Correlative light and electron microscopy (CLEM) is a powerful approach to investigate the molecular ultrastructure of labeled cell compartments. However, quantitative CLEM studies are rare, mainly due to small sample sizes and the sensitivity of fluorescent proteins to strong fixatives and contrasting reagents for EM. Here, we show that fusion of a self-labeling protein to insulin allows for the quantification of age-distinct insulin granule pools in pancreatic beta cells by a combination of super resolution and transmission electron microscopy on Tokuyasu cryosections. In contrast to fluorescent proteins like GFP organic dyes covalently bound to self-labeling proteins retain their fluorescence also in epoxy resin following high pressure freezing and freeze substitution, or remarkably even after strong chemical fixation. This enables for the assessment of age-defined granule morphology and degradation. Finally, we demonstrate that this CLEM protocol is highly versatile, being suitable for single and dual fluorescent labeling and detection of different proteins with optimal ultrastructure preservation and contrast.

The chorda tympani nerve, also known as the taste nerve, runs uncovered through the middle ear cavity, a localization that exposes the nerve to pathological processes and surgical trauma in the middle ear. People operated on for otosclerosis tend to complain more about postoperative taste disturbances than those operated on for chronic otitis media. It has been suggested that this difference may be explained by gradual deterioration of chorda tympani nerve function caused by chronic otitis media infection and that further impairment caused by surgery is less noticeable in these patients. This thesis aimed to evaluate the function of the chorda tympani nerve, the effects of middle ear disease on taste and complications resulting from ear surgery for chronic otitis media or otosclerosis. This information will help to improve the ear surgeon’s ability to predict the prognosis of iatrogenic taste disturbances in patients with middle ear disease and after ear surgery. Taste was assessed using electrogustometry and the filter paper disc method before and after surgery for chronic otitis media or otosclerosis. Patients also completed questionnaires about symptoms and quality of life. The status of the chorda tympani nerve upon surgical opening of the ear and grading of the trauma to the nerve during the surgery were recorded. The ultrastructure of the chorda tympani nerve from healthy ears and from ears with chronic otitis media was examined. Electrogustometry and the filter paper disc method were evaluated. The results of electrogustometry and the filter paper disc method were highly reproducible, although their correlation was moderate. Patients with chronic otitis media, patients with a more traumatized nerve, female patients and younger patients were more likely to report postoperative taste disturbances. Most of the patients recovered their taste after 1 year. The quality of life study showed only minor changes after surgery. Electron microscopic observations of nerves from ears with chronic otitis media showed signs of structural degeneration, although signs of regeneration, such as sprouting were also observed. This results may explain the recovery of taste postoperatively and indicate that the nerve should be carefully handled during surgery.

Afin de réaliser des traitements locaux sur des données de diverses natures (volumes, images ou signaux) contenant des éléments informatifs dans certaines bandes de fréquence, nous nous intéressons dans cette thèse à l’étude de bancs de filtres (BdF). Plus précisément, nous étudions l’existence et la synthèse de BdF de réponses impulsionnelles finies (RIF) inverses d’un BdF d’analyse RIF redondant complexe fixé. Nous proposons en particulier des méthodes testant l’inversibilité de la matrice d’analyse et la construction d’un inverse explicite à l’aide de la formulation polyphase. À partir de ce dernier, nous proposons une paramétrisation réduite de l’ensemble des BdF de synthèse permettant d’optimiser leurs réponses selon différents critères. Cette étude est étendue au cas multidimensionnel notamment par l’utilisation de la notion de résultant. Ces outils permettant de représenter efficacement certaines informations structurées dans des données, il devient possible de les préserver tout en rejetant d’éventuelles perturbations. Le premier cadre considéré est celui du bruit gaussien. Nous avons utilisé le principe de Stein pour proposer deux méthodes de débruitage : FB-SURELET-E et FBSURELET-C. Elles sont comparées à des méthodes récentes de débruitage conduisant à de bons résultats en particulier pour des images texturées. Un autre type d’application est ensuite considéré : la séparation des structures orientées. Afin de traiter ce problème, nous avons développé une méthode de filtrage anisotrope. Les algorithmes réalisés sont finalement testés sur des données issues de différents domaines applicatifs (sismique, microscopie, vibrations) / Our main purpose in this PhD thesis is to perform local frequential (or directional) processing in different kind of data (volumes, images or signals). To this end, filter banks (FBs) are studied. More precisely, we first investigate the existence and the construction of synthesis FBs inverse to a given FIR complex analysis FB. Through the study of the polyphase analysis matrix, we are able to propose methods to test the invertibility and to build one inverse FB. Using this inverse, we provide a parametrization of the set of synthesis FB, with which we optimize filter responses with different criteria. The same study is performed in the multidimensional case. Since FBs provide an efficient representation of structured information in data, it is then possible to preserve them while rejecting unwanted perturbations. By associating Stein’s principle and those FB, we proposed two methods to denoise signals and images corrupted by Gaussian noise. These methods, named FB-SURELET-E and FB-SURELET-C, are compared to recent denoising methods and are found to offer good results, especially for textured images. Another type of application is then investigated : separation of oriented structures. To this end, we have developed an anisotropic filtering method. The different proposed methods are finally applied on images and signals from various fields : seismic images and cubes, transmission electron microscopy (TEM) images of catalysts and vibration signals from car engines

A probe for measuring the specific heat of superconductors at low temperatures and in high magnetic fields has been built and commissioned. The probe has been tested using the relaxation method on samples of copper and the accuracy of the data is 1.3 % between 5 K and 30 K, data taken using the long range pulse method has a resolution of 10 mK. Specific heat measurements have been performed on members of the series (Pb(_1)-(_x))Cu(_1.8x)Mo(_6)S(_8), (Sn(_1-x))Eu(_x)Mo(_6)S(_8) and (Pb(_1-x)M(_x))Mo(_6)S(_8) where M = Gd and Eu, from 3 K up to 30 K and in magnetic fields up to 15 T. Additional results from resistivity, susceptibility, magnetisation. X-ray diffraction, transmission electron microscopy and electron dispersive-ray measurements are also presented. These data have been compared to results from other authors and are analysed in terms of the BCS and GLAG theories of superconductivity and the magnetic properties of these materials. The mean field model has been used to calculate numerically the magnetic contribution to the specific heat (cm) of both ferromagnetic and antiferromagnetic systems as a function of temperature and applied field both above and below the ordering temperature. In addition an approximate analytic form for the magnetisation has been used to calculate Cm above the ordering temperature. Expressions have been derived for the saturation value of the peak in C(_m): C(^sat)(_m) = 1.1245n(_cell)RJI(J+1) and the temperature dependence of the peak with applied field ȡ(μ(_o)H(_ext))/ȡT(_peak)=6.540/g(_J)(J+1). They allow the simple calculation of the values of J and g(_J)(J + 1) from specific heat data. The magnetic contribution to the specific heat of the samples (Sn(_0.65)Eu(0.35)Mo(_6)S(_8)) and (Sn(0.50)Eu(_0.50)Mo(_6)S(_8)) have been modelled using these calculations and excellent agreement is found by considering the magnetic ions as free ions. The sample is accurately modelled by including an additional minority phase (Gd(_2)S(_3)). The approximate expressions have also been used to analyse data on high temperature superconductors producing values of J and g(_J)}{J + 1) consistent with a doublet ground state. The properties of Chevrel phase materials have been determined as a function of doping level. The critical temperature is degraded by doping but an increase in the critical current density is observed in the series (Pb(_1-x)Cu(_1-8x)Mo(_6)S(_8) for very low levels of doping. Increases of up to 28 % in the upper critical field, that are probably due to the compensation effect and an increase in the normal state resistivity, are also observed in the series (Sn(_1-x)Eu(_x)Mo(_6)S(_8)) at high levels of doping and in the series (Pb(_1-x)Gd(_x)Mo(_6)s(_8) for low levels of doping.

De nos jours, l'essor des produits électroniques nomades requièrent une capacité de stockage de données croissante et imposent la fabrication de composants mémoire performants, denses et fiables. Cela implique une grande robustesse des cellules mémoires élémentaires dont les dimensions caractéristiques sont régulièrement réduites. L'objectif principal de la thèse est d'appréhender les mécanismes de claquage d'oxydes minces voire ultraminces intégrés dans des empilements métal/oxyde/semiconducteur. Un intérêt particulier a été porté à la croissance d'îlots cristallins épitaxiés se formant lors de certaines sollicitations électriques et associée aux mécanismes de DBIE (Dielectric Breakdown Induced Epitaxy). L'étude des différents dispositifs (cellules mémoires à grille continue ou discrète, transistors, condensateur) a permis de proposer des corrélations entre la défaillance électrique de ces dispositifs et les défauts microstructuraux générés. Ce travail a été réalisé selon une méthodologie intégrant (i) la sollicitation électrique ; (ii) une préparation d'échantillons adaptée ; (iii) l'identification, l'observation et la caractérisation des défauts par microscopie électronique en transmission (TEM). L'ensemble des études menées a permis d'isoler deux paramètres électriques principaux ayant un rôle prépondérant sur la formation d'îlots de silicium épitaxiés, en lien avec le mécanisme de DBIE : la charge injectée et le courant de compliance. Ces deux paramètres apparaissent comme des facteurs limitant l'emballement thermique qui conduit, en général, à un claquage diélectrique franc de l'oxyde et semblent, par conséquent, retarder la défaillance irréversible d'un dispositif. / Nowadays, the microelectronic industry had to take up ambitious challenges to satisfy the strong economic demand because of the mobile electronic products booming like smartphones, tablets, or more recently "phablets". These high added value products requires the growth of data storage capacity and, subsequently, to produce high-performance, dense and reliable components. That implies a great cell memories robustness whose critical dimensions are regularly reduced. In this context, the thesis issue is to better understand the breakdown mechanisms of the thin and ultra-thin oxides embedded in metal/oxide/semiconductor stacks. Actually, epitaxial growth of crystalline silicon hillocks was pinpointed. These hillocks grown under electrical stresses and were associated to DBIE mechanisms (Dielectric Breakdown Induced Epitaxy). Device studies allowed to correlate electrical stress conditions and microstructural defects thanks to a 3-steps methodology : (i) electrical stresses leading to microstructural defects ; (ii) sample preparation including defect localization and extraction ; (iii) identification, observation and characterization of defects by transmission electron microscopy (TEM). Two main electrical parameters were identified with factors responsible for hillocks growth linked to DBIE : the injected charge and the compliance current. These parameters seem to limit the thermal runaway inducing hard breakdown. Consequently, it is possible that delays the irreversible device degradation. In addition, hillocks seem to grow preferentially under polysilicon grain boundaries over the SiO2/Si stacks.

Au cours de ces travaux de thèse, nous avons étudié la redistribution atomique de contaminantsmétalliques dans le silicium et au voisinage d'une interface SiO2/Si. Pour mener à bien cetteétude, trois techniques de caractérisation complémentaires ont été utilisées (TEM,APT,SIMS).Nous avons dans un premier temps étudié la diffusion ainsi que la ségrégation d'équilibre de contaminants à une interface SiO2/Si, et plus particulièrement, la diffusion du W et du Mo. Le Wprésente une cinétique de diffusion extrêmement lente. Les caractérisations réalisées par TEM et APT nous ont permis de discuter les profils de concentrations mesurés par SIMS et nous ont guidés dans le choix du modèle de diffusion proposé. L'étude de la diffusion du Mo révéle que cette espèce présente une limite de solubilité faible dans le silicium et une forte interaction avec des défauts d'irradiation, provoquant sa précipitation.Dans un second volet, nous nous sommes intéressés à l'effet d'une interface mobile, lors d'une réaction, sur la redistribution atomique des contaminants proches de cette interface. Nous avons ainsiréalisé une étude comparative des comportements du Fe et W lors de procédés d'oxydation.Le tungstène précipite dans le volume et est progressivement rejeté par l'oxydation. Le ferprécipite à l'interface SiO2/Si, provoquant un effet de masquage dont nous avons montré qu'il étaitresponsable de la formation de défauts pyramidaux d'interface, caractéristiques d'une contaminationen fer du silicium. Le procédé de germano-siliciuration de nickel, réalisé à basses températures a également été investigué. Cette réaction provoque le rejet 3D du germanium à l'interface NiSiGe/SiGe. / During this thesis work, we studied the atomic redistribution of metallic contaminantsin silicon and near a SiO2/Si interface. To conduct this study, we used three complementary characterisation techniques : transmission electron microscopy (TEM), atomic probe tomography (APT) and secondary ion mass spectrometry (SIMS).We first studied the diffusion and equilibrium segregation of various contaminants at a SiO2/Si interface, and more particularly, the diffusion of W and Mo. W exhibits a very slow diffusion kinetic.Physico-chemical characterizations performed by TEM and APT allowed discussing the concentrationprofiles obtained by SIMS leading to the diffusion model that we proposed. The study of Mo diffusionrevealed that this specy exhibits a low solubility limit in silicon and strongly interacts with irradiation-induced defects, leading to its precipitation.In a second phase, we studied the effect of a mobile interface, during a reaction, on the atomic redistribution of contaminants near this interface. We performed a comparative study of the behaviourof Fe and W during oxidation processes. W precipitates in the silicon substrate and is progressivelyrejected (snowplow) by the oxidation. Fe preferentially precipitates at the SiO2/Si interface. Theseprecipitates mask a part of the silicon substrate and thus hinder its oxidation, leading to the formation of characteristics pyramidal-shaped defects at the interface. Low temperature nickel germano-silicide formation have also been investigated. This reaction leads to the 3D snowplow of germanium atoms at the NiSiGe/SiGe interface.

De nombreuses propriétés étonnantes ont été récemment découvertes à l’interface de deux perovskites isolantes l’une polaire l’autre non polaire. La discontinuité de charge à l’interface LaAlO3/SrTiO3 engendre un gas d’électron quasi-bidimensionnel qui confère un caractère métallique à cette interface. Les mécanismes locaux et la quantification des propriétés ne font pas consensus car l’interdépendance de facteurs structuraux, chimiques et électroniques complexifie la résolution du problème posé. Une catastrophe polaire, des distorsions structurales, des lacunes d’oxygène, une interdiffusion cationique et une non stœchiométrie du film ont été séparément avancées pour expliquer cette conduction. Dans le cadre d’un programme international, nous avons reçu des héterointerfaces conductrices et isolantes élaborées par ablation laser pulsé (PLD). L’origine des porteurs de charge a été recherchée par une approche globale liant procédé, structure et propriétés électriques (mesurées dans le consortium). Nous avons systématiquement analysé les interfaces en combinant imagerie à haute résolution (STEM-HAADF) et spectroscopies électroniques (EELS) et ioniques (MEIS). Une non planéité des couches atomiques, une interdiffusion cationique et un transfert d’électrons permettent de réduire la divergence de potentiel, la catastrophe polaire n’a donc pas lieu. La formation de défauts donneurs à la surface du film devient favorable au-delà d’une épaisseur critique. Les électrons sont transférés à l’interface dans la bande de conduction du STO. Nous avons mis en évidence un mécanisme de compensation concurrentiel de la charge interfaciale par des lacunes de strontium chargées négativement, qui mettent le substrat en compression plane et s’opposent au confinement 2D électrons. La variation des paramètres procédés, tels que la durée du dépôt, la pression partielle en oxygène, la température et la stœchiométrie de la plume déplace l’équilibre des différents mécanismes mis en évidence. Ce travail démontre la relation complexe liant procédé, propriétés électriques et distribution des défauts autour de ces interfaces singulières. / Novel behavior at the interface between two insulating polar/non polar perovskites has been recently discovered. The polarization discontinuity at LaAlO3/SrTiO3 drives the formation of quasi two dimensional electron gas. Both the local mechanism and quantification of such behavior remain unclear due to interplay of structural, chemical and electronic factors. Several mechanisms have been proposed, such as the polar catastrophe, structural distortions, oxygen vacancies, cationic intermixing at the interface and film non-stoichiometry. In the frame of an international project conductive and insulating heterostrucutres have been synthetized by Pulsed-Laser Deposition. In this thesis, we have developed a comprehensive approach to investigate the origin of the charge carriers. The interfaces have been systematically analyzed by combining high resolution imaging (STEM-HAADF) to atomic resolved electron (EELS) and ion (MEIS) spectroscopies. The observed and quantified parameters have been related to the electrical properties of the interfaces measured in the consortium. Buckling of the atomic layers, intermixing and electron transfer reduce the polar divergence. This rules out the polar catastrophe scenario. The formation of donor defects at the film surface is favored above a critical film thickness. Electrons are transferred to interface in the STO conduction band. A competing compensation mechanism of the positive interfacial charge by negatively charged strontium vacancies has been demonstrated that generates an in plane compression of the STO, unfavorable for a strict 2D confinement of the charges. Varying the process parameters such as growth duration, oxygen partial pressure, temperature, and plume stoichiometry shift the equilibrium of the different mechanisms highlighted. This thesis emphasizes the complex relations between the process and the properties through the defects distribution around these singular interfaces.

Les plasmons de surface (SP) sont des excitations mêlant électrons et photons localisées aux surfaceset interfaces métalliques. On peut les voir classiquement comme les modes électromagnétiquespropres d’un ensemble constitué d’un métal et d’un diélectrique. Cette thèse se base sur la capacitéofferte par les techniques de spectroscopie utilisant des électrons rapides disponibles dans un microscopeélectronique à balayage en transmission (STEM), de cartographier, dans une large gammespectrale et avec une résolution spatiale nanométrique, les modes propres SP. Une telle capacitéa été démontrée séparément, durant ces dernières années, par des expériences de spectroscopie depertes d’énergie d’électrons (EELS), qui mesurent l’énergie perdue par des électrons rapides intéragissantavec un échantillon, et de cathodoluminescence (CL), qui mesurent l’énergie réémisepar l’échantillon par l’intermédiaire de photons, toutes deux résolues spatialement. Dans le cas del’EELS, ces résultats expérimentaux sont aujourd’hui interprétables à l’aide d’analyses théoriquesconvaincantes tendant à prouver que la quantité mesurée dans une telle expérience peut être interprétéede façon sûre en terme de modes propres de surface de l’échantillon. Afin d’élargir une telleinterprétation aux techniques de spectroscopies utilisant des électrons rapides en général, j’ai effectuédes expériences combinées d’EELS et de CL résolues spatialement sur une nanoparticle uniquesimple (un nanoprisme d’or). J’ai montré que les résultats offerts par ces deux techniques présententde fortes similitudes mais également de légères différences, ce qui est confirmé par des simulationsnumériques. J’ai étendu l’analyse théorique du signal EELS au signal CL, et ai montré que la CLcartographie, tout comme l’EELS, les modes de surface radiatifs du sytème, mais avec des propriétésspectrales légèrement différentes. Ce travail constitue une preuve de principe clarifiant les quantitésmesurées en EELS et CL sur des systèmes métal-dielectriques. Ces dernières sont démontrées êtrerespectivement des équivalents nanométriques des spectroscopies d’extinction et de diffusion de lalumière. Basé sur cette interprétation, j’ai utilisé l’EELS pour dévoiler les modes propres SP demilieux métalliques aléatoires (dans notre cas, des films semicontinus métalliques avant le seuil depercolation). Ces modes propres constituent une problématique de longue date dans le domainede la nanooptique. J’ai directement identifié ces modes par des mesures et le traitement de leursrésultats. J’ai complètement caractérisé ces modes propres via les variations spatiales de l’intensitéliée à leur champ électrique, une énergie propre et un taux de relaxation. Ce faisant, j’ai montré quela géométrie fractale du milieu, dont la prédominance croit au fur et à mesure que l’on s’approchede la percolation, est responsable de l’existence de modes propres de type aléatoire à basse énergie. / Surface Plasmons (SP) are elementary excitations mixing electrons and photons at metal surfaces,which can be seen in a classical electrodynamics framework as electromagnetic surface eigenmodesof a metal-dielectric system. The present work bases on the ability of mapping SP eigenmodes withnanometric spatial resolution over a broad spectral range using spatially resolved fast electron basedspectroscopies in a Scanning Transmission Electron Microscope (STEM). Such an ability has beenseparately demonstrated during the last few years by many spatially resolved experiments of ElectronEnergy Loss Spectroscopy (EELS), which measures the energy lost by fast electrons interactingwith the sample, and CathodoLuminescence (CL), which measures the energy released by subsequentlyemitted photons. In the case of EELS, the experimental results are today well accountedfor by strong theory elements which tend to show that the quantity measured in an experiment canbe safely interpreted in terms of the surface eigenmodes of the sample. In order to broaden thisinterpretation to fast electron based spectroscopies in general, I have performed combined spatiallyresolved EELS and CL experiments on a simple single nanoparticle (a gold nanoprism). I have shownthat EELS and CL results bear strong similarities but also slight differences, which is confirmed bynumerical simulations. I have extended the theoretical analysis of EELS to CL to show that CLmaps equally well than EELS the radiative surface eigenmodes, yet with slightly different spectralfeatures. This work is a proof of principle clarifiying the quantities measured in EELS and CL,which are shown to be respectively some nanometric equivalent of extinction and scattering spectroscopieswhen applied to metal-dielectric systems. Based on this interpretation, I have applied EELSto reveal the SP eigenmodes of random metallic media (in our case, semicontinuous metal films beforethe percolation threshold). These SP eigenmodes constitute a long standing issue in nanooptics.I have directly identified the eigenmodes from measurements and data processing. I havefully characterized these eigenmodes experimentally through an electric field intensity pattern, aneigenenergy and a relaxation rate. Doing so, I have shown that the fractal geometry of the medium,which grows towards the percolation, induces random-like eigenmodes in the system at low energies.Keywords: Surface plasmons, fast electron based spectroscopies, scanning transmission electronmicroscopy, disordered media

[pt] Os tópicos mais importantes a ser tratados nesta tese de doutorado são os vários problemas envolvidos na síntese de nanotubos contendo nitrogênio. Isto é principalmente motivado pelas possíveis aplicações que podem ser dadas a este tipo de estruturas. A motivação central está relacionada ao fato da possibilidade de fazer dopagens tipo -p e -n em nanotubos de carbono, incorporando átomos como boro ou nitrogênio. Isto está muito longe de ser uma trivialidade devido a que devemos levar em conta que se os nanotubos de carbono forem pensados como bases estruturais para nanocompósitos e dispositivos nanoeletronicos, é necessário controlar cuidadosamente a reatividade das paredes, sua dureza mecânica e o gap eletrônico por meio de um controle da quantidade de átomos inseridos nas paredes ou entre elas. Portanto, do ponto de vista de diferentes aplicações, é importante ter a possibilidade de dopar controladamente os nanotubos. Neste trabalho apresentam-se o quadro descritivo da dependência dos parâmetros de síntese, assim como uma investigação detalhada da formação de outras estruturas co-produto do processo de formação de nanotubos. Como uma idéia pioneira proposta neste trabalho, é enfatizado o uso de fontes puras de C/N em processos de síntese baseados em deposição química na fase de vapor. Desta maneira foi possivel determinar os efeitos da atmosfera de reação e o pretratamento do catalizador como agentes favoráveis ou desfavoráveis para a síntese efetiva de nanotubos de carbono. / [en] The main topic of this thesis is the study of various issues related to the synthesis of nitrogen containing nanotubes. This is mainly inspired in the possible applications such structures can have. The practical background lies in the fact that defined n- and p-doping of carbon nanotubes can be achieved by substituting carbon atoms from the tube walls by heteroatoms such as boron or nitrogen (N). This is far from been a triviality because we must keep in mind that if carbon nanotubes are to be used as future building blocks in nanocomposites and nanoelectronic devices, it is imperative to fine tune their wall reactivity, mechanical strength and electronic band gap by controlling the amount of foreign atoms inserted into the tube lattices. Therefore, from an applications standpoint, it is important to be able to carefully control the insertion of different dopants into nanotubes. In this work, a complete picture of the dependence on the combined synthesis parameters is established and a fundamental insight into the formation of N doped nanotubes and other structures (co- products) is provided. As a pioneering idea of this whole work, the use of pure C/N feedstocks in chemical vapor deposition methods is emphasized. With this, it was possible to determine the effects of the reaction atmosphere and the catalyst pretreatment as either favoring or disfavoring agents towards the synthesis of N-doped nanotubes.

Destacam-se as semelhanças entre os nutrientes boro (B) e cálcio (Ca) quanto às formas e as funções que eles exercem nos vegetais. Partindo do princípio que a interação Ca - B afeta os atributos morfológicos, ultraestruturais e bioquímicos, bem como a nodulação nas plantas leguminosas, objetivou-se com esse estudo determinar a relação Ca:B mais adequada para a produção de grãos e matéria seca de soja, bem como avaliar os efeitos da interação B e Ca na fixação biológica de N2 (atividade da nitrogenase -N-ase [EC 1.18.6.1]), na morfologia de raízes (comprimento e superfície total), anatomia de raízes (nódulos) e de folhas, assim como as avaliações ultraestruturais dos mesófilos foliares, rendimento do teor de óleo nos grãos das plantas de soja (cultivar BRS 284), crescidas em solução nutritiva. Foram avaliadas ainda a produção de massa seca das plantas, as concentrações e os acúmulos de Ca, B, Mg, K e N na parte aérea e raízes. Foi empregado um esquema fatorial 5² incompleto, perfazendo treze combinações de doses de Ca (mmol L-1) e B (µmol L-1), as quais foram distribuídas segundo o delineamento experimental de blocos ao acaso, com quatro repetições. As maiores atividades da N-ase foram obtidas com as maiores doses de Ca e B (7,25 mmol L-1 de Ca e 50 µmol L-1 de B). A menor dose de Ca resultou na inibição da N-ase em cerca de 90%. Os maiores comprimentos e superfícies totais de raízes foram obtidos na combinação de doses de 7,25Ca e 12,5B, enquanto que para a combinação das menores doses de Ca e B observou-se uma redução de cerca de 70% e 60% para comprimento e superficial de raízes, respectivamente. A maior produção de biomassa seca das plantas e de grãos, os maiores acúmulos de Ca, B, N, K e Mg nas folhas e o maior rendimento de óleo foram obtidos com o fornecimento de 4,25 mmol L-1 de Ca e 12,5 µmol L-1 de B, observando-se relação Ca:B na solução nutritiva próxima de 350:1 e nas folhas de 700:1. O teor de óleo nos grãos incrementou cerca de 20% com o aumento das doses de B de 3,125 para 50 µmol L 1. O número de camadas de células do córtex dos nódulos aumentou com o incremento das concentrações de Ca e B na solução. Observou-se ruptura da membrana dos cloroplastos no mesófilo foliar e elevado número de bactérias fixadoras de N2 penetrando o interior dos tecidos nodulares quando as plantas foram submetidas ao menor suprimento de Ca e maior de B. A combinação da menor dose de B junto a maior dose de Ca provocou ruptura da parede celular e o espessamento da lamela média das folhas. O aumento no fornecimento de Ca ameniza os efeitos negativos do baixo suprimento de B na produção de biomassa das plantas e de grãos, no comprimento e superfície de raízes e na atividade da nitrogenase. A relação Ca:B, tanto na solução, quanto no tecido vegetal, não pode ser entendida como parâmetro exclusivo de equilíbrio nutricional / Regarding mineral nutrition of plants, we highlight the similarities between nutrients boron (B) and calcium (Ca) in relation to similar functions they play in plant physiology. However, the mechanisms involved in the interaction between these nutrients are not very well known. Assuming that the interaction Ca - B affects the morphological, ultrastructural and biochemical attributes, as well as the nodulation in leguminous plants, the objective with this study was to determine the optimum Ca:B ratio, in the solution and plan tissue as well for dry matter yield of grain and dry matter of soybean, and to evaluate the effects of B - Ca interaction in the biological N2 fixation (nitrogenase activity - N-ase [EC 1.18.6.1]), root morphology (total root length and total root surface area), root anatomy (nodules), as well as in the ultrastructural evaluation of foliar mesophilic, and yield of oil content of soybean (cultivar BRS 284), grown in nutrient solution. Dry matter yield of plants, Ca-concentrations and amounts of Ca, B, Mg, K and N in shoots and roots were also evaluated. The experimental arrange used was an incomplete 5² factorial, making thirteen combinations of Ca (mmol L-1) and B (µmol L-1) rates, which were set in a randomized block design, with four replications. The highest N-ase activity was obtained with the highest doses of Ca and B (7,25 mmol L-1 of Ca and 50 µmol L-1 of B). The lowest Ca rate resulted in approximately 90% inhibition of N-ase. The highest total root length and total root surface area were obtained with combination of 7,25Ca and 12,5B rates, whereas the combination of the lowest Ca and B rates caused reduction of approximately 70% and 60% on root length and area, respectively. The highest dry matter and grain yield, highest accumulations of Ca, B, N, K and Mg on the leaves and the highest oil content in grains were obtained when 4,25 mmol L-1 of Ca and 12,5 µmol L-1 of B were supplied, resulting in a Ca:B ratio of approximately 350:1 on nutrient solution and 700:1 on leaves. The oil content in grains raised about 20% increasing B rates from 3,125 to 50 µmol L-1. The number of cell layers of nodule cortex raised with increasing Ca and B levels in the solution. It was observed membrane rupture of chloroplasts in leaf mesophyll and high number of N2-fixing bacteria entering the interior of nodular tissues when plants were supplied with the lowest Ca rate (1,75Ca) and the highest B (50B). The combination of the lowest B rate (3,125B) and the highest Ca rate (7,25) caused cell wall rupture and thickening of the middle lamella on the leaves. The increase of Ca levels on the solution mitigates the negative effects of the low B supply in plant biomass and grain yield, in root length and root surface area and in N-ase activity. The Ca:B ratio either in the solution or plant tissue, cannot be understood as an exclusive parameter of nutritional status