New developments in the modeling and analysis of radiation damage in materials II

Resume : Experiments at large-scale facilities as ion accelerators allow the simulations for materials of real working conditions such as radiative environments ranging from nuclear to space applications. Indeed, swift heavy ion irradiation leads mainly to the deposition of intense electronic excitations entailing defect formation and structural modifications that depend on material characteristics. Since few years the development of versatile laboratory X-Ray diffractometers allows high resolution in grazing geometry being suitable for the irradiated material characterizations where the upper irradiated part of samples might be strongly and non-homogeneously modified in depth. Damage depth profiles are thus established in irradiated materials by using combination of in-situ GIXRD, ex-situ HR-XRD, TEM and even UV-Vis absorption. In the presentation, an overview on defect formation and structural modifications (amorphization, strain, grain subdivision) associated to energy deposition processes (Se or Sn) induced by swift heavy ion irradiation in functional materials will be presented.

Resume : The realisation of fusion energy is dependant on the development of resilient plasma facing materials that can withstand the combined effects of high energy neutron irradiation and high operating temperatures. Transmutation in these components leads to the production of Re and Os, and a composition that varies significantly with time [1]. The precipitation of Re and Os, observed experimentally in both ion and neutron studies is a potential life limiting factor [2]. In this research, neutron irradiated W was irradiated to 1.67dpa at 900oC, producing a post irradiation composition of W-1.4at.%Re-0.1at.%Os. Characterisation using a combination of Atom Probe Tomography (APT) and Scanning Tunnelling Electron Microscopy (STEM) found mostly voids decorated with both Re and Os, as well as smaller isolated precipitates. Grain boundaries were also decorated with both Re and Os. Experimental results were supported by a multicomponent, atomistic kinetic Monte Carlo model [3] incorporating both vacancy & interstitial defects, and solute concentration dependant interactions. [1] Gilbert, M.R. and Sublet, J.-C., Nuclear Fusion, 2011. [2] Lloyd, M.J., Abernethy, R.G., Gilbert, M.R., Griffiths, I., Bagot, P.A., Nguyen-Manh, D., Moody, M.P. and Armstrong, D.E., 2019. Scripta Materialia, 173, pp.96-100. [3] Lloyd, M.J., Abernethy, R.G., Armstrong, D.E., Bagot, P.A., Moody, M.P., Martinez, E. and Nguyen-Manh, D., 2019. The European Physical Journal B, 92(10), p.241.

Resume : Brushite CaHPO4.2H2O (Dicalcium Phosphate Dihydrate: DCPD) is considered as a sorbent of strontium (90Sr2+) which is a contaminant present in some acidic effluents produced by cleaning operations during the dismantling of former nuclear facilities. The idea is to check how electron - radiation - sensitive is this material, since structural and chemical changes caused by irradiation might affect significantly its ion exchange capacity. The energy deposited by the projectile electron into the target is known to be approximated as the sum of two contributions, ballistic and electronic. But only their order of magnitude is supposed to be known. As a result, the structural damage caused by electron irradiation cannot be even reasonably estimated and the radiation sensitivity of brushite has to be checked experimentally. In this paper, powder X-ray diffraction and Raman spectroscopy are associated to analyze the damage effects to brushite following electron irradiation. The accelerator NEC Pelletron of the SIRIUS platform is used for electron irradiation at 2.5 MeV (dose rate of order 20 kGy/s). Amorphization of brushite is observed with increasing dose, and was found complete at 6 GGy. The Raman spectroscopy study tells us that the ultimate compound at the highest dose could be of type anhydrous pyrophosphate. Consequences regarding the damage mechanisms are considered.

Resume : Both borosilicate and iron phosphate glasses are used for the encapsulation of high level nuclear waste. A description of how atomistic models can be used, together with molecular dynamics simulations of radiation damage in these glasses to understand the damage processes and to explain why glasses are an especially good choice for waste encapsulation. In the case of crystals the amount of damage produced by an irradiation event is usually categorised by the number of Frenkel pairs produced but this cannot be used in the case of a glass. It will be shown that a useful concept instead is the change in atomic co-ordination induced by the collision cascade. It will also be demonstrated how some modern simulation techniques can be used to determine how the glasses relax over times scale longer than those generally accessible by molecular dynamics.

Resume : Single crystals of 4H-SiC were implanted by noble gases under different conditions (fluence and temperature). The elastic strain buildup was particularly studied at elevated temperature for which the dynamical recombination prevents the amorphous transition. A phenomenological model based on Multi-Step-Damage Accumulation (MSDA) and cascade recovery (Gibbon's model) allows to fit the evolution of the maximal elastic strain according to the fluence and for different temperatures of implantation. In addition, with the help of transmission electron microscopy the evolution of the implantation related defects under annealing was studied. The formation of nanocavities was observed under severe implantation/annealing conditions of Xe implantation. These cavities are found of different nature according to the xenon and damage distribution as well as the related-strain profile (obtained by the simulations of the XRD curves). The surface swelling (step height) under implantation/annealing was also studied leading to different activation energies supporting the idea of different defects contributions to swelling.

Resume : The theoretical basis of object kinetic Monte Carlo, and many related methods, is that the current state of the system consists of set of spatially discrete objects- point defects, dislocation loops and voids - which may evolve and interact with each other with predefined rules. An object is typically described with a small number of parameters - for instance a void may be defined by the number of vacancies contained, but there is no reason why additional shape information should not be retained if such information is useful. The logical limit of the amount of information retained per object is a description of the atomic types and positions. We describe the development of a new object kinetic Monte Carlo code where the elementary defect objects are off-lattice atomistic configurations. An object evolves by internal rearrangements of the atoms, which can be found on-the-fly. Such rearrangements typically transform one object into another, but can also involve displacement of the centre-of-mass, or splitting one object into two. Elastic interactions are handled with interatomic potentials within an object, and with the dipole tensor formalism between objects. We track visited configurations to build up a relevant database of objects plus events. Once a good database is found we can evolve mobile interstitial clusters faster than the equivalent MD simulation, even at high temperature. We apply the model to the evolution of 20 keV PKA cascades in tungsten, and capture complex internal relaxations of sessile objects instigated by object collisions. These cascades are evolved to experimental timescales at technologically relevant temperatures 600-900 K. We discuss the lessons learnt and the challenges for modelling irradiation damage evolution with object kinetic Monte Carlo. This work has been carried out within the framework of the EUROfusion Consortium and has received funding from the Euratom research and training programme 2014–2018 and 2019–2020 under grant agreement No. 633053, and from the RCUK Energy Programme [grant number EP/P012450/1]. The views and opinions expressed herein do not necessarily reflect those of the European Commission. AES acknowledges support from the Academy of Finland through project No. 311472.

Resume : Graphite has been the material of choice in construction of nuclear reactors for many years due to its low neutron absorption cross-section and high scattering cross-section. The physical properties of a graphite moderator can greatly influence the cost, safety and lifespan of a reactor. Neutron collision damage in graphite results in the formation of basal dislocations. The subsequent interaction of basal dislocations with each other and the surrounding lattice causes severe dimensional changes along the basal direction. There has been a lot of interest recently in AB and AC stacking grain boundaries in bilayer graphene. Transition from AB to AC stacking can be described by the glide of partial basal dislocations resulting in expansion of dislocation cores and buckling of the bilayer. Herein we present full ab initio and molecular dynamics calculations of basal dislocation network structures in bilayer graphene and few-layer graphene in large supercells of up to 100 nm.

Resume : Owing to their properties, reduced-activation ferritic/martensitic (RAFM) steels are proposed as structural materials for in-vessel components of future fusion reactors. Neutron irradiation produces defects in the material, leading to detrimental effects on its mechanical properties such as hardening and embrittlement. Dislocation mobility, which is related to the mechanical behaviour of the material, is highly influenced by the defects created due to neutron/ion irradiation. Moreover, it has been reported that dislocations have an influence on the spatial distribution of defects in bcc iron. Elastic interactions between dislocations and SIA clusters lead to dislocation decoration and to an inhomogeneous spatial distribution of SIA clusters. In this work, we present a study of the influence of spatial defect distribution on dislocation mobility in the case of self-ion irradiation on bcc iron. For this purpose, defect distribution in the presence of dislocations is calculated by MEGA code, an object kinetic Monte Carlo code based on GPU programming which allows the study of high irradiation doses, large spatial regions and elastic interactions between SIA clusters and dislocations. Dislocation mobility in the presence of defects, either placed at random positions or at those positions previously calculated by okMC, is studied by means of Dislocation Dynamics technique with MoDELib suit of codes.

Resume : Si detectors built on p-type substrates are suitable for high energy physics applications due to their improved radiation hardness. However, they are still strongly affected by irradiation-induced displacement damage, as it alters the net space charge. It has been reported that the effective doping concentration (Neff) undergoes a fast decay for low neutron fluences, attributed to an acceptor removal process, and that the removal rate decreases as the initial dopant concentration increases. This process is associated to dopant deactivation by damage, but its microscopic origin and the reason for the rate dependence on resistivity are still unknown. We use atomistic simulations to analyze the acceptor removal process in samples with B concentrations ranging from 1013 to 1017 cm-3, irradiated up to 1016 cm-2 1-MeV neq fluences. Si recoils are simulated within a binary collision approximation scheme, following the energy distribution of primary knock-on atoms calculated with the SPECTRA-PKA code. The subsequent dopant-defect interactions are simulated with a kinetic Monte Carlo code with detailed models of defect and dopant diffusion, as well as dopant deactivation. The simulated Neff dependence on fluence and the obtained removal rates are consistent with experiments. Through the analysis of defect-dopant interactions we will shed light on the atomistic mechanisms responsible for the acceptor removal process, and will explain some intriguing features not understood until now.

Resume : The current strategies to reduce beam damage in a transmission electron microscope (TEM) consists in reducing the electron dose, calculated in electrons per area (usually given in square angstrom), that is delivered to the sample either by decreasing the electron beam current or making fast acquisitions of images, resulting in low signal-to-noise ratio images. A new approach from the field of data compression has been proposed to overcome or alleviate the problem of beam damage in TEM. The so-called compressed sensing technique is based on a sub-sampling approach where the information can be retrieved from a reduced data acquisition [1]. In STEM (Scanning TEM) mode, with compressed sensing is possible to retrieve an image scanning a fewer number of pixels than with normal acquisition. This non-conventional method of dose reduction has been recently implemented in STEM [2,3] and it has showed the possibility of atomic-resolution imaging of beam sensitive materials [3]. As with this sub-sampling technique the sample is not irradiated consecutively, we proposed that a dissipation of a dynamic process of beam damage could be involved in damage reduction when the dose is accumulated at a slower rate than the diffusion of this process. Charging or heating of the sample would follow these model, for instance. In the conventional raster scanning where the probe is scanned across the sample line by line, the dose is accumulated in a consecutive order. If only the order of the scanning positions is modified, skipping a fixed number of pixels between two scanning positions in a line and scanning the missing pixels after completing all the scan lines, for instance, the dose in the surrounding area of a specific point in the sample would be accumulated at a different time rate than with the raster scanning. If any dynamic mechanism is involved, this strategy could dissipate or reduce damage on the sample. With the help of a programmable scan engine different alternative scanning strategies were tested on a beam sensitive sample (commercial Linde type A Zeolite) where the total electron dose was kept constant but its temporal distribution modified. Our results showed less damage when scanning in a different order than raster scanning. These findings demonstrate that the way how the dose is distributed plays an important role on beam damage which gave some insights about the mechanism which can be modelled by a diffusion process. Diffusion simulations can be useful to find a suitable scanning strategy according to the sample characteristics to outrun beam damage in STEM. [1] Candes E. J., Romberg J. K. and Tao T., Communications on Pure and Applied Mathematics, 59, 1207-1223 (2006). [2] Béché, A., Goris, B., Freitag, B., Verbeeck, J., Appl. Phys. Lett. 2016, 108, 093103. [3] A. Stevens, L. Luzi, H. Yang, L. Kovarik, B. Mehdi, A. Liyu, M. Gehm, and N. Browning, Appl. Phys. Lett. 112, 043104 (2018).

Resume : One of the main structural candidate materials for advanced nuclear fusion reactors over the last decade are oxide dispersed strengthened reduced activation ferritic steels (ODS RAFS). A fine, homogeneous, dispersion of stable nanoparticles such as Y2O3 would allow increasing their upper operating temperatures and radiation damage resistance. However, currently available results require further research to clarify the effect of irradiation on these materials. The main objective of this work is to investigate the microstructural stability of different ODS RAFS with nominal composition Fe–14Cr–2W–0.3Ti–0.3Y2O3 (wt%) subjected to simultaneous dual (Fe3+, He+) ion irradiation at 400 °C to 30 dpa. The characterization of the steels has been accomplished using Transmission electron microscopy (TEM), Atom Probe Tomography (APT) and Positron Annihilation Spectroscopy (PAS). The stability of the nanoparticles has been investigated. The irradiation induced damage has been characterized, focusing in the study of open volume defects. There is a slight change in the chemical composition of nanoparticles after irradiation. The mean size and mean distance of nanoclusters is lower after the irradiation pointing to dissolution of larger nanoclusters and nucleation of new, smaller ones. PAS investigations show a limited increase of open volume defects after the irradiation suggesting a possible entrapment of all injected He ions into the open-volume defects simultaneously created by the Fe irradiation or/and recombination of vacancies and interstitials due to the intermediate temperature of the irradiation. CDB profiles show no clear relation between positron traps and He.

Resume : Mesoporous silicas have been proposed for the treatment of radioactive effluents by a new strategy called separation / conditioning. This strategy would allow both the separation of the radionuclide using a selective organic ligand, and its encapsulation after collapse of the porosity by “soft” way (sol-gel reaction, heating under stress or irradiation) to obtain a primary wasteform matrix. This low temperature process would be particularly interesting for the treatment of radioactive effluents containing volatile elements (Cs, I). The talk reports recent results on the dimensional change and the collapse of mesoporous silica pellets under electron beams (2MeV) representative of the beta decay of Cs. The irradiated samples were analyzed by density measurement, Small Angle X-Rays (SAXS), nitrogen adsorption-desorption isotherms to evaluate the evolution of the mesoporous network and Nuclear Magnetic Resonance (NMR), infrared spectroscopy (IR) for the silica network. A mechanism will be proposed to explain the evolution of the mesoporous structure under irradiation. P. Makowski, X. Deschanels, A. Grandjean, D. Meyer, G. Toquer and F. Goettmann, New J. Chem., 2012, 36, 531. Y. Lou, S. Dourdain, C. Rey, Y. Serruys, D. Siméone, N. Mollard, X. Deschanels, Micropor. Mesopor. Mater, 2017, 251, 146.

Resume : One of the open issues concerning the quantitative understanding of irradiation hardening in Cr steels is related to the spatial arrangement of irradiation-induced dislocation loops with respect to pre-existing dislocations and its consequences for irradiation hardening. In order to tackle this issue, a number of neutron-irradiated industrial-purity Fe-Cr alloys (Cr between 2.5 and 12%) was studied by means of TEM. Loops in irradiated Fe-Cr alloys of lower Cr content were found to arrange more homogeneously in the matrix. In contrast, loops tend to arrange preferentially along dislocation lines for higher Cr contents [Hernandez-Mayoral et al., JNM 474 (2016) 88]. At the same time, reported tensile stress-strain curves [Matijasevic et al., JNM 377 (2008) 147] exhibit trends from smooth elastic-plastic transitions towards yield drops for both increasing Cr and increasing neutron exposure with an exception to be discussed. The poster is aimed to rationalize both observations, namely dislocation decoration at the nm length scale and yield effects at the macroscale, on the basis of the source hardening model [Trinkaus et al., JNM 251 (1997) 172]. In particular, calculated unlocking stresses are found to correlate with measured values of the upper yield stress.

Resume : Future generation nuclear reactors present intrinsic benefits in terms of economics, safety and waste management. However, innovative material strategies must be designed. Recently, coatings have earned great attention in the nuclear field as a reliable cost-efficient solution. In this framework, amorphous/nano-crystalline oxides have been characterized as corrosion-resistant anti-diffusion barriers. Previously, oxide films have been tested under ion irradiation, proving high stability up to 150 dpa. This radiation tolerance has been related to the amorphousness of the pristine material, which evolves towards a crystalline structure under irradiation. To preserve the radiation tolerance and improve the range of operation, the amorphous state must be conserved as long as possible. In this work, the stabilization of traditional oxides by doping is discussed. Firstly, pure and doped materials have been compared by thermal annealing, showing a significant increase in the crystallization temperature of the mixed systems. Then, irradiation tests have been performed with in situ TEM. Samples have been irradiated with different ions, up to 20 dpa. Experiments have been conducted at 600 °C and at 800 °C, to observe the effects of temperature on the radiation induced crystallization. Dopants retard the crystallization threshold even under irradiation and stabilize the intermediate metastable phases, conserving partially the pristine characteristics. While pure systems suffer from voids formation after crystallization (a traditional phenome for bulk oxides), no swelling appears in the doped counterpart.

Resume : The mechanical and thermal property change of materials under irradiation is an important challenge for nuclear materials. For a typical Pressurized Water Reactor (PWR), radiation damage of the Reactor Pressure Vessel (RPV) is one of the most important issues that determine the operating lifetime of a reactor. Accurate calculation with realistic uncertainty estimates of radiation damage is thus of great importance. In nuclear reactors, radiation damage of materials is mainly induced by energetic neutrons. The accurate uncertainty quantification of neutron radiation damage requires thus the complete covariance matrix of damage cross section and the covariance matrix of neutron flux spectrum. Covariance matrices of different Prompt Fission Neutron Spectra (PFNS) of 235U are propagated to the neutron flux spectra in a simplified PWR RPV (such as the examples shown in Ref. [1]). Covariance matrix of damage cross section is propagated from nuclear reaction model parameters and primary damage model parameters. The total uncertainties of neutron-induced displacement damage from both the uncertainties of PFNS and the uncertainty of damage cross section are deduced using the so-called “sandwich” formula. In addition to the calculation of total uncertainty of primary radiation damage in the RPV, Primary Knock-on Atom (PKA) spectra and the corresponding covariance matrices are obtained by considering the covariance matrix of neutron flux spectra (but without considering the uncertainties from nuclear reaction data). These PKA spectra and the associated covariance matrices can be used together with molecular dynamics or binary collision approximation simulations to accurately determine the radiation damage and the corresponding uncertainty. Using the standard Norgett-Robinson-Torrens formula, numerical results show that the complete covariance matrices of neutron flux (resulting from covariance matrices of PFNS) lead to about 4% relative uncertainty (depending on different PFNS) for the atomic displacement damage of the RPV, whereas only 1% uncertainty is obtained if the correlations of neutron flux spectra in the RPV at different energies are not considered for generating the PKA spectra. [1] L. Berge, “Contribution à la modélisation des spectres de neutrons prompts de fission. Propagation d’incertitudes sur un calul de fluence cuve,” PhD thesis, Université Grenoble Alpes, 2015.

Resume : Ferritic steel is one of the most promising candidate structural material of future fusion reactors. The extreme conditions when exposed to the fusion plasma, with high temperatures and 14 MeV fusion neutron irradiation, lead to a microstructural modification that impacts properties. Besides the generation of dislocation loops and cavities, there is Cr-rich α' precipitation, which leads to hardening and embrittlement, as well as changes in magnetic properties. In order to better understand and model its evolution a detailed description of its structure is needed. Fe-Cr model alloys are used to understand the fundamental mechanisms at play. Fe-20Cr (in wt.%) were considered in this work, with a thermally-driven decomposition induced by an anneal at 500 °C that leads to a high number density of α' precipitates that are few nanometers in diameter. The α' precipitates’ structure, composition and interface with the matrix were investigated using high-end transmission electron microscopy correlated to atom probe tomography and supported by molecular dynamics simulations.

Resume : In a recent study we addressed the formation mechanisms of <100> loops in pure iron. In this model, the migration of the loops did not depend on the local environment. In fact, <100> loops were considered immobile while ½<111> loops migration followed an Arrhenius law using some predefined prefactors and migration barriers depending on their size. With the present study we want to move a step forward by considering long range interactions between dislocation loops of any size. Expressions for elastic interactions between dislocation loops derived from Dudarev et al. are included into our custom version of the kinetic Monte Carlo code MMonCa. Preliminary calculation show that, for high enough concentrations of loops, the mobility is biased by the presence of the other loops. We plan to perform simulations in irradiated pure iron to explore and evaluate the effects of such interactions into the one dimensional migration of ½<111> loops. In a last instance we plan to compare our results with those obtained by Arakawa et al. who obtained activation energies for loop diffusion of 1.3 eV, independent of the loop size, value that is one order of magnitude higher than that obtained with molecular dynamics simulations. This first approach in using elastic interactions are not accounting for thermally fluctuating small changes in the direction of the Burgers vector, as predicted by Boleininger e et al.

Resume : Radiation hard microelectronic components are essential for aeronautics and space applications and the reduced feature sizes at advanced technology nodes make these devices more susceptible to failure due to irradiation. Our objective is to investigate the atomic-level mechanisms of irradiation-induced defect generation in semiconductor/dielectric heterostructures relevant for FinFETs and nanosheet transistors. We focus on the initial generation of radiation cascades and the annealing that takes place over the first tens of picoseconds using large-scale molecular dynamics simulations using accurate reactive interatomic potentials. We modeled Si/amorphous-SiO2 heterostructures and considered 10 keV Si and O primary knock-on atoms. A detailed analysis of the MD trajectories enabled us to characterize displaced atoms and defective ones. As could be expected, we find that the Si/SiO2 interfaces acts as sinks for O defects but, quite surprisingly, Si defects accumulate at interfaces and preferentially anneal in the bulk SiO2 and Si regions.