Lasing of a cavity-based X-ray source
Article Meta
Article Date: 28 January 2026
Article URL: https://www.nature.com/articles/s41586-025-10025-x
Article Title: Lasing of a cavity-based X-ray source
Article Image: 
Summary
Researchers at the European XFEL report the first demonstration of lasing in a cavity-based X-ray free-electron laser (CBXFEL). The experiment used four undulator segments inside a 132.8 m Bragg-reflecting cavity built from diamond (400) crystals and orthogonal Kirkpatrick–Baez mirrors, synchronised to 2.23 MHz electron-bunch trains. When the cavity length was tuned to match the bunch repetition rate, recirculating monochromatised X-ray pulses seeded subsequent bunches, producing multi-pass gain and a strong, spectrally narrowed output. The team measured per-pulse outcoupled energies of 4–13 μJ, an average round-trip gain of around 0.4, and a transmitted seeded bandwidth down to ~0.24 eV FWHM (limited by the spectrometer resolution). Thermal load on the diamond crystals introduced intensity oscillations within trains; cryogenic cooling and improved outcoupling are expected to raise performance by orders of magnitude.
Key Points
- First experimental lasing demonstration in a cavity-based X-ray FEL (CBXFEL) at the European XFEL.
- Cavity: two diamond crystals in Bragg (400) back-reflection plus focusing mirrors forming a 132.8 m round-trip retro-reflector.
- Electron bunch trains at 2.23 MHz were synchronised to the cavity; micrometre cavity-length tuning and nanoradian alignment were essential.
- Seeding by recirculating photons produced strong multi-pass gain, spectral narrowing and stable, highly coherent pulses compared with single-pass SASE.
- Measured outcoupled pulse energies 4–13 μJ; average round-trip gain ≈ 0.4; transmitted seeded spectrum ≈ 0.24 eV FWHM (spectrometer-limited).
- Thermal effects in diamond crystals (heat bumps and differential lattice shifts) caused intra-train oscillations; cryogenic cooling should mitigate this and boost output.
- Simulations indicate potential pulse-energy increases to millijoule levels and up to three orders-of-magnitude higher spectral flux versus SASE/self-seeded XFELs.
- Enables future high-resolution and nonlinear X-ray experiments (meV spectroscopy, resonant inelastic scattering, quantum X-ray optics).
Content Summary
The paper places this demonstration in historical and technical context: while FEL oscillators are routine from infrared to UV, hard X-ray oscillators were prevented by lack of suitable cavity optics. Diamond crystals with high reflectivity and thermal properties make Bragg cavities feasible. The experiment used the last four segments of the SASE1 undulator at 14 GeV, with retro-reflector assemblies 66.41 m apart and a cavity round-trip of 132.8 m.
Alignment and synchronisation were performed to micrometre and nanoradian tolerances. When the cavity length matched the 2.23 MHz bunch spacing, in-cavity diagnostics showed clear ring-up and ring-down behaviour and an orders-of-magnitude rise in spectrometer signal, proving oscillator lasing. The seeded pulses became much narrower and more intense than the initial single-pass SASE output. The observed cavity round-trip reflectivity was lower than ideal (≈67–75% vs theoretical ≈97%) due to optical imperfections and finite mirror apertures.
Although the demonstrator did not yet reach peak fluences of top SASE/self-seeded systems, measured performance and simulations suggest substantial room for improvement via: optimised electron beam and cavity parameters, cryogenic cooling of diamond crystals to suppress thermal bumps, and advanced outcoupling schemes. With these, CBXFELs could produce fully coherent, highly stable and spectrally pure hard X-rays in comparatively compact assemblies.
Context and Relevance
This result is a watershed moment for X-ray sources. It experimentally validates a two-decade-old CBXFEL concept and shows that synchronising a long X-ray cavity to MHz electron bunch trains is technically achievable. For users of XFELs, the promise is more stable, fully coherent X-ray pulses with dramatically improved spectral brightness — a step-change for spectroscopies that require high resolution and stability (for example, meV-resolved inelastic and nuclear-resonance techniques), and for nonlinear or quantum X-ray experiments that need high coherence.
The work also highlights practical engineering challenges: nanoradian-level alignment, micrometre-scale cavity tuning, and thermal management of diamond optics under megahertz heat load. Solving these will be decisive for scaling CBXFELs to millijoule pulse energies and opening new experimental regimes.
Why should I read this?
Short and sweet: if you care about the next generation of X-ray tools — better coherence, far narrower spectra and much more stable pulses — this is the demo that proves the idea works. It tells you what’s solved, what still trips people up (hello, heat bumps), and how far performance could jump with sensible engineering. Worth skimming unless you only use visible light labs.
Author style
Punchy: this is a major experimental milestone. The team’s work turns a long-standing theoretical proposal into reality and points clearly to practical upgrades that could make cavity-based X-ray oscillators a routine, transformative user technology.