Concrete deterioration is a natural process, which has to be carefully monitored over a the service life of a structure. The Alkali-Silica Reaction (ASR) is a slow-process degradation, which devel-ops over decades and is therefore difficult to predict. Since its recognition 80 years ago, progress in understanding the fundamental mechanisms has been limited due to many parameters involved in the product formation, and the difficulty to isolate and test them separately. Such parameters include the concentrations of chemical elements in solution, the temperature and the water supply, which can evolve with time. This thesis focuses on characterizing the initial stages of ASR, in different conditions. First, early-stage products from OPC concrete in standard accelerated conditions are analyzed. The existence of two different ASR products was found. One has a granular morphology with an amorphous structure, and the second has a platey morphology with a partially crystalline structure. Both products show similar chemical compositions. Such a study at the nanometer scale was made possible using powerful mi-croscopy techniques which are Scanning Electron Microscopy (SEM) to localize the area of interest, Focused Ion Beam (FIB) to prepare ultra-thin lamellae that can be studied in Transmission Electron Microscopy (TEM) in Scanning mode (STEM) to carry out Energy Dispersive X-ray (EDX) measurements and calculate the compositions, coupled with Selected Area Electron Diffraction (SAED) to analyze the structure. Besides the first analyses, the influence of three parameters was tested. The temperature in the standard accelerated conditions, were lowered from 60°C to a more field realistic one of 38°C, to evaluate the impact on ASR procuct characteristics. At 38°C, only amorphous product was found, with different compositions, presumably due to the mixed analysis of a precursor form of the final ASR product (observable at low temperature) and the ASR product itself. In a second experiment, Supple-mentary Cementitious Materials (SCM) (fly ash and calcined clay), which are known to hinder concrete expansion, have been used as a partial replacement of OPC in the concrete mix. SCM showed no ex-pansion in the first accelerated months, and rare occurences of ASR products were found. The product composition didn't change (no aluminium uptake), although its Ca/(K+Na) ratio slightly increased in comparison to OPC samples. A third experiment studied the addition of alkali to a concrete mix con-taining calcined clays. No major change was observed in the early stage ASR product formation, de-spite the fact the product was more easily found in the sample. Finally, samples from the field and from synthetic production were also analyzed. The comparison be-tween field results and accelerated results was necessary to 1) validate the relevance of the acceler-ated tests and 2) evaluate the differences between early-stage and late-stage products found after the expansion and structure cracking has occurred. They were found to be similar in morphology (granular and platey), composition and structure, the field samples exhibiting an overall stronger crystalline behavior. The synthetic samples showed a similar morphology and composition but with mainly nano-crystalline to crystalline structure. The comparison validated the relevance and similarity with field and accelerated products. In is advised to focus in the future on amorphous to slightly nanocrystalline products, in