Imagerie Nanométrique Femtoseconde par diffraction X cohérente – I-NanoX

Developing a photon source compatible with coherent diffractive imaging requires working at every stage of the beamline, from the initial laser to the source itself, with the continual support of numerical simulations. A new wave front sensor was developed for short wavelengths and was used to increase the photon flux and the beam coherence on sample. In the framework of active collaborations, we implemented original diffractive imaging techniques adapted to our beamline. Coherent diffractive imaging is based on iterative algorithms to retrieve the spatial phase of the field diffracted by an isolated sample, in order to reconstruct its image. Holographic techniques lead to a direct reconstruction without any ambiguity, and are very efficient when the signal is weak. The training of the group’s PhD students to focused ion beam techniques allows them to create samples with sub 100nm details. Finally, the Attophysique group expertise in pump-probe setups was used to implement the two color infrared/XUV delay line. It offers a femtosecond resolution, while each arm of the interferometer is around 10 meter long.

The extensive optimization work on the whole XUV radiation source and on the beam transport line lead to 10^9 available photons per pulse (20 fs each) at 32nm wavelength (H25) in a 5 microns focal spot. The resulting intensity is up to 10^12 W/cm². Thanks to the use of a XUV wave front sensor, the beam spatial coherence is better than 0.8. A new technique was developed to measure the beam complete spatial coherence map in a single laser shot. We also optimized the beamline towards shorter wavelengths. The photon flux at 20nm (H39) is thus 10^8 photons per laser shot.
This unique radiation source allowed us to demonstrate lensless imaging on a laboratory source with a spatial resolution of 78nm (twice the illuminating wavelength) in a single laser shot, i.e. with 20fs temporal resolution. We also implemented Fourier transform holography with extended references, a technique more robust and better suited to our beamline characteristics than iterative algorithms. We developed a new 3D imaging technique, compatible with the single shot regime, based on the stereoscopic human vision principle.

Our beamline unique capability to probe matter with both sub 100nm spatial and 20fs temporal resolutions opens new ways to study dynamics in solid state physics. The interaction between a sample and a pump laser pulse usually destroys the sample. It is therefore often mandatory to record the sample image in a single XUV pulse. While this possibility was previously limited to the very large facilities that are free electron lasers, it is now accessible to any laboratory with a commercially available TW laser. The first demonstration experiment will study the expansion dynamics of a laser created nanoplasma. Other studies, such as nanomagnetism or phase transitions, are already planned. Our laboratory is open to other researchers (through French and European call for proposals), other research teams have already mentioned their interest in using our beamline.
Beyond applications, new developments will be pursued. We will implement lensless imaging in a reflective geometry. Because of the very high absorption of many materials in the XUV range, this new geometry will allow the study of thick samples and of those grown on opaque substrates.

The project results were published in prestigious journals (5 published articles, 2 submitted and 3 in preparation), among which Physical Review Letters and Optics Express. Moreover, they were presented in numerous international conferences, half of which were invited contributions. The published results ranged from the work on the XUV source (Ge Opt. Express 2013, Ge JMO 2013, Mahieu submitted to Appl. Opt.) to imaging techniques (Gauthier PRL 2010, Ge Opt. Express 2013,…).