Phase Change Random Access Memories (PCRAM), which are based on the reversible amorphous-crystalline transition in phase change materials (PCMs), constitute a very promising alternative to Flash technology, which is reaching fundamental limits. One of their key advantages is their scalability but, for ultimate miniaturization, energy consumption is critical and a promising solution is the geometrical confinement of the memory points. Mastering this with PCMs at ultimate dimensions (typically 5 nm) is, however, a real challenge, which calls for a fundamental understanding of the interplay between strain (the amorphous-to-crystal transition is accompanied by density increase of several %) and interface energies at the nanoscale. The objective of the SESAME project is to study the influence of strain and size on the PCM phase transition at ultimate dimensions. To address these issues we will use advanced in situ characterization techniques applied to ultra-thin layers, confined nanostructures and nanoclusters in order to investigate the early stages of phase transition and also to measure local strains and microstructure changes at crystallization. Five partners with complementary know-how will participate in the project: IM2NP-Marseille, CEA-LETI-Grenoble, CEA-INAC-Grenoble, synchrotron SOLEIL – St Aubin, CINaM-Marseille. The SESAME project will be organized along 5 tasks: 1. Coordination, 2. Sample preparation and characterization, 3. High resolution synchrotron X-ray scattering, 4. Transmission Electron Microscopy (TEM), 5. Simulation. Thin/ultra-thin GeTe and Ge2Sb2Te5 (GST) PCM films and PCM materials in confined structures will be prepared at CEA-LETI. Various thickness (100 to 5 nm), size (down to 10 nm width) and capping materials (Ta, TaN, Ta2O5, SiN, SiO2, Ti, TiN …) will be studied. CEA-INAC has the unique capability of preparing sub-10 nm GeTe and GST clusters by gas phase condensation. This will allow us to address the ultimate sizes, far beyond existing capabilities of lithography. Clusters with different composition or doping will be embedded in matrices with various thermo mechanical properties in order to evaluate the impact of mechanical stress on PCM clusters properties. Preliminary in situ sample characterizations will be performed at CEA: in situ ellipsometry, reflectivity, Raman spectroscopy or four-point-probe resistivity measurements. On these well-characterized samples unique in situ High-resolution synchrotron x-ray scattering and state-of-the-art transmission Electron microscopy (TEM) investigations will be performed. An original combination of resistance, X-ray diffraction and X-ray reflectivity that allows correlating structural and electrical PCM properties upon crystallization has been developed jointly by IM2NP and ESRF and will be used at synchrotron SOLEIL to characterize in situ the phase transition of ultrathin PCMs. Also the in situ combination of X-ray diffraction and optical curvature measurements developed jointly by IM2NP and DiffAbs beamline at SOLEIL will allow for an in-depth understanding of the mechanics involved in the amorphous-to-crystal transition. State-of-the-art TEM performed at CEA-INAC and CEA-LETI will bring valuable knowledge on local distribution of elements, defects and strains. In situ TEM observations during crystallization will offer invaluable information on the nucleation sites for crystallization. It is worth noting that these highly original in situ techniques (based either on TEM or Synchrotron radiation) will be used also to investigate structural changes in the amorphous phase. The issue of resistance drift in the amorphous phase is a key point for the stability of stored information in the memory cell. Atomistic simulations (Density Functional Theory, Molecular Dynamics) will be performed at CINaM in order to simulate the atomistic structure and the properties (structural, electronic, spectroscopic) of phase change materials in amorphous and crystalline form.

The objective of the XMicroFatigue project is to build a synchrotron based 3D x-ray microscope and combine it with a state-of-the-art device for in situ deformation of objects at the micrometre length scale. This advanced characterisation tool will provide unprecedented insights into the structural evolution of damage in materials under low cycle fatigue loading. For this purpose, the Laue microdiffraction station at the French BM32 beamline at the ESRF will be upgraded to put the performance at a higher level by achieving an additional sub-micrometre resolution along the sample depth. The so called DAXM method will be accelerated enabling the investigation of a wide range of heterogeneous and complex materials regarding their fundamental and industrial aspects, such as reliability and endurance. The fatigue damage accumulation will be studied in the vicinity of well-selected grain-boundaries of micron-sized cantilevers in two distinct cases: close to a free surface and in the material interior. The x-ray microscope will measure 3D spatially-resolved microstructural quantities relevant to understand the underlying interaction mechanisms between dislocation and grain-boundary, such as the deviatoric and hydrostatic strains, the orientation and the density of the so-called geometrically-necessary-dislocations. Thereby, the project enables the formulation of mechanism-based material laws, which brings us one step closer to the future of material science: grain- boundary engineering. The two partners (INAC BM32 beamline and the nano- and micromechanics group of the MPIE) have a recognised expertise in synchrotron based structural characterisation and in understanding the mechanical behaviour of micron sized objects by in situ deformation.

Mesophon project is devoted to the study of thermal transport in nano-materials where the characteristic sizes are comparable to the phonon wavelength Lph (100nm at 3K & 1nm at 300K). In order to reach this goal a multidisciplinary research project is proposed. It lies on the elaboration of specific nano-structures and on the study of their thermal properties by using state of the art measurement and modeling. The main asset of the project is to develop an original approach based on very low temperature measurements and numerical simulations to observe wave effects on phonon transport. In this low temperature limit, thermal transport in nano-structured semiconductor has been scarcely studied, as measurements are very challenging at these length scales. Nevertheless, recent experimental works pointed out significant contributions of intermediate phonon wavelength to heat transport at the nanoscale, indeed questioning existing textbook thermal models. Furthermore, there is an increasing interest for understanding phonon scattering in newly developed phononic devices that could be relevant for several applications, particularly in the fields of electronic and optic. Within this collaborative research program, specialists in the fields of elaboration, thermal measurement, theory and numerical simulation will work together. The scientific program involves 3 workpackages that address specific issues. • First, unique thin Ge films with Ge:Mn or Ge:Sn:Mn nano-inclusions will be grown at SiNaPS. Their design, concerning size and dispersion of clusters, make them the most suitable objects to be used as model materials to study wave effects due phonon-cluster interaction. • Second, thermal properties measurements of the elaborated thin films, from 300K to 4K, will be held at Institut Néel. Through dedicated devices, thermal conductivity and phonon drag effect on Seebeck coefficient are going to be measured. Collected data will help to develop a theoretical model for phonon scattering by nano-inclusions as a function of their wavelength. • Third, at the same time within the LEMTA, models and numerical simulations will be developed to appraise thermal transport properties and compare them with measurements. Feedback to the growing procedure will improve the targeted window in terms of size and dispersion of nano-inclusions. New numerical models targeting wave effect in the phonon transport will be investigated in order to appraise properties such as a phonon scattering phase function. This work is strongly collaborative and interdependent. From the material viewpoint, thin films elaboration for new devices with a controlled tailoring of the nano-structures is very challenging. Thus, developing characterization tools (experiments, models and simulations) that can measure or predict optimal properties give the necessary feedback for improving materials. From the advanced characterization standpoint, high sensitivity metrology development requires model materials with well defined properties as well as theoretical and numerical model to extract and appraised reliability of measured properties. Lastly, in what concerns modeling and numerical simulation at the atomic and mesoscales, realistic materials as well as accurate properties measurements are extremely valuables to assess models and tools. In this frame, the Mesophon project will provide useful knowledge for several research communities. Furthermore, as it is exposed in the detailed scientific document, preliminary studies including fabrication of Ge:Mn films, thermal conductivity measurements at 300K and simulations on nano-structured Ge were already conducted by the consortium members to appraise the project feasibility. The obtained promising results demonstrate that there are interesting scientific and technological challenges to tackle. Results of this study should foster the development of “phononic” science and its applications (thermal management, heat recovery, heat diodes, etc).

Highly disordered superconductor are the core element of very promising devices, both from a fundamental and an applied point of view: first, they are predicted to provide non dissipative high impedances in the microwave domain, giving rich possibilities for quantum information processing and for mesoscopic physics. Second, their strong non linearities, high kinetic inductance and / or high critical temperature enable the realization of new instrumentation such as high quantum efficiency detectors, or quantum limited amplifiers for astroparticle physics and cosmology. Exploitation of the full potentialities of disordered superconductors requires one to understand deeply their electrodynamics, which often deviates from that of a conventional BCS superconductor. In our project we focus on the Kinetic Inductance Detectors (KIDs), an infra-red photon sensor. These devices have been commonly implemented with Aluminium thin films, but highly disordered superconductors are predicted to yield far more sensitive and faster detectors. We will also develop lossless high inductances designed for Josephson photonics devices where high characteristic impedances allow exploring a quantum electrodynamics regime with fine structure constant > 1, offering a host of possibilities for generation and measurement of microwave quantum states. Despite exciting proof of concept realizations, previous attempts to use them in KIDs have been hindered by high frequency losses, a photon absorption weaker than anticipated, showing unexpected power or temperature dependence, and unexplained noise and quasiparticle relaxation. Other unconventional effects were reported: for instance the increase of the kinetic inductance with current (non linearity) is not always accompanied by an increase in dissipation, which is very desirable for a number of applications. Furthermore, the decay rates measured in resonators after the absorption of a photon are not identical for the dissipative and for the reactive response, two quantities which are supposed to be complementary in the BCS theory. In short, disordered superconductors have shown up to now a highly unusual electrodynamics, whose understanding is one of the goals of our project. The mechanisms at the root of this anomalous electrodynamics remain poorly understood, and lack a unifying theoretical framework, which prevents one from reaching the ultimate performances of such devices. One of the difficulties to comprehend these limitations is that highly disordered materials display inhomogeneities on the scale of the Fermi wavelength, making the usual approaches for standard metals inaccurate. Moreover, key microscopic parameters such as the superconducting coherence length are inhomogeneous and poorly known. The aim of this project is to fill this gap in understanding by detailed and complementary experimental and theoretical studies. The knowledge acquired throughout the project will allow us to solve technological bottlenecks in kinetic inductance detectors and will open numerous other interesting possibilities for mesoscopic physics.

Radical S-Adenosyl-L-Methionine (SAM) enzymes are capable of catalyzing chemically difficult reactions such as S atom insertions, cyclizations and unusual methylations. One-electron transfer from a conserved Fe4S4 cluster to SAM induces its cleavage and leads to the production of a 5'-deoxyadenosyl radical species. This radical species can extract a hydrogen atom from various substrates allowing for the propagation of the reactive unpaired electron. SAM can be either a cofactor or a substrate depending on whether the electron returns after the reaction to regenerate the reduced Fe4S4 cluster and the cofactor, or this is definitively consumed after the reaction. Several radical SAM enzymes have been characterized in terms of structures, substrates and catalytic mechanism. However, no systematic study of the factors that modulate specific radical formation and subsequent product generation has been carried out. To perform such a study we have selected four radical SAM enzymes: HydG (a FeFe-hydrogenase maturase), ThiH (involved in thiamine synthesis), FO (didemethyl-hydroxy-deazariboflavin) synthase, (involved in the methanogenic F420 coenzyme synthesis) and NosL, (an enzyme involved in antibiotic synthesis), all of which bind an aromatic amino acid and catalyze the same intermediate reaction but generate widely different products. Tyrosine is the substrate of HydG, ThiH and FO-synthase whereas the substrate of NosL is tryptophan. In all cases, the radical reaction results in the cleavage of the amino acid Ca-Cß bond. One of the fragments liberated by this reaction is either dehydroglycine (ThiH and FO-synthase) or a glycyl radical (NosL and possibly HydG). HydG, ThiH and NosL subsequently use this fragment to generate their respective products: CO/CN-, the thiazole ring and the antibiotic nosiheptide. Conversely, the CofH component of FO-synthase uses the p-cresyl radical, the other product of the cleavage reaction when dehydroglycine is produced, for the synthesis of FO. Thus, p-cresol can either be a secondary, unproductive final by-product of the reaction, or an integral component of the product. In a related reaction, NosL carboxylates the tryptophan-derived fragment using the glycyl radical to generate its product. The goal of this project is to understand how each protein channels the reaction toward a given pathway, making this radical reaction specific. Our intention is not to stop at deciphering the mechanism of one specific enzyme but, rather, to extend our analyses to extract general principles about the underlying radical-based chemistry. To achieve this goal we will develop a global approach combining biochemical, crystallographic, spectroscopic and computational techniques to identify both the active radical species involved in the reactions and the protein-substrate interactions that play key roles in orienting the selectivity. The four partners have highly complementary competences that will be variously applied to the different target proteins: protein production and detailed functional studies with characterization of products and kinetic catalytic parameters by Partner 2; protein production, X-ray protein structure determination including complexes with substrates, inhibitors and products, time-resolved kinetic studies in the crystals and theoretical calculations on these systems by Partner 1 (a crystallization robot and automated crystallization plate monitoring setup are operational inside a glove box; this is one of the very few setups for automated anaerobic protein crystallization worldwide); Mössbauer and EPR spectroscopic experimental and theoretical studies by Partners 3 and 4 (Partner 3 has the only Mössbauer platform dedicated to biological macromolecules in France, and one of the few existing in the world; Partner 4 is a member of an “Advanced EPR” team that performs pulsed ENDOR and ESEEM experiments).