s → 3 p transition in neon.

Techniques:

Journal: Nature

Article Title: Super-resolution stimulated X-ray Raman spectroscopy

doi: 10.1038/s41586-025-09214-5

Figure Lengend Snippet: The super-resolution s-SXRS has linewidth smaller than the SASE spike bandwidth 0.1 eV when fitting at regions away from the resonances that are distorted. (A) The Gaussian fitting of the 10 lineouts of the super-resolution s-SXRS with 1.0 bar neon pressure (2D map shown in the main text) in the range ω 1 ∈ [866.1, 866.8] eV. (B) The FWHM of the fitted Gaussian functions at different ω 1 lineouts. The FWHM data is divided by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\sqrt{2}$$\end{document} 2 to show the expected energy-loss axis projected resolution.

Article Snippet: The layout of the s-SXRS experimental setup is shown in Fig. . An incident SASE XFEL pulse (7.5 eV full width at half-maximum (FWHM), 40 fs) from the European XFEL was focused into a gas cell containing 1–4 bar neon with the central photon energy resonant with the 1 s → 3 p transition in neon.

Techniques: