A B S T R A C TIn this study, 316L parts were fabricated with the selective laser melting additive layer manufacturing process using unidirectional laser scan to control their texture. The melt pool shape, microstructure and texture of three different cubic samples were analyzed and quantified using optical microscopy and electron back-scattered diffraction. The samples scanned along the shielding gas flow direction were shown to exhibit shallow conduction melt pools together with a strong {110} < 001 > Goss texture along the laser scanning direction. The sample prepared with a laser scan perpendicular to the gas flow direction had deeper melt pools, with a weaker {110} < 001 > Goss texture in addition to a < 100 > fiber texture parallel to the scanning direction. Correlations were proposed between the melt-pool geometry and overlap and the resulting texture. The decrease of the melt pool depth was assumed to be linked to local attenuation of the laser beam effective power density transmitted to the powder bed.

Metal powder bed fusion techniques can be used to build parts with complex internal and external geometries. Process parameters are optimized in order to obtain parts with low surface roughness and porosity, while maintaining a high productivity rate. The goal of this work is to quantify the sensitivity to internal and surface defects on the fatigue endurance of additively manufactured metallic parts. 316L Stainless Steel samples were fabricated through powder bed fusion using identical contour parameters, but three different hatching strategies were applied by varying the scanning speeds in the internal portions of the parts. Samples were subsequently mirror-polished to smooth the rough as-built surface. X-ray computed tomography analysis revealed several defect populations in samples from all three parametric sets due to lack of fusion in the bulk, with a nearly fully dense external "shell". High cycle fatigue tests at R = 0.1 were then performed on the specimens and combined with the X-ray computed tomography scans, helping to identify the largest and the critical defect size at which crack initiation occurred. Most fatigue failures initiated within the external contour zone for small (< 100 μm) defects, even when larger (> 200 μm) lack of fusion defects were widely present below the surface. It was determined that the high porosity (1% in volume or above 5% in area at some fabricated layers) observed in the bulk of parts manufactured with high scanning speeds had little impact on the fatigue limit of the material.

Thermal silicon oxide-to-oxide bonding was investigated at the nanometer level using X-ray reflectivity, transmission electron microscopy, and infrared absorption spectroscopy. The measurements reveal the stages of the closure mechanism, which are different from standard silicon bonding. Upon annealing, interface water pockets are formed, the contents of which are further dissolved into the oxide, demonstrating that the buried thermal oxide-silicon interface acts as a barrier against water reaction with silicon.Direct wafer bonding consists in joining two wafers at room temperature without any adhesive or additional materials, followed by annealing. Despite a wide technological interest, the detailed mechanism of the sealing of such model solids is poorly known due to the lack of experimental tools to investigate a nanometer-wide interface buried under millimeter-thick materials. Yet, the bonding technology is increasingly used in microelectronics, e.g., for siliconon-insulator mass production, microelectromechanical systems manufacturing, or in three-dimensional level integration. In many cases, it represents an alternative to deposition for building heterostructures with additional capabilities. For most applications, high quality bonding is required, driving many studies on bonding interfaces. 1-4 Water removal has been widely described in the literature through H 2 outgassing due to a low temperature oxidation occurring through the native oxide. We show here that although macroscopically oxide/oxide, silicon/oxide, and silicon/silicon bondings show similar behavior, the interface evolutions at the nanometer scale are entirely different. We used interfacial X-ray reflectivity ͑XRR͒, 5 high resolution transmission electron microscopy ͑HR-TEM͒, and Fourier transform infrared spectroscopy in multiple internal reflection ͑FTIR-MIR͒ 6 to demonstrate our point. The results of similar techniques applied to silicon/silicon or silicon/oxide bonding can be found for comparison in Ref. 7.The wafers used were 100 mm in diameter ͑001͒-oriented Czochralski-grown Si wafers. HF-etch was first performed on the wafers to remove the native oxide from the surface. Then the oxide films were thermally grown on the Si-H terminated wafers at 800°C in dry O 2 flow. The typical SiO 2 thickness used in this study was 10 nm. Wafers were then cleaned in sulfoperoxide mixture ͑H 2 SO 4 ,H 2 O 2 ͒ and rinsed in deionized water. An RCA cleaning treatment was then performed, followed by a deionized water rinse and spin-drying. Finally, hydrophilic wafers were bonded at room temperature in a clean room atmosphere and annealed at various temperatures in the range from room temperature to 400°C for 2 h. Figure 1 presents the evolutions of ͑a͒ bonding interface density and ͑b͒ width as a function of the annealing temperature, obtained from XRR analyses. Above 100°C, together with the bonding energy, our data show that the interface density sharply increases corresponding to interface closure. This mechanism of interface sealing in wafer bonding h...

The development of microcracks in hydrogen-implanted silicon has been studied up to the final split using optical microscopy and mass spectroscopy. It is shown that the amount of gas released when splitting the material is proportional to the surface area of microcracks. This observation is interpreted as a signature of a vertical collection of the available gas. The development of microcracks is modeled taking into account both diffusion and mechanical crack propagation. The model reproduces many experimental observations such as the dependence of split time upon temperature and implanted dose.

Hydrogen implanted silicon has been studied using high resolution X-ray scattering. Strain induced by implantation has been measured as a function of implantation dose. The dependence of strain with implanted dose shows different regimes starting from linear to quadratic and saturation. The observed strain is consistent with ab-initio and elasticity calculations. Strain rate changes can be associated to the predominant location of hydrogen in bond center location.

International audienceCrack propagation in implanted silicon for thin layer transfer is experimentally studied. The crack propagation velocity as a function of split temperature is measured using a designed optical setup. Interferometric measurement of the gap opening is performed dynamically and shows an oscillatory crack "wake" with a typical wavelength in the centimetre range. The dynamics of this motion is modelled using beam elasticity and thermodynamics. The modelling demonstrates the key role of external atmospheric pressure during crack propagation. A quantification of the amount of gas trapped inside pre-existing microcracks and released during the fracture is made possible, with results consistent with previous studies. (C) 2015 AIP Publishing LLC

Interface defects formed during the wafer bonding process upon annealing have been studied. Based on the hydrogen diffusion in SiO2 and the stability of the bubbles at the bonding interface, models of the growth and further dissolution of the defects are presented. Considering the hydrogen diffusion through the interfacial oxide, diffusion coefficients and activation energy (Ea=0.25 eV) are obtained. It has been shown that the defect dissolution is driven by the exodiffusion of the hydrogen toward the silicon substrates.

No abstract