Search

International audience

The mechanisms for the formation of the systems 〈Au/InP(100)〉 and 〈Au/InSb/InP(100)〉 are connected with their electrical properties. By using quantitative AES, a precise knowledge of the atomic mechanisms allows us to perform the desired component such as a Schottky diode. The best results are obtained for the 〈Au/heated InSb/InP〉 system with a barrier height of 0.7 eV.

We report the first synthesis of GaAs nanowires (NWs) by Au-assisted vapor-liquid-solid (VLS) growth in the novel hydride vapor phase epitaxy (HVPE) environment. Forty micrometer long rodlike monocrystalline GaAs nanowires exhibiting a cubic zinc blende structure were grown in 15 min with a mean density of 10(6) cm(-2). The synthesis of such long figures in such a short duration could be explained by the growth physics of near-equilibrium HVPE. VLS-HVPE is mainly based on solidification after direct and continuous feeding of the arsenious and GaCl growth precursors through the Au-Ga liquid catalyst. Fast solidification (170 microm/h) is then assisted by the high decomposition frequency of GaCl. This predominant feeding through the liquid-solid interface with no mass and kinetic hindrance favors axial rather than radial growth, leading to twin-free nanowires with a constant cylinder shape over unusual length. The achievement of GaAs NWs several tens of micrometers long showing a high surface to volume ratio may open the field of III-V wires, as already addressed with ultralong Si nanowires.

The model of quantitative interpretation of Auger electron spectroscopy (AES) results is described and some complementary electron energy loss spectroscopy (EELS) results are also reported. It is shown that the InP(100) surfaces perturbed by the Ar+ cleaning treatment can be ordered by Sb deposition. The variations of ditferent Auger lines are interpreted, and the transformations 3D → 2D of the initially formed In clusters can be well-followed during the frrst stages of the deposition. PACS numbers: 68.35.Βs, 68.35.Fx, 72.80.Ey, 82.80.Pv

International audience

Quantitative and qualitative analysis techniques were employed to study the first stages of ultra-high vacuum plasma nitriding of the 42CrMo4 steel. At constant treatment temperature, maintained for all samples at about 360 °C, we have established the influence of treatment time on the chemical composition, thickness and electrical properties of the nitrided layer. In this purpose it was used a stacking atomic layer model describing the sample surface, which takes into account the attenuation depth of photoelectrons by the atomic monolayers. So, we have found that after 2 h of nitriding in laboratory conditions, 70% of the nitrided layer was composed of iron oxide. Also, I – V measurements indicate an influence of the nitride overlayer with increasing treatment time.

International audience; The nitridation of InP(1 0 0) surfaces has been studied using synchrotron radiation photoemission. The samples were chemically cleaned and then ion bombarded, which cleaned the surface and also induced the formation of metallic indium droplets. The nitridation with a Glow Discharge Cell (GDS) produced indium nitride by reaction with these indium clusters. We used the In 4d and P 2p core levels to monitor the chemical state of the surface and the coverage of the species present. We observed the creation of In-N and P-N bonds while the In-In metallic bonds decrease which confirm the reaction between indium clusters and nitrogen species. A theoretical model based on stacked layers allows us to assert that almost two monolayers of indium nitride are produced. The effect of annealing on the nitridated layers at 450 °C has also been analysed. It appears that this system is stable up to this temperature, well above the congruent evaporation temperature (370 °C) of clean InP(1 0 0): no increase of metallic indium bonds due to decomposition of the substrate is detected as shown in previous works [L. Bideux, Y. Ould-Metidji, B. Gruzza, V. Matolin, Surf. Interface Anal. 34 (2002) 712] studying the InP(1 0 0) surfaces.

International audience; The nitrides of group III metals: AlN, GaN and InN are very important materials due to their applications for short wavelength opto-electronics (light-emitting diodes and laser diodes). It is essential for the realization of such novel devices to grow high-quality nitride single crystals. In this paper, we report the first stages of the InP(1 0 0) surfaces nitridation in order to grow high-quality nitride films. Indeed, the nitridation process is an important step in the growth of nitrides [J. Vac. Sci. Technol. A 17 (1999) 2194; Phys. Status Solidi A 176 (1999) 595]. Previous works [Synth. Met. 90 (1997) 2233; Appl. Phys. Lett. 63 (1993) 1957] have shown that in situ Ar+ ions bombardment is useful on the one hand to clean the surface, and on the other hand to create droplets of metallic indium in well-controlled quantity. Then the indium metallic enrichment of the surface, monitoring by elastic peak electron spectroscopy (EPES) and Auger electron spectroscopy (AES) allows to prepare the III-V semiconductors surfaces to the nitridation step. The nitridated process has been performed with a high voltage plasma discharge cell and has been studied using quantitative Auger electron spectroscopy, elastic peak electron spectroscopy and electron energy loss spectroscopy (EELS), in order to optimize the conditions of InN layers formation.

Porous GaAs layers were formed by electrochemical etching of p-type GaAs(1 0 0) substrates in HF solution. A surface characterization has been performed on p-type GaAs samples using X-ray photoelectron spectroscopy (XPS) technique in order to get information about the chemical composition, particularly on the surface contamination. According to the XPS spectra, the oxide layer on as-received porous GaAs substrates contains As2O3, As2O5 and Ga2O3. Large amount of oxygen is present at the surface before the surface cleaning. Compared to untreated GaAs surface, room temperature photoluminescence (PL) investigations of the porous layers reveal the presence of two PL bands: a PL peak at ∼871 nm and a “visible” PL peak at ∼650–680 nm. Both peak wavelengths and intensities varied from sample to sample depending on the treatment that the samples have undergone. The short PL wavelength at 650–680 nm of the porous layers is attributed to quantum confinement effects in GaAs nano-crystallites. The surface morphology of porous GaAs has been studied using atomic force microscopy (AFM). Nano-sized crystallites were observed on the porous GaAs surface. An estimation of the mean size of the GaAs nano-crystals obtained from effective mass theory and based on PL data was close to the lowest value obtained from the AFM results.

EPES (elastic peak electron spectroscopy) and AREPES (angle resolved elastic peak electron spectroscopy) are non destructive methods and very sensitive to the surface region. These techniques allow to measure the percentage η e of elastically backscattered electrons from the surface excited by an electron beam. Both methods are combined with Monte-Carlo (MC) simulations to interpret experimental results. In this work, we underline the importance of a micrometric scale roughness at the surface. The use of an original Monte-Carlo method was fruitful for the simulation, moreover 3D representations have been developed for visualization and qualitative interpretation of the results. For a more precise quantitative study, a 2D representation was necessary. The calculated results have been compared with published experimental ones got for different incidence angles and primary energies, on a silicon surface having triangular saw-tooth aspect (crenels) obtained by photolithography. We have observed that the effect due to the roughness increases with the incidence angle.

Quantitative X-ray photoelectron spectroscopy was used to characterize the native oxide film formed on 42CrMo4 steel surface by air exposure in normal conditions. In order to determine the thickness and composition of the oxide layer we have used a stacking layer model together with experimental XPS sputtering depth profiling. At a nanoscale study, to obtain quantitative results one must take into account fundamental parameters like the attenuation depth of photoelectrons. We have found that both lepidocrocit (γ-FeOOH) and magnetite (Fe3O4) were present and the total thickness of the oxide layer was 16 monolayers.

In this study, InP(1 0 0) surfaces were bombarded by argon ions in ultra high vacuum. Indium metallic droplets were created in well controlled quantities and played the role of precursors for the nitridation process. A glow discharge cell was used to produce a continuous plasma with a majority of N atomic species. X-ray photoelectron spectroscopy (XPS) studies indicated that the nitrogen combined with indium surface atoms to create InN thin films (two monolayers) on an In rich-InP(1 0 0) surface. This process occurred at low temperature: 250 °C. Synchrotron radiation photoemission (SR-XPS) studies of the valence band spectra, LEED and EELS measurements show an evolution of surface species and the effect of a 450 °C annealing of the InN/InP structures. The results reveal that annealing allows the crystallization of the thin InN layers, while the LEED pattern shows a (4 × 1) reconstruction. As a consequence, InN related structures in EELS and valence bands spectra are different before and after the annealing. According to SR-XPS measurements, the Fermi level is found to be pinned at 1.6 eV above the valence band maximum (VBM).

The gas sensing, electrical and chemical properties of a novel toxic gas sensor structure based on n-InP epitaxial layers have been investigated. The gas sensing properties were investigated under extremely low concentration of toxic gas flow, nitrogen dioxide, in dry pure air. Responses, resistance changes ∃R, towards NO2 of the structures with different active InP layer thickness were compared. The influence of the surface states at InP surface on the initial resistance of the sensor structure was shown from numerical simulations. Furthermore, the changes of InP surface chemical properties due to gas exposition were stud-ied using the X-ray photoelectron spectroscopy (XPS).

The percentage, η e ′ of electrons elastically reflected by a substrate is a fundamental parameter for elastic peak electron spectroscopy (EPES). A Retarding Field Analyser with its large acceptance angle can be used for this spectroscopy. However, distortion is observed in the low energy side of the recorded elastic peak, increasing with the energy of the impinging electrons, which introduces difficulty for η e measurement. To eliminate this difficulty, we propose in this paper a method to obtain η e using a reference curve. The process is applied to the study of the Ag Mo system, as well as experimental determination of the mean free path of electrons in indium samples.

5-9 Septembre 2005

The structure and sensing properties of a novel resistive NO2 sensor based on n-type InP epitaxial layers have been presented. The studies of sensor resistance changes due to adsorbed gas NO2 under exposures in the range from 20 to 100 ppb at a temperature of 80°C were performed. The thickness of the active InP layer changed from 0.2 to 0.4 Μm. The response time and signal stability was also investigated. Furthermore, the influence of surface states and near-surface region on sensor parameters in terms of the resistance relative changes was shown from numerical simulations. The analysis of the measured photoelectron spectroscopy (XPS) spectra confirmed the complex chemical composition of the InP oxides, which give rise to the high density of surface states.

This article investigates the nitridation effect of InP(100) semiconductor surfaces performed using a glow discharge cell. Auger electron spectroscopy and XPS were used to understand the different steps of this process. An important point is the initial quantity of metallic indium on the InP(100) surfaces. Indeed the indium droplets, created in well-known quantity, play the role of precursor. At a relatively low temperature of 523 K the system undergoes surface restructuring, which includes removal of the In droplets and the formation of two InN monolayers. Phosphorus-nitrogen bonds have been detected by the analysis of P LMM Auger peaks, and In-N bonds by analysis of the In 4d XPS peak. However, the presence or not of metallic indium inside this InN overlayer is crucial for passivation of the substrate. Ex situ photoluminescence measurements correlated to the electron spectroscopies results have shown the good passivation effect of the InP(100) surfaces by InN overlayers for 40 min of nitridation.

International audience; Auger electron spectroscopy (AES) was used to investigate the processes taking place during the initial stages of InP(100) surfaces nitridation. This AES study combined with electrical measurements (intensity-potential) shows that the processes greatly differ depending on the nitridation angles. Results show that with grazing angle for nitrogen flow, the nitridation process is more efficient. Results obtained with AES spectra are coherent with electrical measurements : Hg/InN/InP(100) Schottky diodes present better electrical characteristics in the case of a grazing flow. That means, the adsorption of nitrogen on the surface is more important for this configuration.

The cleaning and the nitridation of GaAs(1 0 0) substrates have been studied in this work by electron energy loss spectroscopy (EELS), elastic peak electron spectroscopy (EPES) and X-ray photoelectron spectroscopy (XPS). The nitridation of GaAs(1 0 0) was performed using a double differential pumping RF cell. We have studied in a first part, the surface modifications of the substrate during ionic cleaning. In a second part, we have determined the λ i ( E ) K ( E , E n ) curves from EELS spectra using Tougaard's procedure for cleaned-GaAs(1 0 0) and nitrited-GaAs(1 0 0) substrates. The thickness of the GaN overlayer was deduced from these results.

A double chamber vacuum system was constructed for the electron spectroscopy study of GaN layer growth. The samples are prepared by vacuum deposition from Knudsen-type Ga source in the afterglow atmosphere of active nitrogen produced by high-pressure RF-plasma source. The growth chamber is connected to the analysis chamber equipped with electron spectroscopy (XPS, AES) analysis, permitting to study the chemical composition of layers as a function of growing conditions. It is shown that the new RF-source is suitable for preparation of pure stoichiometric GaN layers.

The paper deals with porous Si (PSi) samples, prepared with different parameters (substrate, current density, solution, etc.) of anodization and anodic oxidation. The porosity of samples varied between 36 and 78%, their thickness was 1–5 p.m. The samples were studied in six laboratories using complementary methods. The spectra of elastic reflection coefficients r e (E p ) are characteristic of the composition and they are affected by the structure of the surface layer. They were measured with a retarding field analyser using spectrometer correction and deduced from the elastic peak versus E p primary energy. The r e (E p ) showed a dramatic decrease in the 50–100 eV range with H adatoms on the surface, produced by HF treatment. It is caused by attenuation and multiple elastic reflections of electrons on the porous surface with H adatoms. An oxidized sample showed a contrary effect, which can be explained by the elastic reflection cross-sections of Si and O in this low energy range. Valence band XPS spectra were measured with a KRATOS XSAM 800 analyser. They showed changes with HF treatment and porous structure. To avoid difficulties with STM due to the oxide, the surface topography was mapped by atomic force microscopy, resulting in 0.3–0.9 nm roughness on samples of 36% porosity.

The retarding field analyser (RFA) proved to be advantageous in electron spectroscopy due to its high luminosity and by providing LEED simultaneously with AES and EPES. RFA is very efficient for determining the inelastic mean free path of electrons by EPES. Dolinski et al measured the elastic peak by repelling the electron beam using a bias potential on the sample. In this paper a further development of this procedure is described supplying the true percentage of elastically scattered electrons collected by RFA. The percentage has been determined by our RFA with optimized bias potential of the sample for primary energies between 100 and 2000 eV. It was deduced from the high energy side of the elastic peak because the background affects the detected signal on the low energy side. The experimental curve of elastic reflection coefficient is presented for indium. The measured elastic peak is distorted by the analyser. The spectrometer response of an RFA was determined by analysing the peak shape of deflected electrons after optimization of the bias voltage of the sample. The peak shape shows asymmetry since the low energy side differs from the Gaussian like high energy side. The distortion produced by RFA is similar to that observed by Taylor. It is enhanced by the background adjacent to the elastic peak. A computer algorithm was elaborated for reconstruction of the elastic and Auger peaks. Corrected curve is presented for indium Auger peak.

Our Monte Carlo simulation for electron trajectories in the crystal used for EPES has been improved for all angular studies. The percentage of reflected electrons can be obtained for different directions of the primary beam and also for different backscattering angles. The simulation is compared with experimental results performed on a pure copper sample.

Ion beam nitridation of indium phosphide (100) substrates was studied using x-ray photoelectron spectroscopy (XPS) and Auger electron spectroscopy (AES). Argon ion bombardment of the InP(100) surface removes the native oxides and induces the formation of three-dimensional indium clusters. After cleaning, the samples were nitrided in an ultrahigh vacuum chamber using a home-made radiofrequency plasma source (13.56 MHz) that allows nitridation at low pressures (10−4 Pa). We have studied the influence of temperature on the nitridation mechanisms. For low temperature (T < 200°C), no variation of the metallic indium droplets was observed; for high temperature (T > 270°C), the first layers of the substrate were damaged. The optimal temperature for the nitridation of InP(100) was found to be 250°C.

This paper deals with both the theoretical and experimental analysis of electronic properties of n-type indium phosphide (n-InP) epitaxial layers covered by native oxides for gas sensor applications. The rigorous calculations of the influence of continuous surface states with U-shaped density distribution NSS(E) as well as surface fixed charge (QFC) representing adsorbed ionized species or surface ?-doping on the conductivity of n-InP layers were carried out. The surface state density minimum (NSS0) was changed from 1011 to 1013 eV-1cm-2. From the Hall effect measurements upon NO2 exposition the concentration of charge carriers was determined as a function of gas concentration. Additionally, the roughness of the n-InP surface before and after gas action was investigated by using AFM.

The response time and signal stability were also investigated. Furthermore, the influence of surface states and near-surface region on sensor parameters in terms of the resistance relative changes was shown from numerical simulations. The analysis of the measured photoelectron spectroscopy (XPS) spectra confirmed the complex chemical composition of the InP oxides, which gives rise to the high density of surface states.

Many key technologies, to overcome some crucial steps, are needed for space missions. One of them concerns the re-entry: when the vehicle enters the atmosphere layer, its high velocity relative to the ground will generate significant friction and an increase in pressure. The kinetic energy of the object will be converted into heat energy and heat the gas forming a shock wave. Temperatures depend on the initial velocity, the atmosphere composition and its pressure. The gas is ionized and become plasma, it will therefore transfer its heat to the body of the probe not only by convection but also by radiation. To protect itself, the vehicle has a heat shield that can withstand extreme phenomena encountered during the descent. Additional constraints will impose a particular geometry and a heavy shield. So the goal is to have the lightest possible and effective protections to allow the probe to maximize its payload. For this, one of the key parameters is to know the behavior of the plasma and radiation produced during re-entry into the atmosphere. This thesis is positioned in this area of ​​study: experiments to create a re-entry plasma are intended to be observed by a spectrometer., La mise en œuvre de missions spatiales demande à développer de nombreuses technologies clés afin de surmonter certaines étapes cruciales. L’une d’entre elle concerne la rentrée atmosphérique : lorsque le véhicule pénètre dans la couche d’atmosphère, sa vitesse relative très grande par rapport au sol va engendrer des frottements importants ainsi qu’une augmentation de pression. L’énergie cinétique de l’objet va donc être transformée en énergie thermique et chauffer le gaz en formant une onde de choc. Les températures dépendent de la vitesse initiale, de la composition de l’atmosphère et de sa pression. Le gaz ainsi chauffé va s’ioniser et devenir un plasma, il va donc transférer sa chaleur au corps de la sonde non seulement par convection mais aussi par rayonnement. Afin de se protéger, le véhicule dispose d’un bouclier thermique pouvant résister aux phénomènes extrêmes rencontrés lors de la descente. Les contraintes supplémentaires vont imposer une géométrie particulière et un lourd bouclier. Ainsi l’objectif est d’avoir les protections les plus légères et efficaces possibles afin de permettre à la sonde de maximiser son emport en équipement scientifique. Pour cela, un des paramètres clés est de connaître le comportement et le rayonnement du plasma produit lors de la rentrée dans l’atmosphère. Cette thèse se positionne dans ce domaine d’étude: les expériences montées et réalisées ont pour objectif d’observer par moyens spectroscopiques un plasma correspondant à celui rencontré par les sondes spatiales.

This thesis focuses on two aspect of GaAs surface treatment using plasma in order to deposit controlled oxide and nitride layers. In both cases, various means of analysis were used to control the chemical composition and the thickness and the structure of produced layers. Theoretical calculations of the XPS signals established on a schematic representation of the samples allow the understanding the involved phenomena. The first study reported in this thesis has demonstrated the efficiency of a microwave plasma composed of O2 and SF6 for the cleaning of GaAs substrates having undergone one or several technological processes. The variation of plasma parameters such as composition, power and exposure time enables us to better understand the mechanisms of oxidation occurring on the surface. Particularly, the presence of SF6 in the plasma induces the decrease of the concentration of arsenic on the surface and thus improves the stability of the oxide. The second study of this work presents the nitridation of GaAs substrates with different orientations. More the benefits of using a low power GDS source to create high quality GaN layer on the surface is highlighted. This optimized process leads to an As-free nitride layer formed by the inter-diffusion of arsenic and nitrogen, and shows a near saturation of the nitride layer thickness as a function of exposure time. Furthermore, this layer is crystallized by annealing in a cubic lattice parameter close to that of c-GaN.; Ce mémoire porte sur deux aspects du traitement des surfaces de GaAs faisant appel à des plasmas, dans le but de déposer des couches contrôlées d’oxyde et de nitrure. Dans les deux cas divers moyens d’analyse ont été utilisés pour contrôler la composition chimique ainsi que l’épaisseur et la structure des couches créées. Des calculs théoriques des signaux XPS établis sur une représentation schématique des échantillons ont été à la base de la compréhension des phénomènes mis en jeu. La première étude exposée dans ce mémoire a permis de démontrer l’efficacité d’un plasma micro-ondes composé d’O2 et de SF6 pour le nettoyage de substrats de GaAs ayant subis un ou plusieurs processus technologiques. La variation de paramètres comme la composition ou la puissance du plasma et le temps d'exposition à celui-ci nous a permis de mieux comprendre les mécanismes d’oxydation se déroulant sur la surface. Notamment, la présence de SF6 dans le plasma diminue la concentration d’arsenic sur la surface et améliore ainsi la stabilité de l’oxyde. La seconde étude réalisée dans ce travail, qui présente la nitruration de substrats de GaAs suivant différentes orientations, a mis en évidence les avantages que procure l’utilisation d’une source GDS à faible puissance pour la création d’une couche surfacique de GaN de très bonne qualité. Ce processus optimisé amène à une couche de nitrure exempte de composés dus à l’inter-diffusion de l’arsenic et de l’azote, et montre une quasi-saturation de l’épaisseur de la couche de nitrure en fonction du temps d’exposition. De plus cette couche se cristallise par recuit en une maille cubique de paramètre proche de celui du GaN.

EPES (Elastic Peak Electron Spectroscopy) allows measuring the percentage he of elastically backscattered electrons from the surface excited by an electron beam. This is a non destructive method which is very sensitive to the surface region. The aim of this work is to model the trajectory of elastic electrons in the matter with a computer simulation based on Monte Carlo method. This simulation allows interpreting experimental results of the EPES spectroscopy. We have moreover adapted this simulation for different surfaces transformed to micrometer and nanometer scales. Using an original method, based on a description of material layer by layer, I realized a computer program (MC1) that takes into account the path of elastic electrons in different layers of material. The number of electrons emerging from the surface depends on many parameters such as: the electron primary energy, the nature of the material, the incidence angle and the collection angles of the analyzer. In addition, I was interested in the effect of surface roughness and I showed that it plays an important role in the intensity of the elastic peak. Then, through an association of the EPES and the Monte Carlo simulation results, I deduced the growth patterns of gold on silver and copper substrates. The effects of the atomic arrangement and the surface excitations were then studied. For this, a new simulation MC2 that takes into account these two parameters has been developed to study nanoscale surfaces. These parameters not previously included in our MC1simulation play a important role in the elastic intensity. Then I have got a simple formula for interpreting the results obtained by the simulation for a RFA analyzer. To validate the different results of the simulation MC2, I realized nano-structured silicon surfaces, using aluminium oxide masks. Nano-pores have been created by Ar+ ions bombardment in UHV chamber on silicon surfaces.To control the morphology of the surfaces, I realized SEM images (Techinauv Casimir) ex-situ. The Monte Carlo simulations, developed here, associated with the EPES experimental results can estimate the depth, the diameter, the morphology of pores without the help of other ex-situ techniques., La spectroscopie EPES (Elastic Peak Electron Spectroscopy) permet de mesurer le pourcentage he d’électrons rétrodiffusés élastiquement par la surface d’un échantillon soumis à un bombardement électronique. C’est une méthode non destructive et extrêmement sensible à la surface. L'objectif de ce travail est de modéliser le cheminement des électrons élastiques dans la matière grâce à une simulation informatique basée sur la méthode Monte Carlo. Cette simulation contribue de manière essentielle à la connaissance et à l'interprétation des résultats expérimentaux obtenus par spectroscopie EPES. Nous avons, de plus, adapté cette simulation à différentes surfaces transformées à l’échelle micrométrique et nanométrique. A l’aide d’une méthode originale, basée sur une description couche par couche du matériau, j’ai réalisé un programme informatique (MC1) rendant compte du cheminement des électrons élastiques dans les différentes couches du matériau. Le nombre d’électrons ressortant de la surface dépend de nombreux paramètres comme : la nature du matériau à étudier, l’énergie des électrons incidents, l’angle d’incidence, les angles de collection des analyseurs. De plus, je me suis intéressé à l’effet de la rugosité de la surface et j’ai démontré qu’elle joue un rôle déterminant sur l’intensité du pic élastique. Ensuite, grâce à l’association de la spectroscopie EPES et de la simulation Monte Carlo, j’ai déduit les modes de croissance de l’or sur substrat d’argent et de cuivre. Les effets de l’arrangement atomique et des pertes énergétiques de surfaces ont ensuite été étudiés. Pour cela, une deuxième simulation MC2 tenant compte de ces deux paramètres a été réalisée permettant d’étudier les surfaces à l’échelle nanométriques. Ces paramètres jusqu’alors non pris en compte dans notre simulation MC1, joue un rôle essentiel sur l’intensité élastique. Ensuite, j’ai obtenu une formulation simple et exploitable pour l’interprétation des résultats obtenus par la simulation MC2 pour un analyseur RFA. Afin de valider, les différents résultats de la simulationMC2, j’ai réalisé des surfaces de silicium nanostructurées, à l’aide de masques d’oxyde d’alumine réalisés par voie électrochimique. J’ai pu créer des nano-pores par bombardement ionique sous ultravide sur des surfaces de silicium. Afin de contrôler la morphologie de la surface, j’ai effectué de l’imagerie MEB ex-situ. La simulation Monte Carlo développée associée aux résultats EPES expérimentaux permet d’estimer la profondeur, le diamètre et la morphologie des pores sans avoir recours à d’autres techniques ex-situ.Cette simulation MC2 permet de connaître la surface étudiée à l’échelle nanométrique.

EPES (Elastic Peak Electron Spectroscopy) allows measuring the percentage he of elastically backscattered electrons from the surface excited by an electron beam. This is a non destructive method which is very sensitive to the surface region. The aim of this work is to model the trajectory of elastic electrons in the matter with a computer simulation based on Monte Carlo method. This simulation allows interpreting experimental results of the EPES spectroscopy. We have moreover adapted this simulation for different surfaces transformed to micrometer and nanometer scales. Using an original method, based on a description of material layer by layer, I realized a computer program (MC1) that takes into account the path of elastic electrons in different layers of material. The number of electrons emerging from the surface depends on many parameters such as: the electron primary energy, the nature of the material, the incidence angle and the collection angles of the analyzer. In addition, I was interested in the effect of surface roughness and I showed that it plays an important role in the intensity of the elastic peak. Then, through an association of the EPES and the Monte Carlo simulation results, I deduced the growth patterns of gold on silver and copper substrates. The effects of the atomic arrangement and the surface excitations were then studied. For this, a new simulation MC2 that takes into account these two parameters has been developed to study nanoscale surfaces. These parameters not previously included in our MC1simulation play a important role in the elastic intensity. Then I have got a simple formula for interpreting the results obtained by the simulation for a RFA analyzer. To validate the different results of the simulation MC2, I realized nano-structured silicon surfaces, using aluminium oxide masks. Nano-pores have been created by Ar+ ions bombardment in UHV chamber on silicon surfaces.To control the morphology of the surfaces, I realized SEM images (Techinauv Casimir) ex-situ. The Monte Carlo simulations, developed here, associated with the EPES experimental results can estimate the depth, the diameter, the morphology of pores without the help of other ex-situ techniques.; La spectroscopie EPES (Elastic Peak Electron Spectroscopy) permet de mesurer le pourcentage he d’électrons rétrodiffusés élastiquement par la surface d’un échantillon soumis à un bombardement électronique. C’est une méthode non destructive et extrêmement sensible à la surface. L'objectif de ce travail est de modéliser le cheminement des électrons élastiques dans la matière grâce à une simulation informatique basée sur la méthode Monte Carlo. Cette simulation contribue de manière essentielle à la connaissance et à l'interprétation des résultats expérimentaux obtenus par spectroscopie EPES. Nous avons, de plus, adapté cette simulation à différentes surfaces transformées à l’échelle micrométrique et nanométrique. A l’aide d’une méthode originale, basée sur une description couche par couche du matériau, j’ai réalisé un programme informatique (MC1) rendant compte du cheminement des électrons élastiques dans les différentes couches du matériau. Le nombre d’électrons ressortant de la surface dépend de nombreux paramètres comme : la nature du matériau à étudier, l’énergie des électrons incidents, l’angle d’incidence, les angles de collection des analyseurs. De plus, je me suis intéressé à l’effet de la rugosité de la surface et j’ai démontré qu’elle joue un rôle déterminant sur l’intensité du pic élastique. Ensuite, grâce à l’association de la spectroscopie EPES et de la simulation Monte Carlo, j’ai déduit les modes de croissance de l’or sur substrat d’argent et de cuivre. Les effets de l’arrangement atomique et des pertes énergétiques de surfaces ont ensuite été étudiés. Pour cela, une deuxième simulation MC2 tenant compte de ces deux paramètres a été réalisée permettant d’étudier les surfaces à l’échelle nanométriques. Ces paramètres jusqu’alors non pris en compte dans notre simulation MC1, joue un rôle essentiel sur l’intensité élastique. Ensuite, j’ai obtenu une formulation simple et exploitable pour l’interprétation des résultats obtenus par la simulation MC2 pour un analyseur RFA. Afin de valider, les différents résultats de la simulationMC2, j’ai réalisé des surfaces de silicium nanostructurées, à l’aide de masques d’oxyde d’alumine réalisés par voie électrochimique. J’ai pu créer des nano-pores par bombardement ionique sous ultravide sur des surfaces de silicium. Afin de contrôler la morphologie de la surface, j’ai effectué de l’imagerie MEB ex-situ. La simulation Monte Carlo développée associée aux résultats EPES expérimentaux permet d’estimer la profondeur, le diamètre et la morphologie des pores sans avoir recours à d’autres techniques ex-situ.Cette simulation MC2 permet de connaître la surface étudiée à l’échelle nanométrique.