A helium cell as a lens: shaping free-electron laser pulses without glass

An international collaboration led by the Max-Planck-Institut fuer Kernphysik in Heidelberg has used an atomic gas as a time-dependent lens, shaping an extreme ultraviolet pulse both spatially and in its spectral composition ^[1]. That matters because the optics that route visible light are not available at these photon energies: XUV photons are strongly absorbed by standard glass lenses, mirrors and prisms rather than reflected or refracted ^[3], and while free-electron lasers can now deliver ultrashort, high-brilliance XUV pulses, tools to shape those beams remain scarce ^[4]. The work is published in Science Advances ^[2].

The mechanism is not a passive material property. When an intense XUV pulse crosses an optically dense atomic gas, it rapidly drives electrons up and down between atomic energy levels, a Rabi oscillation ^[5]. A beam is brightest at its centre and weaker at the edges, so the strength of that interaction varies across the beam profile, producing a self-induced change in the medium's refractive index that is most pronounced near an atomic resonance ^[6]. According to Yu He, the study's first author, the interplay between this intensity-dependent interaction and macroscopic propagation effects along the medium "turns the atoms effectively into a refractive element, such as a lens or prism" ^[7]. He adds that the gas deflects the resonant light outward more effectively than other frequencies, and so reshapes the spectrum ^[8].

The measurement was made by focusing high-intensity XUV pulses from FLASH at DESY in Hamburg into a gas cell of helium ^[9]. Peak photon energy was 21.2 eV, matching the fundamental 1s-2p transition in helium ^[10].

Two constraints follow directly from how it works. First, the element is active rather than passive: because the index change is created by the pulse's own intensity driving Rabi oscillations, a weak beam gets no lens ^[1]. Second, the operating point is set by the gas. The effect peaks near resonance, and the demonstration used a photon energy chosen to sit on helium's transition, so the working energy of any such element is tied to the chosen species ^[3].

It is also worth being precise about the claim being made. MPIK frames the gas element as a route toward better XUV and X-ray pulse control, with chemical reaction steering, quantum computing and spectroscopy as targets ^[13], and the argument for caring is that many atomic transitions sit in this energy range ^[12]. But the experiment reported here is XUV at 21.2 eV in helium; X-ray control is a stated prospect, not a demonstrated result ^[2].

What to watch: throughput and conversion efficiency, which the supplied account does not quantify ^[15]; which gas-resonance pairs cover useful photon energies given that each species fixes its own operating point ^[3]; and whether the approach survives translation to shorter wavelengths, where the same absorption problem that motivates it is worse ^[3]. Beamline designers should also want to know the intensity threshold below which the lens stops existing ^[1].