Arthur MOISSET, Christophe PETIT, Pascal ANDREAZZA, Damien CHAUDANSON, Olivier MARGEAT, Suzanne GIORGIO *,

• Sorbonne Université,CNRS, MONARIS, UMR 8233, 4 place Jussieu, 75005 Paris
• Université d’Orléans, CNRS, ICMN, UMR 7374, 1 rue de la Ferollerie, 45100 Orléans
• Aix-Marseille Université, CNRS, CINaM, UMR 7325, Campus de Luminy, 13288 Marseille

The control of the shape and structure of metal nanoparticles or nanoalloys is important for applications. It mainly depends on the crystal growth mechanism. Co nanoparticles (NPs) and Co-Ni nanoalloys, grown from solution, exhibited a shape transformation from nanospheres to nanorods [1]. The growth mechanism from nanospheres to nanowires as in figure 1, was studied by the association of two in situ techniques, liquid microscopy and SAXS. In situ SAXS measurements have shown a textured media between the oleylamine (OAm) and the cobalt precursor. The shape transformation of pure Co NPs and Co-Ni nanoalloys was observed by TEM in a drop of OAm solution diluted in toluene or heptane, encapsulated between graphene layers, as described in previous works [2]. The growth of the NPs and alloys in an OAm visquous solution allows a growth mechanism by direct adsorption of the monomers and an homogeneous size and shape distribution, as seen in figures 2 a-c (a), after 5mn (b), 15 mn(c) . At RT, after the growth of NPs in the encapsulated solution, the OAM has a tendency to texture while the metal NPs are dissolved (figure 2d-e, 17 mn (d) and 18 mn (e)). Then, Co re-crystallizes as nanorods as in figures 2 f-g-h (19 mn (f), 23 mn (g), 27 mn (h)). The same mechanism was observed with Co-Ni nanoalloys. Careful observation of NPs in diluted OAm during the NPs dissolution in the electron beam, shows residual metal clusters with sizes below 1 nm, as in figure.2i (scale bar 50 nm). By the following, these remaining small clusters are the results of the beginning of the Co NRs’ formation and tend to align along the OAm texture then, they grow as nanorods. Références: [1] L.Meziane, C.Salzemann, C.Aubert, H.Gérard, C.Petit , M.Petit, Nanoscale, 8 (2016), p.18640 [2] De Clercq, W. Dachraoui, O. Margeat; C.R.Henry, S.Giorgio; J. Phys. Chem. Letters, 5 (2014) p.2126 [3] A. Vivien, M. Guillaumont, L. Meziane, C. Salzemann, C. Aubert, S. Halbert, H. Gérard, M. Petit and C. Petit, Chem.Mater., 31, 3, (2019), p. 96

Jianhan XIONG, Nicolas DUPRE, Philippe MOREAU *, Bernard LESTRIEZ

• Université de Nantes, CNRS, Institut des Matériaux Jean Rouxel, IMN, F-44000 Nantes, France

Negative electrodes based on silicon have been considered for some time (due to their theoretical high capacity) [1], but present long term performance not meeting the current requirements. This is due to the large change of volume of the material with its reaction with lithium ions and its unstable passivation against the electrolyte. In order to mitigate this downfall evolution, mixtures of silicon with graphite (a long used and efficient material) are now being developed and studied [2]. This is the objective of the ANR project SILMARILION concerning the negative electrode of the complete lithium-ion battery to be produced. Since interactions between the different constituents of the composite electrode (Si, graphite, binder) are essential to the behaviour during cycling and occurs at very small scale, (S)TEM experiments are quite useful. In this study, we have used valence electron energy-loss spectroscopy to map the different constituents of the pristine electrodes as well as the reaction products along cycling. The use of this method was shown [3] to be very effective even on highly beam sensitive materials such as lithium fluoride or polymeric entities. Thanks to a method developed by M. Boniface in his PhD thesis [4], it was even possible to grasp the evolution of these constituents along cycling in a quantitative way. To that end, the Gatan 969 spectrometer (equipped with a K2) of the Nant’Themis, some PCA post-processing and MLLS fitting routines were used. They allow identifying and, visualising the role of the PAA polymer binder (Figure 1) and the presence of a hairy silicon (Figure 2) as soon as 30 cycles, despite the presence of graphite [5].

Pedro Lourenco *, Nadine WITKOWSKI, Geoffroy PRÉVOT, Hervé CRUGUEL, Romain BERNARD

• Institut de Nanosciences de Paris, Sorbonne Université, Paris, France

Transition-metal dichalcogenides (TMDC), such as WS2 and WSe2, are layered two dimensional materials, having a wide number of applications in optoelectronics, energy storage or catalysis. Epitaxial films are usually synthesized by chemical vapor deposition (CVD), leading to high-quality stoichiometric crystals with large domain sizes. Recently, there have been efforts to synthesize high-quality WS2 and WSe2 films through the use of reactive magnetron sputtering,[1] a versatile physical vapor deposition (PVD) process already implemented in the semi-conductor industry. As compared to CVD, more structural defects and smaller domain sizes are however observed. This could be explained by the impact of energetic species at the surface of the film during the growth. These structural defects play an important role in the characteristics of the TMDC. For example, chalcogenide vacancies are known to decrease the carrier mobility and thus to deteriorate the device performance. [2] To better understand the impact of energetic particles on the TMDC surface, we have generated single defects on the surface of WS2 and WSe2 crystals using low energy (50 eV to 500 eV) Ar+ ions with a controlled fluence. The irradiated samples were then analyzed by Scanning Tunneling Microscopy (STM). When imaged at negative sample bias, the defects appear as a bright protrusion due to the presence of a localized electronic state, associated with chalcogenide vacancies (see Fig. 1). Whereas a strong dependance with Ar+ energy of the probability of defect creation is observed, the shape and size of each defect, when imaged by STM, appears to be independent of the energy (see Fig. 1). In order to have an atomistic view of the collision phenomena, we have performed Molecular Dynamics (MD) simulations. These simulations show that there is a higher probability of creating defects at higher Ar energies, agreeing with the experimental results. They also show that there are more defects created which are located into the layers underneath the surface due to the collision cascade. This might explain why there is no clear difference in defect morphology with varying Ar+ energy, as the effect of the generated structures in the layers underneath cannot be fully probed by STM. Références/References : [1] Villamayor, M. et al, Vacuum, 188 (2021). [2] Yu, Z. et al. , Nat Commun 5, 5290 (2014)

Mahdi Jaber *

• Univ. Grenoble Alpes / CEA / CNRS, SPINTEC, France.
• CEA, IRIG/MEM, France.

A major driving force in applied spintronics is magnetic memories. While existing products rely on uniformly-magnetized elements, data storage using the motion of magnetic domain walls in two-dimensional strips using spin-polarized current is also investigated [1]. The three-dimensional geometry like in cylindrical nanowires and nanotubes is researched [2]. Magnetic domain walls in the latter are distinct from those in flat films and strips, opening perspectives for new physics. In comparison to cylindrical nanowires, isolated ferromagnetic nanotubes have given rise to relatively few experimental reports. Here, we describe the characterization of high-aspect ratio CoNiB nanotubes, grown chemically by electroless plating, using various TEM techniques such as Energy-dispersive X-ray spectroscopy EDX, Electron energy-loss spectroscopy EELS, and in-situ Electron Holography. Previous experimental studies revealed the possibility for orthoradial magnetization (i.e. azimuthal domains), a phenomenon not expected by theory, but interesting for the application side [3] (Figure 1.a,b). Based on that, we performed TEM electron holography to explore the magnetic configuration in nanotubes, as well as their magnetic history under applied fields (Figure 1.c,d). We also evidence that Nickel tends to segregate on both side of the CoNi layer, which is supposed to affect the system magnetic behavior (Figure 2). To help the physical and the theoretical interpretation of the measurements, reasonable efforts on data analysis and micromagnetism simulations will be made. As a result, a deeper understanding of the spin textures in nanotubes will contribute to enhance the productivity in spintronics. References: [1] J. Hurst et al. Theoretical study of current-induced domain wall motion in magnetic nanotubes with azimuthal domains. Phys. Rev. B 103, 024434 (2021). [2] Magnetic nanowires and nanotubes, M. Staňo, O. Fruchart, Handbook of Magnetic Materials. Ed. Ekkes Brück vol. 27, 2018, Pages 155-267 [3] Staňo, M. et al. Imaging magnetic flux-closure domains and domain walls in electroless-deposited CoNiB nanotubes. SciPost Phys. 5 (2018) p 038.

Naoures HMILI *, Swapneel THAKKAR, Quentin WEINBACH, Patrick allgayer, Marc SCHMUTZ, Laure BINIEK, Alain CARVALHO

• CNRS – Institut Charles Sadron – 23 Rue du Lœss 67034 Strasbourg, France.

Porous bulk polymers are of great interest for their advantageous properties like low bulk density, flexibility and high surface area. These properties among many others, open new doors for diverse applications such as the hot topic of thermoelectricity. In that field, the porous polymer should be electrically conductive and thermally isolating. Ideally, an anisotropic structure with well oriented pores is crucial to reach high thermoelectric efficiency. Porous conducting polymers can be prepared by freeze-drying a hydrogel (PEDOT:PSS) leading to a porous cryogel. But, with a classic flash freeze-drying a non-controlled porous structure is obtained as shown in the SEM image (fig.1) (taken by Hitachi SU 8010 Ultra High Resolution Field Emission Scanning Electron Microscope). In this project, our team applies the ice templating technique [1] on PEDOT:PSS gels to control the porous structure. This involves the unidirectional growth of ice crystals in the gel as a texturing agent for the porosity. We demonstrate by cryoSEM that while freezing, the ice crystals grow and push apart the polymer fibers to create the walls of the pores. Once freezing is complete, the sample is dried by sublimation. A porous cryogel is then obtained in which the pores are the replica of the ice crystals. Via this method, and by controlling the cooling rate and temperature [2], we have succeeded to channel the pores and obtain a well aligned honeycomb structure as seen in fig.2 (a) and (b). Finally, the pore size as well as the directionality of the channelled pores were analysed by ImageJ (using respectively “analyse particles” and the plug-in “orientation J”). References: [1] Deville, S. (2013). Ice-templating, freeze casting: Beyond materials processing. Journal of Materials Research, 28(17), 2202–2219. doi:10.1557/jmr.2013.105 [2] S. Deville, Ice-Templating: Processing Routes, Architectures, and Microstructures. 2017. This Master internship work has been supported by a Sfμ grant.

Sumit Kumar *, Frédéric Fossard, Gaëlle Amiri, Jean-Michel Chauveau, Vincent Sallet

• GEMaC, CNRS - UVSQ, Université Paris-Saclay
• LEM, ONERA-CNRS, Université Paris-Saclay

Unique growth mechanisms involved in semiconductor nanowires (NWs) pave the way to the achievement of new crystallographic phases and remarkable material properties. Interestingly, in the case of 1D nanostructures, polytypism can arise due to the particular growth mode below a catalyst droplet, which may induce stacking faults along the length of the nanowire. Moreover, these stacking faults can be correlated and form ordered arrays, until giving rise to new phases (polytypes) with distinct properties[1]. Hence, 4H, 6H, 8H, and 10H (so-called high order polytypes) can be observed in nanowires[2]. Hence, studying polytypism in semiconductor NWs arouses a strong interest for the next generation of electronic and photonic applications. In this framework, ZnS is an important II-VI semiconductor that has a wide range of optoelectronic applications including luminescent devices, infrared windows, and UV-photodetectors[3]. In this work, Au-assisted ZnS nanowires were grown by MOCVD on GaAs (111B) substrate (VLS mode or VSS mode). The idea is to provide a change in the growth mechanism via the physical state of catalyst droplets (liquid or solid) and hence, study the induced polytypism in ZnS nanowires. ZnS nanowires with length up to 1.4 µm and diameter in the range 10–34 nm was successfully achieved. The obtained morphologies and densities of the NWs have been systematically inspected by scanning electron microscopy (SEM) directly on the substrate (illustrated in Fig 1). Transmission Electron Microscopy has been also used to investigate the crystallographic structures and compositions of both catalysts and nanowires. HRTEM observations revealed that the NWs exhibits periodic stacking faults, and the resulting structure was accurately identified as 3 sequences of 5 planes ABCBA-BCACB-CABAC (i.e. 5th order superstructure), giving rise to an astonishing 15R crystal structure (Fig 2)[4]. This structure is highlighted for the first time in ZnS nanowires. We modeled this 15R structure in the framework of the classical nucleation theory and axial-next-nearest-neighbour-Ising model (ANNNI). References [1] G. Priante et al. Phys Rev B. 89 (2014) 241301 [2] Y. Jiang et al. Adv. Mater. 15 (2003) 1195 [3] S. Premkumar et al. Sci. Rep. 9 (2019) 18704 [4] S. Kumar et al. Nano Res. (2021) https://doi.org/10.1007/s12274-021-3487-8

Marie-Jose SALEH AFIF *, Stephanie JUBLOT-LECLERC, Joel RIBIS, Aurelie GENTILS, Marie LOYER-PROST

• Université Paris-Saclay, CEA, Service de Recherches de Métallurgie Physique, 91191 Gif-sur-Yvette, France
• Université Paris-Saclay, CNRS/IN2P3, IJCLab, 91405 Orsay, France

Because of industrial development and the growth of the world population, nuclear reactors are developed to cover the increase in energy consumption. The nuclear energy produced through fission and fusion reactions is indeed capable of producing a very large amount of electrical power. One of the problems encountered during these nuclear reactions is the irradiation swelling of fuel cladding materials. Hence nanostructured materials such as Oxide Dispersion Strengthened (ODS) steels were especially developed to resist to irradiation [1] and high temperature [2]. These steels are composed of a metallic body-centered cubic matrix (FeCrW or FeCrMo) reinforced with nano-oxides of titanium and yttrium. As the ODS steel matrix is mainly a Fe-Cr system, a Cr-rich phase, alpha’, can precipitate under irradiation [3] and brittle the system. This study focuses on the impact of the nano-oxides dispersion and the irradiation conditions on the formation of the alpha’ phase in an ODS steel: MA957 [4]. A MA957 steel and a corresponding steel without nano-oxides are irradiated on the JANNuS platform [5] at 400°C with Fe ions at different damage doses (from 0,5 to 9 dpa). Samples are characterized at the nanometric scale using Transmission Electron Microscopy (TEM), Energy Filtered TEM (EFTEM) and Atomic Probe Tomography (APT). The drastic impact of the nano-oxide distribution and the influence of injected ions on the alpha’ distribution are revealed and discussed. References: [1] R.L. Klueh et al., Journal of Nuclear Materials 341 (2005), p.103 [2] T.R. Allen et al., Journal of Nuclear Materials 375 (2008), p. 26 [3] N.A. Bailey, Journal of Nuclear Materials 459 (2015), p. 225 [4] J. Ribis et al., J. Mater. Res. 30 (2015), p.2210 [5] A. Gentils, C. Cabet, Nucl. Instrum. Methods B 447 (2019), p.107

Vitomir Sever *, Nicolas Bernier, Pierre Noé, Jean-Luc Rouvière

• Université Grenoble Alpes, CEA, LETI, F-38000 Grenoble

GeTe-Sb2Te3 superlattices (SL’s) are the most promising materials for iPCMs, but the physical switching mechanism is still under debate. To improve this emerging technology, it becomes mandatory to describe the iPCMs structures at an atomic scale since the involved mechanism upon switching is expected to rely on local atomic motion. Using atomic resolution HAADF STEM images, acquired on a FEI Titan Themis microscope, two analyses were performed. However, before studying the more complex case of iPCM SL’s, we explored the experimental approaches on a reference Si-SiGe multilayer sample calibrated by XRD. Various sets of data were acquired and analyzed using Python programming. First, images recorded with an HAADF detector were interpreted quantitatively by means of comparison between experimental and simulated intensities. We used the STEMsim software from the Rosenauer’s group [1] for HAADF image simulations. From this comparison, the local thickness can be obtained. If the thickness is known, the unknown chemical concentration of a single element can be measured. Here, PACBED measurements give the thickness and Ge concentration in SiGe layers was extracted [1]. Additionally, the position of atomic columns can be precisely determined utilizing a template-matching approach [2]. After acquiring a stack of low-dose images, rigid and non-rigid alignments were performed using the SmartAlign software. Then the center of gravity method is used to find the precise position of the correlation peaks between a template and an experimental image. Hence, a reference lattice can be calculated. Therefore, atomic displacements and strain maps were obtained on the same reference sample. For robust and fast analysis, two Python libraries have been developed and corresponding notebooks were deployed at the CEA Grenoble. After validating these methods on the reference sample, these analyses were applied to numerous HAADF images of iPCM SL’s and the first results on Sb2Te3 blocks will be presented. References: [1] Rosenauer et al., Ultramicroscopy 111 (2011), p.1316-27. [2] Zuo et al., Ultramicroscopy 136 (2014), p. 50-60

Grégoire Breyton *, Hakim Amara, Jaysen Nelayah, Damien Alloyeau, Guillaume Wang, Jérôme Creuze, Christian Ricolleau

• MPQ, Univ. de Paris
• LEM, CNRS / ONERA
• ICCMO, Univ. Paris Saclay

Nanoparticles (NPs) are primarily known for their high surface to volume ratio making them perfect candidate as catalyst for reactivity process. Another attractive aspect of these nano-objects lies in the unexpected reactivity of some materials. For example, gold is known to be inert at bulk scale but very reactive in small sized nanoparticles (< 5 nm) for the CO to CO2 transformation [1]. Interestingly, this phenomenon is enhanced by alloying gold with copper [2]. However, little is known on the surface structure of the Au-Cu alloy and more precisely of their surface composition which plays a crucial role during reactivity process. In this context, our work aims at characterizing the AuxCu1-x nanoparticles structure in order to highlight surface segregations or not by using both a theoretical and experimental approaches. Atomic scale Monte Carlo simulations based on semi-empirical potential [3] have been performed (Fig 1 Left). A strong segregation of gold over the entire surface is observed whatever the size of octahedron NPs (Fig 1 Middle). Simulations of nanocubes seem to strengthen these results where only (100) facets are present. Experimentally, octahedron NPs and nano-cubes have been synthesized by Pulsed Laser Deposition (PLD) on a substrate of NaCl. This substrate has a cubic structure which favors cubo-octaedron morphologies. Their chemical analysis and structural characterization on an aberration corrected JEOL ARM 200F in TEM and STEM modes are still ongoing (Fig 1 Right). References: 1. Lopez Nuria et al. Journal of Catalysis223.1 (2004): 232-235. 2. Liu Xiaoyan et al. Catalysis today 160.1 (2011): 103-108. 3. Rosato Vittorio et al. Philosophical Magazine A59.2 (1989): 321-336.

Farah Chafaï *, Loïc Patout, Khalid Hoummada, Lotfi Bessais, Najeh Mliki, Ahmed Charaï,

• Université de Tunis El Manar, Faculté des Sciences de Tunis, Laboratoire Matériaux Organisation et Propriétés, 2092, Tunis, Tunisie
• Aix-Marseille Université, CNRS, Université de Toulon, IM2NP - Campus de St Jérôme, 13397 Marseille cedex 20, France
• ICMPE UMR CNRS 7182 – Université Paris Est Créteil, 94320, Thiais, France

Les intermétalliques à base de terre-rares (RE) et d’éléments de transition sont parmi les meilleurs candidats pour une utilisation comme aimants permanents. Les composés nanocristallins ont été très étudiés notamment par rapport aux techniques de préparation et l’état approprié du matériau pour obtenir le champ coercitif optimisé [1]. En particulier, le composé Pr5Co19 constitue un alliage attractif dû à ses propriétés significatives donnant lieu à une grande variété d’applications technologiques telles que la réfrigération magnétique à haute température [2]. Quelques travaux montrent l’importance de réduire les défauts microstructuraux qui peuvent influencer localement les forces de pinning [3]. Une poudre nanocristalline Pr5Co19 de structure rhomboédrique a été préparée par fusion à arc et broyage à haute énergie. Les analyses MET montrent une structure en forme de plaquettes avec des alternances de contraste séparées par des distances non périodiques le long de l’axe c (c = 48,75 Å). Une projection HREM [100] tiltée le long de c* met en évidence des décalages d’empilement de part et d’autre des plaquettes sombres qui sont en partie intrinsèques à la structure (Fig. 1). En effet, l’analyse des empilements dans une image en axe de zone [210] et des réflexions FOLZ dans la FFT correspondante montrent que des fautes apparaissent dans la séquence des blocs A-B-C (ou 3R) (Fig. 2). La présence de polytypes, notamment les structures hexagonales Pr5Co19 et Pr2Co7, forment des structures d’inter-croissance où un bloc est retiré pour former localement des séquences bicouches de blocs A-B (ou 2H). Ces défauts semblent difficiles à éviter dû à la gamme de solubilité solide très limitée de la phase Pr5Co19 dans le système binaire et aux très faibles différences d’énergie entre les différentes stœchiométries et entre les structures stables 3R et métastables 2H. References [1] R. Fersi, et al.. J. Alloys Compd. 522 (2012) 14-18 [2] V. K. Pecharsky, et al. Phys. Rev. B, 64 ; 144406, 2001 [3] M. Duerrschnabel, et al. Nat. Commun. 8 (2017) 54

Temperature is a crucial parameter in the liquid-phase synthesis of metal nanoparticles (NPs) that directly impacts all the atomic-scale processes that drive the size dispersion of colloidal assemblies and the shape of nanostructures1-2. As the temperature concomitantly affects the kinetics of chemical reactions and the thermodynamic equilibrium of nanomaterials in solution, the full understanding of thermal effects on the nucleation and growth processes requires direct in situ observations at the nanoscale.   

Here, we exploit for the first time temperature controled liquid-cell TEM to study thermal effects on the radiolysis-driven formation of gold nanocrystals in water between 25 °C and 85 °C (Figure 1)3. The huge impacts of temperature on the nucleation and growth rates of nanostructures measured using automated video processing are quantitatively explained in the framework of the classical theories. Thus, we show that the increase of molecular diffusion and nanoparticle solubility governs the drastic changes in the formation dynamics of nanostructures in solution with temperature. In contradiction with the common view of coarsening processes in solution, we also demonstrate that the dissolution of nanoparticles and thus the Ostwald ripening are not only driven by size effects (Figure 2). Furthermore, visualizing thermal effects on faceting processes at the single nanoparticle level reveals how the competition between the growth speed and the surface diffusion dictates the final shape of nanocrystals. Our method and data-analysis workflow can be applied to other dynamical processes in nanochemistry where temperature plays an important role.

Acknowledgments:

We gratefully acknowledge the help of Maxime Moreaud in the processing of our videos using the PlugIM plateform developed at IFPEN. 

References:

[1] N. T. Thanh, N. Maclean and S. Mahiddine, Chemical reviews, 2014, 114, 7610-7630.

[2] X. Xia, S. Xie, M. Liu, H.-C. Peng, N. Lu, J. Wang, M. J. Kim and Y. Xia, Proceedings of the National Academy of Sciences, 2013, 110, 6669-6673.

[3] A. Khelfa, J. Nelayah, G. Wang, C. Ricolleau, D. Alloyeau, JOVE Vis. Exp. (168), e62225, doi:10.3791/62225 (2021).

Josephine REZKALLAH *, Simona MOLDOVAN, Bernhard WITULSKI, Xavier SAUVAGE

• GPM Laboratory, CNRS UMR 6634, Rouen University, Normandy
• LCMT Laboratory, CNRS UMR 6507, ENSICAEN & Caen University, Normandy.

Owing to their high reactivity, Pt nanoparticles (NPs) are commonly used in catalysis.[1,2] Shape, size, morphology and catalytic performance of NPs can be engineered by different synthetic methodologies;[3] and one of the strategies aiming at the active surface maximization is the use of hollow nanospheres.[4] Since catalytic reactions often occur under a well-controlled environment including conditions of high temperature and high pressure, a deeper knowledge on the catalytic materials behavior is necessary to understand, control and improve materials properties under reaction conditions. In this context, the approach proposed herein concern the use of Environmental TEM to probe microstructural changes in real time and under real-life conditions with notably atomic resolution. Two syntheses that differ each other in the well-defined ratio of Pt salts (H2PtCl6) and Co nanoparticles as precursors have been carried out for the elaboration of discrete Pt hollow nanospheres. [4] From a morphological point of view, the Pt-based NPs are spherical with a diameter of 15-20nm. They exhibit a core-shell structure with a polycrystalline shell (Fig. 1a and 2a). The chemical composition of these NPs was measured by EDX (Fig. 1(d-g) and 2(d-g)). It is strongly influenced by the Pt/Co ratio during the synthesis. The cobalt content of these hollow-sphere NPs produced from a Pt/2Co = 1.5 and a Pt/2Co = 0.75 ratio is 10% and 20%, respectively. Here we report on the influence of temperature, pressure and exposed gas (H2, Ar) on the morphological (coarsening, agglomeration and/or collapsing) and structural stability (phase transformation) of those hollow-sphered Pt nanoparticles. A windowed gas cell in-situ TEM holder (atmosphere-Protochips) was used in the 20-300°C temperature range and up to 1 atm pressure for this study. Acknowledgment The authors acknowledge the financial support of the French Agence Nationale de la Recherche and LabEx EMC3 through the Project ZeoMah (Grant No. ANR-10-LABX-09-01) References: [1] T.S. Ahmadi et M.A. El-Sayed, Science, 272 (1996), 1924. [2] Z. Zhou et Q. Xin, Chem.Commun., (2003). [3] T. Altantzis and S. Bals, Nano Letters, 19 (2019), 477. [4] H.‐P. Liang and C.‐L. Bai, Angew. Chem. Int. Ed., 43 (2004),1540.

Mariana Timm *, Rongrong Zhang, Karine Masenelli-Varlot, Jérémie Margueritat, Lucile Joly-Pottuz

• Univ-Lyon, INSA, UCBL, CNRS, MATEIS UMR 5510, 69621 Villeurbanne, France
• Institut Lumière Matière, Université de Lyon, Université Claude Bernard Lyon 1, UMR CNRS 5306, 69622 Villeurbanne, France

The mechanical behavior of materials at the nanometer scale is strongly different from the one observed in bulk materials [1]. However, the measurements at the nanoscale remain extremely challenging, and the same is true concerning their interpretation. Thus, we combine two different techniques, nanocompression and Brilloun spectroscopy, on the same system to reach a high characterization level. Initially, the elastic properties of monocrystalline Au nanoparticles with sizes around 100 nm was assessed with low-frequency Raman/Brillouin spectroscopy using the dependency of the acoustic vibrational modes. The measurement of acoustic vibrational modes is fundamental to the full comprehension of the mechanical properties of nano-objects, since they are connected to the intrinsic characteristics of the material, such as crystallinity, size, shape and elasticity [2]. We show that changes in the environment around the particles modify the distribution of the electrical field inside the NPs, therefore affecting their acoustic response. The same particles were submitted to nanocompression in the TEM to characterize their deformation behavior and to obtain the stress-strain curve. The in-situ nanocompression experiments were performed with a Cs-corrected FEI-Titan microscope using a Hysitron pico-indenter. Finally, using the pico-indenter sample-holder adapted to the Brillouin experimental set-up, the measurement of the acoustic vibrational modes of single Au NPs was performed during compression tests. References : [1] O. Kraft et al., Annu. Rev. Mater. Res 40 (2010) 293-317 [2] Hodak, J. H.; Martini, I.; Hartland, G. V. J. Phys. Chem. B 102 (1998), 6958−6967

Rongrong ZHANG *, Lucile Joly-Pottuz, Gaëtan Laurens, Tristan Albaret, Epicier Thierry, Karine Masenelli-Varlot

• Univ Lyon, INSA-Lyon, MATEIS
• Univ Lyon, UCB Lyon 1, ILM

As one of the most important ceramic materials, cerium oxide is widely used in many applications, such as in solid oxide fuel cell electrodes, catalysis, and is also used as superior abrasive particles in chemical mechanical planarization [1]. Ceria is sensitive to electron beam irradiation leading to a reduction [2]. Nevertheless, the use of the environment allows to control or avoid this reduction [publi Matthieu]. It is, thus, possible to test a CeO2 structure or a structure CeOx (1.5

Nicolas Folastre *, Kirill CHEREDNOCHENKO , François CADIOU , Matthieu BUGNET , Edgar RAUCH , Jacob OLCHOWKA, Stefan STANESCU, Sufal SWARAJ , Rachid BELKHOU , Antonella Iadecola, Christian MASQUELIER , Laurence CROGUENNEC , Arnaud DEMORTIERE

• Laboratoire de Réactivité et de Chimie des Solides (LRCS), CNRS UMR 7314, Amiens, France
• MATEIS Laboratory – CNRS UMR 5510/University of Lyon, Lyon, France
• Laboratoire Science et Ingénierie des Matériaux et Procédés (SIMaP) – Grenoble INP/CNRS/UJF, Saint Martin D’Hères, France
• Institut de Chimie de la Matière Condensée de Bordeaux (ICMCB), CNRS/Université de Bordeaux, Pessac, France
• SOLEIL Synchrotron - Science Division - Gif-sur-Yvette, Ile-de-France, France
• Laboratoire de Réactivité et de Chimie des Solides (LRCS), CNRS UMR 7314, Amiens, France
• Institut de Chimie de la Matière Condensée de Bordeaux (ICMCB), CNRS/Université de Bordeaux, Pessac, France

In situ liquid electrochemical TEM [1] is one of the most powerful analytical tools allowing to follow structural and chemical transformations in Li-ion battery active materials at nanoscale with high spatial resolution and in conventional liquid electrolytes. For instance, very recently, the changes in the unit cell structure of cathode materials (LFP) during electrochemical cycling in liquid electrolyte were determined using in situ electron diffraction tomography [2]. The same compound was earlier comprehensively studied by 4D-STEM ASTAR (parallel beam) [3] and other techniques as electron ptychography [4]. Sodium vanadium (III) fluorophosphate, Na3V2(PO4)2F3 (NVPF) is attracting great interest as a potential positive electrode for Na-ion batteries due to its exceptional rate and electrochemical cycling capabilities. During the charging process, the disinsertion of two Na+ ions leads to the change of vanadium oxidation state and corresponds to structural transformations which can be registered with 4D-STEM ASTAR system. In our work we employed Poseidon in situ electrochemical liquid cell TEM holder (Protochips) and TEM equipped with Oneview camera (GATAN) and 4D STEM ASTAR system. Thanks to 4D-STEM system we succeeded to follow the structural and, thus, phase transformations of NVPF grains during the charge/discharge process in coin-cell battery. The results of obtained phase mapping are in good agreement with corresponding galvanostatic curves and available X-ray diffraction synchrotron data. More recently, we mapped by STXM-XANES (SOLEIL, beamline HERMES) (energies of V and O) then by STEM-EELS (energies of Na) the same individual grains of NVPF de-sodiated ex-situ at different charge states[5] (Fig. 1). These results Added to 4D-STEM mappings will provide some answers concerning the (dis)-sodiation mechanisms in the NVPF material. Based on this multimodal study, we correlated structural and chemical behaviors in high resolution maps to improve reliability and draw evidence of Na diffusion pathway inside single grains. References: [1] Protochips, Micros. Today. 25 (2017) 13–13. [2] O.M. Karakulina, et al. , Nano Lett. 18 (2018) 6286–6291. [3] E.F. Rauch et al. Zeitschrift Für Kristallographie. 225 (2010). [4] Y. Jiang et al. Nature. 559 (2018) 343–349 [5] S. Adams, et a. Springer Berlin Heidelberg, Berlin, Heidelberg, 2014: pp. 129–159.

Julie Poulizac *, Adrien Boulineau, Emmanuel Billy, Karine Masenelli-Varlot

• CEA Grenoble
• MATEIS Lyon

Liquid in situ TEM is a powerful technique to observe various phenomena in their native environment such as nanoparticles nucleation and growth, alive biological samples or battery cycling, with high spatial and temporal resolution [1]. In this study, we applied this technique to investigate the dissolution kinetics of LiFePO4 cathode material for Li-ion battery recycling by hydrometallurgy. As the specimen studied here is a highly agglomerated powder, we applied the ultramicrotomy technique to prepare samples for the in situ experiment in order to control the sample size and avoid breaking the Si/SiN liquid cell. Thanks to this sample preparation method, we were able to observe in situ a complete dissolution of LiFePO4 by sulfuric acid, as shown in figure 1. We used BF-STEM technique to image the dissolution and the resulting images were post-processed with ImageJ software [2] to extract the reaction kinetics (figure 2). In parallel, ex situ dissolution experiments were done on LiFePO4 in order to obtain the ex situ reaction kinetics. Both ex situ and in situ kinetics were compared. Even though the in situ kinetics appears to be much slower than the ex situ one, by taking into account the samples geometry (spheres vs. disks) and the reacting surfaces, we show that the reaction kinetics are equivalent and could be overlaid. We demonstrate here that we are able to investigate in situ dissolution kinetics by the combination of the sample preparation method and image-processing routine. References : [1] S. Pu, C. Gong, and A. W. Robertson, “Liquid cell transmission electron microscopy and its applications,” R. Soc. open sci., vol. 7, Jan. 2020. [2] M. D. Abràmoff, P. J. Magalhães, and S. J. Ram, “Image Processing with ImageJ,” Biophotonics international, vol. 11, 2004.

Nathaly ORTIZ PEÑA *, Kondareddy CHERUKULA, Alberto BIANCO, Cecilia MENARD-MOYON, Florence GAZEAU, Damien ALLOYEAU

• Laboratoire Matériaux et Phénomènes Quantiques, UMR 7162 CNRS/Université Paris Diderot, 75013 Paris, France
• Laboratoire Matière et systèmes complexes, CNRS UMR 7057, Université de Paris, Paris Cedex 13, 75205, France
• CNRS, Institut de Biologie Moléculaire et Cellulaire, Laboratoire d'Immunopathologie et Chimie Thérapeutique, 67084 Strasbourg, France

In the booming of 2D materials for a variety of applications, transition metal dichalcogenides, in particular molybdenum disulfide, have raised as an attractive subject thanks to their exceptional electronic, optical, mechanical and chemical properties.1 Notably, the large surface area with tunable electronic properties, the intercalable layers and the readiness for functionalization makes them good candidates for biomedical proposes such as biosensing, bioimaging, and drug delivery, among others. Furthermore, MoS2 have shown to have a better biocompatibility and stability in comparison with their carbon-based analogues. Prior publications assessing the degradation products from the biotransformation of MoS2 point to the oxidation of the sheets leading to the formation of free Mo+4 ions and molybdenum oxides.2 Such degradation process appears to be innocuous for the cell life. Nonetheless, additional insight in the biotransformation mechanism of exfoliated MoS2 nanosheets is key in the assessment of its viability for biomedical applications. Herein, we have implemented a complementary approach by using in situ liquid phase transmission electron microscopy and ex-vivo studies to unveil a new aspect of the behavior of MoS2 patches in conditions mimicking those within cells, in particular those with high concentrations of ROS and H2O2. We have observed in direct the scrolling of MoS2 sheets in situ and verified the presence of the same structures ex vivo (Figure 1-2). Additionally, we determined that freestanding unrolled sheets can undergo oxidation and etching. The oxidized fragments were found in post-in situ samples and extra-cellular vesicles recovered from ex vivo experiments. Further studies enquiring on the effect of the scrolling in the stability of the nanoscrolls should be conducted in order to prove that this can constitute a protective mechanism of the structure. We would like to acknowledge the financial support of the French national research agency for the CYCLYS project. References: (1) Bazaka, K.; Levchenko, I.; Lim, J. W. M.; Baranov, O.; Corbella, C.; Xu, S.; Keidar, M. J. Phys. D. Appl. Phys. 2019, 51, 1–39. (2) Kurapati, R.; Muzi, L.; de Garibay, A. P. R.; Russier, J.; Voiry, D.; Vacchi, I. A.; Chhowalla, M.; Bianco, A. Adv. Funct. Mater. 2017, 27.

Nathaly ORTIZ PEÑA *, Charlotte DEJEAN, Benedicte MENEZ, Cyril GADAL, Damien ALLOYEAU, Alexandre GELABERT

• Laboratoire Matériaux et Phénomènes Quantiques, UMR 7162 CNRS/Université Paris Diderot, 75013 PARIS, France
• Université de Paris, Institut de physique du globe de Paris, CNRS UMR 7154, Paris, France

Recent direct observations in liquid phase of mineral nucleation and growth on bacterial cells highlighted, at the nanoscale, the critical role of chemical functions in cell surfaces and exopolymers in metal biomineralization [1]. Harnessing the possibilities of liquid-cell scanning transmission electron microscopy (LC-STEM), and using the incident electron beam to trigger and observe the precipitation of Mn-bearing minerals, differences in the morphology and distribution of Mn precipitates were observed. Differences in nucleation site density and accessibility on bacterial cell surfaces and exopolymers were correlated to the in situ observations. Herein, we use functionalized polystyrene beads of 1 µm diameter as simple analogues of bacteria cells, in order to better understand and characterize independently the role of the various chemical functions encountered on biological surfaces on Mn precipitation. Ten representative types of functionalization were selected for LC-STEM experiments, ranging from basic ones, like carboxylic groups (–COOH) or amine functions (–NH ), to chelating agents like nitrilotriacetic acid (NTA) or to protein compounds like streptavidin. This allows to characterize the specific impact of individual chemical function taken independently. Manganese mineralization was tracked for each type of functionalized beads, displaying significant difference in growth rates and mineralization patterns. As illustrated in figure 1, at the same electron dose rate, Mn precipitation for COOH-bearing beads is much faster (Figure 1b), compared to non-functionalized beads (Figure 1a) and, exhibit massive dendritic precipitates at the bead surface. Bulk measurements of electrophoretic mobility for each type of bead will be use to correlate surface charge to mineral growth rate on beads. The bulk surface charge explains only partially Mn mineralization properties at the beads surface but other parameters like steric effect need to be explored. We would like to acknowledge the financial support of the French national research agency for the MAMBA project. References [1] Couasnon, T., Alloyeau, D., Ménez, B., Guyot, F., Ghigo, J. M., & Gélabert, A. (2020). In situ monitoring of exopolymer-dependent Mn mineralization on bacterial surfaces. Science advances, 6(27), eaaz3125

Isabel Gómez-Recio *, Ram Kumar, Dris Ihiawakrim, Almudena Torres-Pardo, Jose M. González-Calbet, Ovidiu Ersen, David Portehault

• Sorbonne Université, CNRS, Laboratoire Chimie de la Matière Condensée de Paris (LCMCP), 4 place Jussieu, 75005 Paris, France
• Institut de Physique et Chimie des Matériaux de Strasbourg (IPCMS), UMR 7504 CNRS-Université de Strasbourg, BP 43 Strasbourg Cedex 2, France
• Dpt. of Inorganic Chemistry, Faculty of Chemistry, Universidad Complutense de Madrid, Madrid (28040) Spain

Metal oxides exhibit a wide range of properties combined to stability under ambient conditions that have given rise to an extensive range of applications. The decrease of oxides particle size at the nanoscale is well known to modify or enhance their properties. Developing synthesis methods is then important to widen the compositional range of oxides accessible at the nanoscale. Among several strategies, synthesis in molten salts is particularly appealing as a liquid-mediated pathway that allows extensive nucleation with limited grain growth to enable particle size control.[1] Nonetheless, the reaction mechanisms underlying molten salts syntheses are currently unknown. Real time observations of nucleation and growth steps can unveil the key mechanisms of nanocrystals formation, as well as provide guidelines towards selected morphologies. [2] Of particular interest is the search of self-assembly processes driving the crystallization and that could play an essential part of the toolkit for designing inorganic nanocrystals. In situ techniques, as in situ transmission electron microscopy, strongly contribute to the advancement of such knowledge, but have never been applied to the specific conditions of molten salts syntheses that encompass high temperatures and liquid phase. In this work the self-assembly is suggested to be one of the MgO crisytallization mechanisms in molten salts. Oriented attachment is a characteristic feature of colloidal syntheses, never observed before in molten salts medium. This growth mechanism gives rise to enthalpy drops in a thermodynamic spontaneous process due to crystal surface energy reduction. [3] TEM low magnification images evidence that the number of primary nanoparticles belonging to a cluster increase over the reaction time (Fig. 1), and that particle attachment occurs preferentially face-to-face. TEM images (Fig. 2) show the same crystallographic orientation at both sides of primary nanoparticles boundary. This information is complemented by in situ TEM studies which allow direct monitoring of MgO crystallization. For the in situ TEM study an environmental gas holder (Atmosphere from Protochips) was used.

Vinavadini Ramnarain *, Nathaly ORTIZ PEÑA, Dris IHIAWAWKRIM, Mariana LONGUINHO, Tristan GEORGES, Thierry AZAIS, Clément SANCHEZ, Ovidiu ERSEN

• Institut de Physique et Chimie des Matériaux Strasbourg (IPCMS)
• CBPF, Rio de Janeiro, Brasil
• Laboratoire de Chimie de Matière Condensé de Paris, Sorbonne Université, Paris, France
• USIAS, Université de Strasbourg, France

Biominerals have complex hierarchical structures, often combined with exceptional properties, such as high mechanical, electrical or magnetic properties, which are achieved under the direct control of biomolecules. This level of control originates from an intimate association of an organic matrix comprised of macromolecules and inorganic components in a localized environment of the organisms1. Thus, scientists seek inspiration from the properties of biominerals to produce advanced materials by means of biomimetic material design. The key factors which influence the final properties of the crystal are predominantly proteins of the organic matrix which may consist of various amino acids2. Whilst several studies have been carried out on the growth of biominerals controlled by macromolecules, the early steps before growth are not fully understood. Therefore, it is of great interest to study the impact of these amino acids on the early nucleation stages of biomineralization and to provide a direct mechanistic insight into these phenomena. Herein, we present the study of the mineralization processes of calcium carbonate. This system is chosen because of its abundance in nature and it is a relatively simple model. Furthermore, we have investigated the role of L-Aspartic acid on biomineralization. We have used the Liquid Phase Transmission Electron Microscopy to assess the incipient steps of mineralization within a liquid environment confined between two silicon nitride windows. In this study, we have been able to unravel, in real time, novel visual insights on the incipient steps, that is pre-nucleation and nucleation of CaCO3. In particular, we could experimentally determine the presence of pre-nucleation clusters and their dynamic behavior (Fig 1), Moreover we could observe a real time nucleation taking place where the pre-nucleation species are transformed to liquid nanodroplets upon crossing a binodal limit (Fig 2). The liquid nanodroplets further aggregated to form hydrated amorphous calcium carbonate by an annealing process. These results allowed us to understand the influence of organic matter during biomineral formation and hence establish a mechanistic picture. References: [1] Weiner, S.; Addadi, L. Annu. Rev. Mater. Res. 2011, 41 (1), 21–40. [2] Tavafoghi, M.; Cerruti, M. J. R. Soc. Interface 2016, 13 (123), 20160462

Ibrahim Koita *, Xiaoyan Li, Luiz H. G. Tizei, Jean-Denis Blazit, Nathalie Brun, Etienne Janod, Julien Tranchant, Benoît Corraze, Laurent Cario, Marcel Tencé, Odile Stéphan, Laura Bocher

• Université Paris-Saclay, CNRS, Laboratoire de Physique des Solides, 91405, Orsay, France
• Institut des Matériaux Jean Rouxel (IMN), Université de Nantes, CNRS, 2 Rue de la Houssinière, 44322 Nantes, France

V2O3 présente différentes transitions métal-isolant (TMIs) induites en température, sous pression ou par dopage chimique[1]. Ces TMIs présentent des scenarii complexes du point de vue de leurs évolutions électronique et structurale qui rendent encore leurs compréhensions incomplètes[2]. Lorsqu’il est refroidi en dessous de 160 K, V2O3 subit une transition structurale (rhomboédrique vers monoclinique) et une TMI avec une hausse de résistivité de 7 ordres de grandeur. Cette TMI activée en température a été largement étudiée à l’échelle macroscopique (Fig.1a)[1], récemment la coexistence de domaines isolant/métallique (I/M) a été cartographiée in situ par photoémission[3] et nano-IR[4] avec une résolution respective de 1 µm et de 25 nm au mieux. Or cette séparation de phases électroniques peut être due localement à des mécanismes structuraux d’où la nécessité de coupler cartographies structurale et électronique à l’échelle locale et sous températures variables. Ici nous avons réalisé des expériences STEM/EELS monochromatées de ultra-haute résolution spectrale sur le NION CHROMATEM 200 MC associé au porte-objet cryo double-tilt Henny Z refroidi à l’azote et utilisant des MEMS (Fig.1b) pour accéder à des conditions de températures variables à partir de 120 K. Lors des cycles thermiques à travers la TMI, les spectres EELS acquis dans le domaine des pertes proches présentent une signature plasmonique caractéristique à 1.1eV uniquement pour la phase métallique (Fig.2a)[5]. A température intermédiaire, la coexistence de nanodomaines I/M est cartographiée à l’échelle nanométrique (Fig.2b et 2c). Des analyses par K-means clustering confirment l’homogénéité électronique de chaque domaine (Fig.2d). En parallèle, des expériences 4DSTEM de nanodiffraction permettent de révéler l’évolution structurale des domaines I/M en température. Ces expériences visent à sonder localement les mécanismes physiques mis en jeu lors de la transition et en particulier la relation entre les degrés de libertés structuraux et électroniques. Les auteurs remercient l’EDPIF, l’ANR IMPULSE No. ANR-19-CE42-0001 et le programme de futur investissement TEMPOS No. ANR-10-EQPX-50. References: [1] McWhan, et al., PRB,2 (1970), 3734 [2] Kalcheim, et al., PRL.,122 (2019), 057601. [3] Lupi,et al., Nat. Commun.,1(2010),105. [4] McLeod, et al., Nat. Phys.,13 (2017),80. [5] Abe, et al., Jpn. JAP., 37 (1998), 584

Kilian Gruel *, Raphaël Serra, Aurélien Masseboeuf, Christophe Gatel *, Martin Hÿtch *

• CEMES-CNRS, Toulouse, France
• Université de Toulouse, Toulouse, France

Electron holography has the potential to become a powerful technique to measure electric fields in microelectronics devices such as capacitors during their operation. However, we first need to show that measurements can be performed quantitatively, and in operando conditions, for model systems [1]. The experimental results need to be compared with modeling that takes into account factors such as specimen geometry, FIB damage, stray fields and charging [2]. The model dielectric capacitor that has been studied experimentally is composed of an insulating layer of silicon nitride (Si3N4) between a bottom electrode (substrate) of highly P-doped silicon (Si) and a top electrode of deposited titanium (Ti). Electron holography experiment has been performed on this sample using the I2TEM microscope (Hitachi HF3300-C) operating at 300 kV whilst applying different voltages from – 10V to 10V. Finite element method (FEM) was used to model the electrostatic potential (COMSOL Multiphysics). Parameters were adjusted to produce an excellent fit with the experimentally measured phase shift (Figure 2). The curvature of the phase in the electrodes is well reproduced and is caused by the stray field around the sample. In the silicon nitride, there is a slight disagreement with the experimental curvature and can be modelled by introducing charges in the insulating layer. This work was supported by the French national project IODA (ANR-17-CE24-0047). The research leading to these results has received funding from the European Union Horizon 2020 research and innovation programme under grant agreement No. 823717 – ESTEEM3. Références/References : [1] AC Twitchett et al., Phys. Rev. Lett. 88 (2002), p. 238302 [2] S Yazdi et al., Ultramicroscopy 152 (2015), p. 10

Charles Sidhoum *, Mateusz Odziomek, Dris Ihiawakrim, Doru Constantin, François Schosseler, Clément Sanchez, Ovidiu Ersen

• Institut de Physique et Chimie des Matériaux de Strasbourg, 23 rue du Lœss, BP 43, Strasbourg Cedex 2, France
• Laboratoire de Chimie de Matière Condensé de Paris, Sorbonne Université, Paris, France
• Institut Charles Sadron, UPR22, 6 rue Boussingault, 67083 Strasbourg cedex, France

The electrochromic and photocatalytic properties of tungsten oxide (WO3) find broad applications in display devices, smart windows or degradation of pollutants. The synthesis of WO3 does not require high temperatures, and soft methods fitting into “chimie douce” concept are well-established. Following the sol-gel routes, it is possible to obtain WO3 gels by acidification of tungsten salt solutions. The acidification leads to hydrolysis and consecutive condensation first to polyoxotungstate species and further to gel-like structure. Although the process has been studied by various techniques1, we still lack the understanding on how the charged polyoxometallate species react with each other and form a gel. Several models have been proposed, based on the condensation of a hypothetical “zero-charge complex” by hydroxylation and oxolation2. The recent developments in TEM, in particular the emergence of the in-situ Liquid Phase TEM (LP-TEM), may provide new insight into the nucleation and growth processes of WO3-based nanostructures, with a particular focus on the early steps of gel formation. Herein, using the ability of LP-TEM to monitor nanometric processes in real space, we report a first picture of the growth of WO3 gels. This in situ analysis highlights several possible pathways for the growth mechanism, including the diffusion-limited cluster aggregation process. Indeed, a closer look into the reaction environment (Figure 1) reveals the presence of well-defined clusters, with a nanometric size, which apparently plays the role of elemental bricks for gel formation3. To further improve the comprehension of the first growth events, we have also carried out time-resolved DLS measurements for different tungsten salt concentration. As expected, a decrease in concentration (Figure 2b) leads to a decrease in intensity but, at the same time, we find similar sizes (Figure 2a), indicating the presence of precursors of few nm. In complement to DLS data obtained during the first 30 minutes, cryo-TEM studies are ongoing in order to complete the mechanistic picture and to provide a complete description of the growth process. References: [1] Pope, M. T.Heteropoly and Isopoly Oxometalates;Springer:Berlin, 1983. [2] J-P. Jolivet. (2015). De la solution à l’oxyde. EDP Sciences [3] Chemseddine, A. J. Solid State Chem. 2008, 2731–2736.

Suzanne Vernier *, Pierre Joly, Lydia Laffont, Eric Andrieu

• CIRIMAT - INPT
• FRAMATOME
• CIRIMAT - INPT/ENSIACET

L’acier 20MND5 (0,2%C-1,4%Mn-0,7%Ni-0,5%Mo) est un acier faiblement allié utilisé dans l’industrie nucléaire pour fabriquer des générateurs de vapeur. Ces pièces en 20MND5 sont obtenues par forgeage de lingots de très grandes dimensions qu’il est difficile d’homogénéiser chimiquement. Ainsi, des micro-ségrégations, héritage des ségrégations entre bras de dendrite secondaires, sont classiquement observées dans ces pièces. Les micro-ségrégations sont responsables de la microstructure en « bandes » obtenue après un refroidissement de l’acier à vitesse modérée [1] : des bandes de bainite supérieure (matrice non ségrégée) alternent avec des bandes de bainite inférieure (micro-ségrégation) (Figure 1). La structure en bandes de l’acier 20MND5 est suspectée d’être à l’origine d’une dispersion des valeurs de résilience à basse température, en particulier dans la zone de transition ductile-fragile. Or, l’étude de la microstructure a montré que les grains parents, c’est-à-dire les anciens grains austénitiques, sont plus gros dans les micro-ségrégations que dans la matrice non ségrégée. Cette plus grande taille de grains austénitiques combinée à la ségrégation des éléments d’alliage et des impuretés pourrait expliquer une plus grande fragilité des micro-ségrégations à basse température. Afin de comprendre les raisons d’une telle croissance des grains austénitiques dans les micro-ségrégations, l’évolution de la taille des grains parents a été suivie dans chaque type de zone au cours d’un traitement d’austénitisation (montée lente puis maintien en température – Figure 2.a). Des images en microscopie optique et électronique, des essais de micro-dureté et des analyses EBSD ont été réalisés à différentes étapes du traitement (Figure 2). En particulier, les analyses EBSD ont permis de reconstruire les cartographies des grains parents en se basant sur la relation d’orientation entre grains enfants et grains parents [2]. Ainsi, il est apparu que l’écart de taille de grains entre les deux types de zone se creuse une fois la transformation austénitique terminée, lors du maintien en température. Des nitrures d’aluminium de taille nanométrique pourraient être à l’origine du ralentissement de la croissance des grains dans la matrice non ségrégée. Références : [1] R.A. Grange, Metall. Trans., vol. 2 (1971), p. 417 [2] T. Nyyssönen, Metall. Mater. Trans. A, vol. 49 (2018), p. 6426

Emre Yörük *, Holger Klein, Stéphanie Kodjikian

• CNRS Néel Institute, University of Grenoble Alpes

3D electron diffraction (3D ED) has recently emerged as an alternative to x-ray diffraction to elucidate atomic structures of nano-sized crystals [1]. High scattering cross-sections of electrons allows for a diffraction data quality on par with synchrotron x-rays, albeit at a reduced cost. However, electron beam damage is still a major issue in the 3D ED field [2], and exposure conditions need to be optimized for beam sensitive compounds. LD-EDT [3] is a low dose 3D ED technique for ab initio structure determination of beam sensitive crystals such as hydrated minerals or MOFs. High quality diffraction data can be obtained from individual single crystalline particles in a powder without damaging the structure, and a precise sampling of the reciprocal space is assured by beam precession. We recently applied LD-EDT on Bulachite [4], a hydrated Al arsenate mineral, and solved its atomic structure. Difficulties related to the small size of crystals as well as beam sensitivity due to the presence of H2O molecules inside the lattice were overcome by LD-EDT, where synchrotron x-ray powder diffraction previously failed. The resulting structure [5] is comprised of layers containing edge-sharing Al-O octahedra, inter-connected with As-O tetrahedra by corner sharing. The localization of light atoms in the lattice showcases the potential of electron crystallography for yielding high quality diffraction data even under low dose conditions. References : [1] M. Gemmi et al., ACS Cent. Sci., 5 (2019), p. 1315−1329. [2] J. Hattne et al., Structure, 26 (2018), p. 759-766. [3] S. Kodjikian, H. Klein, Ultramicroscopy, 200 (2019), p. 12-19. [4] K. Walenta, Aufschluss., 34 (1983), p. 445-451. [5] I.E. Grey., E. Yoruk et al., Mineralogical Magazine, 84 (2020), p. 608-615.

Christian Turquat *, Mateusz JĘDRUSIK, Łukasz CIENIEK, Agnieszka KOPIA, Christine LEROUX

• Université de Toulon, AMU, CNRS, IM2NP (UMR 7334), CS 60584, Toulon, F- 83041, France
• Wydział Inżynierii Metali i Informatyki Przemysłowej, Akademia Górniczo – Hutnicza, al. Mickiewicza 30, 30-059 Kraków, Polska

In this paper, our attention has been focused on lanthanum iron oxide (LaFeO3) as thin films because of its potential in silicon microelectronics-based gas sensing technologies [1]. More precisely, the purpose of this research is to investigate the influence of process temperature on the morphology and crystallographic structure of LaFeO3 (LFO) thin films obtained by a Pulsed Laser Deposition (PLD) method on silicon (Si) substrates, in particular the nature of the exposed crystallographic facets. Three deposition temperatures, 750°C, 850°C, and 1000°C were investigated. Morphologic information was secured via Scanning Electron Microscopy and Atomic Force Microscopy while crystallographic information was gained via grazing incidence X-ray Diffraction and Transmission Electron Microscopy. At all temperatures, the films are crystallized in the Pnma orthorhombic structure. Thin films with a temperature deposition of 750°C and 850°C exhibit a microstructure of columnar grains with different LFO thicknesses, respectively 80 and 150 nm, and a similar intriguing surface structuration (see Figure 1). However, films grown at 750 °C show cracks, which disqualify them for applications. Two types of columnar grains are observed, with a tip termination and a columnar width from 15 to 20 nm, and broader flat terminated grains (see Figure 2). They correspond to two different growth direction [101] (flat) and [200] (tip), but in both cases {101} facets are exposed. The microstructure of the sample with a temperature deposition of 1000°C contrasts with the lower temperature samples and shows an active layer thickness of around 130 nm with a random arrangement of grains. HREM evidenced a reaction between the silicon substrate and the LaFeO3 layer, with the formation of La2Si2O7, indicating diffusion of Si through the native SiO2 barrier. This phenomenon was also observed in LaCoO3/Si [2]. These disparities will be discussed in terms of the structure and composition variation with the LFO layer. Références/References: [1] H. Zhu, P. Zhang, S. Dai, ACS Catalysis, 5 (2015) 6370-6385. [2] M. Jedrusik, Ł. Cieniek, A. Kopia, Ch. Turquat, Ch. Leroux, Arch. Metall. Mater. 65 (2020), 793-797

Romain Facchinetti *, Cyril Langlois, Sophie Cazottes, Thierry Douillard, Claire Maurice, Juliette Chevy, Christine Nardin

• MATEIS – Institut National des Sciences Appliquées (INSA Lyon) - Lyon – France
• Mines Saint-Etienne, Univ Lyon, CNRS, UMR 5307 LGF, Centre SMS, F-42023 Saint-Etienne France
• Constellium – Voreppe - France

Understanding the relationship between microstructure and properties is a necessary step for the improvement of existing materials or the development of new materials. The microstructure can be characterised using orientations maps, from which information like grain size distribution and deformation state can be extracted. Apart from EBSD (Electron Backscattered Diffraction), a promising approach to obtain orientation maps with the SEM (Scanning Electron Microscope) is the eCHORD method [1], which is based on electron channelling contrast. This method has the advantage that the sample tilt is only about 10°, and that the accelerating voltage can be lowered to a few kV, resulting in an improved spatial resolution. A series of BSE images is recorded during the rotation of the sample around its tilted normal direction, from which the intensity profiles of each pixel can be extracted as a function of the rotation angle. By comparing the obtained intensity profiles with the one calculated from ECP (Electron Channelling Patterns) simulations [2], the orientation of each pixel can be determined. The angular resolution achieved on Aluminium alloys is about 0.1°. In addition to orientation maps, we present here an original method to characterize the local misorientation using eCHORD. One objective is to determine the fraction of recrystallized grains in a partially recrystallized Aluminium. In fact eCHORD identify the internal misorientation by comparing the distances between profiles of adjacent pixels, without the need to compute pixels orientation. The results are compared with EBSD Kernel maps obtained on Aluminium with different levels of strain and different fractions of recrystallized grains. Références/References : [1] C.Lafond, T.Douillard, S.Cazottes, P.Steyer, and C.Langlois. Electron CHanneling Orientation Determination (eCHORD): An original approach to crystalline orientation mapping. Ultramicroscopy, 186 :146–149, (March 2018). [2] Singh, S., Ram, F., & De Graef, M. EMsoft: Open source software for electron diffraction/image simulations, Microscopy and Microanalysis, 23(S1), 212-213. (July 2017)

Nicolas Gautier *, Anne-Claire Gaillot

• Institut des Matériaux Jean Rouxel (IMN), UMR6502 CNRS, 2 rue de la Houssinière BP32229, 44322 Nantes cedex3, France

Depuis une dizaine d’années, la résolution de la structure d’un monocristal par diffraction électronique s’est fortement développée [1]. Différentes techniques ont été mises au point, basées sur la diffraction électronique en mode tomographie (EDT), associée ou non à la précession électronique, permettant un meilleur échantillonnage du réseau réciproque en ne se limitant plus aux principaux axes de zones [2]. Des programmes tels que PETS2, implémenté dans Jana2006, permettent de résoudre et d’affiner les structures cristallines à partir des données de diffraction électronique en mode tomographie avec précession (PEDT) en prenant en compte également les effets dynamiques [3]. Mais les techniques classiques d’EDT sont difficilement applicables aux cristaux très sensibles au faisceau d’électrons. Des conditions d’acquisition spécifiques doivent être mises en œuvre pour limiter le courant du faisceau d’électrons (« low dose ») ou augmenter la vitesse d’acquisition pour limiter le temps d’exposition. C’est ce que permet le script “Continuous Rotation movie acquisition” (CRmov) intégré au programme libre SerialEM [4]. CRmov permet de contrôler la platine goniométrique pour tourner le porte-objet à l’angle de départ, collecter une série de clichés de diffraction en continu jusqu’à l’angle final et sauvegarder les données automatiquement en quelques minutes. Les données acquises sur un monocristal peuvent ensuite être traitées à l’aide de logiciels tels que PETS2 et JANA, afin de déterminer les paramètres de maille du composé, jusqu’à la résolution structurale. Installé sur Nant’Themis (S/TEM ThermoScientific Themis Z), ce dispositif nous a permis grâce à cette rapidité d’acquisition, de collecter des données de diffraction électronique en mode tomographie avec précession (PEDT) sur des échantillons organiques. Il a également été mis à profit sur des échantillons moins sensibles sous le faisceau, pour enregistrer dans un temps limité plusieurs jeux de données dans des conditions expérimentales différentes. Références : [1] Gemmi et al. ACS Central Science 2019, 5 (8), p.1315– 1329, [2] Kolb et al., Ultramicroscopy 107 (2007) 507–513 [3] Palatinus, L., Petricek, V. & Correa, C. A. (2015). Acta Cryst. A71, 235-244 [4] De la Cruz et al., Ultramicroscopy, 201 (2019), 77-80

Priscille Cuvillier *, Eric Derniaux, Rémi Mercier, Yannick Thébault, Marylou Boisson, Roch Menand, Virginie Boulay

• EDF – DI, CNPE de Chinon - BP23 - 37420 Avoine, France
• EDF – DI, 2 rue Ampère – 93206 Saint-Denis, France
• EDF UNIE, 1 place Pleyel 93200 Saint-Denis, France
• EDF – DT, 19 rue Pierre Bourdeix – 69007 Lyon, France

Radiation-Induced Segregation (RIS), causing in particular Cr depletion at grain boundaries (GB), is a phenomena suspected in the embrittlement of nuclear component during service [1]. RIS in a bolt stayed 26 years under neutron irradiation in a civil nuclear reactor (PWR) has been studied by TEM using a FEI TECNAI Osiris equipped with a energy dispersion X-ray spectrometers (four SDD detectors). This is the first time EDF investigates a deposited baffle-former bolt from PWR at such levels of irradiation (≈ 45dpa) [2] thanks to new sampling preparation [3] using a FIB-SEM Lyra3 TESCAN in a high-activity cell [4] making it possible to prepare remotely lamellae from an initially highly radioactive sample. The bolt, with no indication of cracking, and its anti-rotation lock-bar (Figure 1) are respectively in a 316L and 304L austenitic stainless steel. As RIS is subjected to be affected by several parameters such as dose, temperature, material composition or crystallographic orientation of the GB, three samples at different pairs of dose/temperature, in different material and for different disorientation of GB have been studied: 45dpa/295°C (304L anti-rotation strip), 40dpa/305°C (316L bolt head) and 24dpa/340°C (316L bolt shank). Depletions of Cr, Fe, Mn and Mo and enrichment of Ni and Si have been observed at GB and maximum Cr depletion reaches ΔCr≈6 wt% (Figure 2). Some GB decorated with precipitates (probably γ’ (Ni3Si) or G (M6Ni16Si7) phases) are also observed. RIS is larger for the pair “higher” dose / “lower” temperature and seems to be more pronounced for highly disoriented GB and more disadvantageous for 316L than 304L (initially richer in Cr). Further work on higher dose and cracked bolts is planned to better understand the evolution of RIS and its influence on GB embrittlement. References: [1] A. J. Ardell and P. Bellon, Solid State and Materials Science 20 (2016), p. 115-139 [2] É. Fargeas et al, Fontevraud 9, (2018) [3] L. Legras et al., European Microscopy Congress (2016) p. 443-444 [4] S. Miloudi et al., European Microscopy Congress (2016) p. 368 The authors would like to thank all the persons of EDF hot laboratory that contributed to the examinations.

Guillaume Noircler *, Fabien Lebreton, Etienne Drahi, Patricia De Coux, Bénédicte Warot-Fonrose

• CEMES-CNRS
• Total SA
• Institut Photovoltaïque d’Ile-de-France (IPVF)
• LPICM

Amorphous aluminum oxide (a-AlOx) capped with amorphous silicon nitride (a-SiNx:H) is nowadays a common passivation stack for p-type silicon surfaces and is a good candidate to passivate the rear side of two-terminal perovskite/silicon tandem cells. The combination of atomic layer deposition (ALD) for a-AlOx with plasma-enhanced chemical vapor deposition (PECVD) for the capping made its success. Chemical passivation is insured by hydrogen from a-SiNx:H that can easily diffuse to the c-Si/AlOx interface. The field effect passivation is provided by a-AlOx which contains a high negative fixed charge density. The origin of the latter one is still a matter of debate and therefore requires additional research to be more controlled. 3 different samples have been studied with different process for the annealing step and a-SiNx:H composition (figure 1). In an attempt to better understand these negative fixed charges, local scale investigation by STEM-EELS has been performed at the c-Si/a-AlOx interface on 3 different samples. However, the extreme sensitivity of a-AlOx and a-SiNx:H to the electrons beam has led us to carry out a detailed study of the electron irradiation process. The c-Si/a-AlOx interface can undergo several electron-beam irradiation damage like sputtering, knock-on or radiolysis which have a direct impact on the structure and the chemical composition if precautions are not taken. We have shown that the radiolysis process was the dominant radiation damage and that it could drastically be limited by reducing the acceleration voltage and by modifying the scan orientation. Once the radiation damage has been understood and brought under control, STEM-EELS analysis has been done using Si and Al L2,3 and O K edges to determine the composition of the c-Si/a-AlOx interface (figure 2). Just above the c-Si, an ultrathin a-SiOx layer was found. Between the latter and the a-AlOx, a non-stoichiometric aluminum silicate has been characterized containing tetrahedrally coordinated Al in its first layers. The tetrahedral coordination of Al is known to have a net negative charge and high catalytic activity. Thus, this coordination could strongly participate in the production of negative fixed charge which improves the field effect passivation of this amorphous stack.

Nathalie Brun *, Guillaume Lambert, Laura Bocher

• CNRS Université Paris Saclay Laoratoire de Physique des Solides
• Université Paris Saclay M1 Mathématiques Appliquées

Le développement de la microscopie électronique analytique en mode STEM banalise l'acquisition de spectres-images représentant un grand volume de données, de l'ordre de 10k pixels x 1000 canaux en énergie. Il est possible d'appliquer les techniques classiques de traitement des données EELS à chaque spectre. Toutefois il est préférable de considérer le spectre-image comme un tout et d'exploiter le caractère redondant de l'information. Les techniques d'analyse statistiques commencent ainsi à être couramment utilisées, notamment l'analyse par composantes principales (ACP) mise en œuvre par l'intermédiaire de la Toolbox Hyperspy [1]. Néanmoins l'ACP est en général utilisée pour débruiter le spectre-image, auquel on applique ensuite le processus classique (soustraction du fond continu/somme des canaux correspondants au seuil caractéristique). En effet le but du traitement est de passer d'un cube de données difficilement visualisable à des données 2D exploitables sous forme de cartes. Le moyen d'obtenir directement ces cartes en exploitant pleinement l'information contenue dans le spectre-image est d'utiliser des techniques de démélange spectral [2]. Le démélange spectral repose sur l'hypothèse que chaque spectre individuel d'un spectre-image peut être décrit comme une combinaison linéaire d'un petit nombre de n spectres représentatifs des composants présents dans l'échantillon. Le spectre-image peut être alors complètement décrit par ces n composantes et les n cartes correspondantes. Outre d'être souvent beaucoup plus rapides, ces méthodes ont l'avantage de s'affranchir de problèmes tels que la superposition de seuils ou la difficulté de soustraire le fond continu, voire de cartographier indirectement certains éléments dont le seuil n'est pas détecté avec la dispersion utilisée. Je montrerai des exemples obtenus avec différentes techniques de démélange; on s'intéressera en particulier à l'intérêt de l'utilisation du Deep Learning pour ce type de problème. Références: [1] https://hyperspy.org/index.html [2] Dobigeon, N., & Brun, N. Spectral mixture analysis of EELS spectrum-images. Ultramicroscopy, 120 (2012) p.25-34. Remerciements : nous remercions André Thiaville (LPS, Orsay), William Legrand, Nicolas Reyren et Vincent Cros (UMP CNRS/Thales, Palaiseau) pour l’échantillon Pt/Co/Ru.

Yves Auad *, Cyrille Hamon, Marcel Tencé, Vahagn Mkhitaryan, Odile Stephan, Javier Garcia de Abajo, Luiz Galvão Tizei, Mathieu Kociak

• Laboratoire de Physique des Solides, Orsay
• ICFO, Barcelona

Whispering gallery mode resonators confine light using total internal reflection and take advantage of its several and narrowband circulating resonances for applications in optomechanics [1], quantum electrodynamics [2] and sensing [3]. The spherical symmetry and the low radiation loss make them difficult to study under free-space light. The ability of fast electrons to excite confined light modes in the near-field of photonic objects paves the way to probe these circulating modes with high spatial resolution. Electron Energy Loss Spectroscopy (EELS) has already been performed in small ( 1 μm) silica spheres with high quality factors. References: [1] Tobias Kippenberg et al. Analysis of radiation-pressure induced mechanical oscillation of an optical microcavity. Physical Review Letters, 95, 2005. [2] D. Vernooy et al. Cavity qed with high-q whispering gallery modes. Physical Review A, 57, 1998. [3] Yu Zheng et al. Sensing and lasing applications of whispering gallery mode microresonators. Opto-Electronic advances, 1, 2018. [4] Jerome Hyun et al., Measuring far-ultraviolet whispering gallery modes with high energy electrons. Applied Physics Letters, 12 2008. [5] Ofer Kfier et al. Controlling free electrons with optical whispering-gallery modes. Nature, 2020.

Alessandro PUGLIARA *, Lydia LAFFONT, Teresa HUNGRIA, Claudie JOSSE, Jacques LACAZE

• CIRIMAT-ENSIACET, Université de Toulouse, Toulouse, France
• Centre de Microcaractérisation Raimond Castaing, Université de Toulouse, UPS – CNRS – Toulouse INP – INSA, Toulouse, France

Spheroidal graphite cast irons are alloys with excellent ductility (mechanical characteristics close to those of mild steels). In these alloys, spheroidizing and inoculation treatments, using Fe-Si-Mg and Fe-Si alloys respectively, interact to give the final particles acting as nuclei for generating the spheroidal graphite nodule [1]. Many observations of these nuclei reported the presence of multi-phase nanometric structure containing oxides, sulfides, oxysulfides [1], and sometimes nitrides [2, 3]. Accordingly, the chemical characterization of nuclei is of prime importance to establish the sequence of phase transformations leading to the formation of spheroidal graphite nodules. Scanning transmission electron microscopy (STEM) coupled with energy dispersive X-ray spectroscopy (EDS) is carried out on a thin diametric section of a graphite nodule. Bright field (BF) STEM image shows the presence of a nucleus in the middle of the graphite nodule (Fig. 1a). This nucleus appears composed of two parts: a round head loosely connected to an elongated tail (Fig. 1b). High-angle annular dark-field (HAADF) STEM images evidence the presence of at least one element much heavier than carbon in both parts (Fig. 1c). Moreover, the presence of different contrasts in the head indicates a multi-phase nature with a fan-like faceted precipitate (Fig. 1d). STEM-EDS cartography shows that the tail is composed of Al, Mg, Si and N, with however a few isolated small iron rich precipitates. In contrast, the head exhibits at least four different phases: 1) an inner facetted center rich in Fe; 2) a shell rich in S and Mg surrounding the center; 3) the fan-like precipitate mainly composed of Ti and developing outwards from this shell; and 4) a structure having the same composition as the elongated tail (Fig. 2). In this study, transmission and analytical electron microscopy allowed determining an organized multi-phase graphite nucleus at nanometric scale in a spheroidal graphite cast irons and a formation sequence could be proposed [4]. References: [1] T Skaland, Metall Trans A, 24A (1993), p. 2321 [2] H Nakae, Mater Trans, 43 (2002), p.2826 [3] JK Solberg, Mater. Sci. Tech., 17 (2001) p.1238 [4] L Laffont, J Mater. Res. Technol., 9 (2020) p.4665

Eric Maximilian Weikum *, Georges Beainy, Jonathan Houard, Ivan Blum, Simona Moldovan, Jean-Michel Chauveau, Maxime Hugues, Denis Lefebvre, Nolwenn le Biavan, Angela Vella , Lorenzo Rigutti

• UNIROUEN, CNRS, Groupe de Physique des Matériaux, Normandie Université, 76000 Rouen, France.
• Centre de Recherche sur l’Hétéro-Epitaxie et ses Applications, UPR10 CNRS, 06560 Valbonne, France.

The Photonic Atom Probe (PAP) [1] uses the pulsed laser of a Laser-assisted Atom Probe Tomograph (La-APT) to excite electronic states in semiconductor materials. The radiation caused from the relaxation of these excited electronic states is analyzed by a Photoluminescence (PL) set-up. Simultaneously the sample can be field-evaporated by the La-APT set-up. This destructive analysis method offers compositional information on the atomic scale. The combination of both analysis methods allows to correlate the PL spectra with the evaporation state of the sample. This instrument allows performing super-resolution PL analysis correlated with 3D chemical mapping [2] and stress measurements on tip specimens [3]. In this contribution we report about the effect of the dynamical development of the tip specimen morphology during evaporation on intensity of the collected PL. The experiment shows indeed that the collected PL intensity from a single MgZnO/ZnO-QW (quantum well) shows a dependence on the amount of field-evaporated atoms. Using the FTDT method of the LUMERICAL software (Fig. 1), the emission of PL was simulated for different evaporation states of the tip and compared to the experiment. The calculation was performed taking into account the microscopic APT-reconstruction of the tip, which allows determining the location, the composition and the geometry of the light-emitting quantum well (Fig. 2). The numerical results show that the change in the tip’s geometry, due to the removal of atoms by field evaporation, leads to a change in the spatial distribution of the emitted light, resulting in a change in the measured PL intensity. A good agreement between experimental and numerical results is observed. In addition to the analysis of the intensity as a function of the evaporation state, the Polarization dependence of the measured data was compared to the simulated results. It was found experimentally that the emission from the ZnO-QW shows a different favored Polarization direction (σ) than the MgZnO-barrier (π), which is not to be expected, since the ZnO-QW shares its Wurzite crystal structure with the MgZnO-barrier. References [1] Review of Scientific Instruments 91.8 (2020): 083704. [2] Nano Letters 20.12 (2020): 8733-8738. [3] Physical Review Applied 15.2 (2021): 024014.

Lydia Laffont *, Adrien Barroux, Thomas Duguet, Nadège Ducommun, Eric Nivet, Julien Delgado, Christine Blanc

• CIRIMAT/Toulouse INP-Ensiacet
• CIRIMAT/CETIM Nantes
• CETIM Nantes

L’acier inoxydable martensitique 17-4PH (17%Cr, 5%Ni, 5%Cr) est utilisé dans l’industrie nucléaire, aéronautique et biomédicales en raison d’une bonne résistance à la corrosion intergranulaire et de très bonnes propriétés mécaniques (précipités nanométriques riches en Cu). Ces pièces sont obtenues par forgeage (REF) mais depuis quelques années par fusion laser sur lit de poudre (procédé de fabrication additive métallique -FA) [1]. La microstructure de ces pièces obtenues après le traitement thermique de référence H900 est équivalente mais présente néanmoins des différences en ce qui concerne la quantité d’austénite, la taille des lattes de martensite et celle des précipités de NbC (Figure 1) [2]. Dans cette étude, la spectroscopie de photoélectrons (XPS) et la microscopie électronique en transmission (MET/STEM) associée à la spectroscopie de dispersion en énergie des rayons X (EDS) ont été utilisés afin de caractériser le film passif nanométrique formé après immersion en milieu NaCl sur l’acier obtenu par FA en comparaison à l’acier de référence. Ce film présente une structure duplex constituée d’hydroxydes/oxydes de Fe et de Cr suivie d’une couche intermédiaire enrichie en Cr, Cu et Ni. La couche d’oxyde externe est composée principalement d’oxy/hydroxydes de fer alors que la couche interne est enrichie en chrome au niveau de l’oxyde de fer (Figure 2). Des différences significatives au niveau de l’épaisseur et de la distribution en chrome dans le film passif ont été observées pour les deux aciers. De plus, un film d’oxyde particulier est formé sur les précipités de NbC présents à la surface. Ces différences de composition chimique et de structure au niveau du film passif pour ces deux aciers permettent d’expliciter les différences de comportement en corrosion obtenues en ce qui concerne la résistance à la corrosion par piqûres [3]. Références : [1] X. Lou J. Nucl. Mater. vol. 499 (2018) p.182 [2] A. Barroux, Corr.Sci., vol. 169 (2020), p. 108594 [2] A. Barroux, Surf Int., vol. 22 (2021), p. 100874