In recent years, superhydrophobic surfaces have attracted a considerable amount of attention from the academic and industrial communities. This growing interest is mainly caused by the fundamental physico-chemical theoretical aspects that remain obscure and the number of promising practical applications emerging in a wide range of fields, such as textiles, self-cleaning coatings, and micro-fluidic systems. Superhydrophobic surfaces exhibit simultaneously high static water contact angles and low resistance to liquid motion on the surface (i.e. low contact angle hysteresis). These properties result from the combination of the chemical hydrophobicity of the topmost layers of the surface and its roughness, the latter being the dominant factor. In this PhD work, advantage has been taken of recent advances in micro-/nano-processing technologies to fabricate microstructured surfaces with specific and controlled roughness. This has enabled systematic experimental investigations to be carried out to address some of the still unanswered questions relating to superhydrophobic phenomena. Silicon wafers were microstructured by photolithography in order to obtain periodical distributions of well-defined flat-top obstacles. A gas-phase silanization process was used to cover the prepared microstructured surfaces with a hydrophobic dense mono-layer of perfluorodecyltrichlorosilane. Several series of samples in which each roughness parameter (distance between obstacles, obstacle height, obstacle size, obstacle shape, etc.) was individually varied were fabricated, and static and dynamic contact angle variations as a function of each parameter were studied. The results obtained by water contact angle measurements were compared to the classic Wenzel and Cassie models. The first assumes that the liquid wets the asperities of the rough substrate completely (referred to as wetted state), whilst the second describes the liquid as sitting on a mixture of air and solid (referred to as composite state). By systematically varying a given parameter, a transition between the composite and the wetted regime was observed. Simple thermodynamic considerations based on the energy minimization of the drop-substrate system showed that Cassie and Wenzel contact angles correspond to two energy minima of the system. Experimentally, the Cassie regime is observed when the Cassie angle is the absolute energy minimum, while when the Wenzel angle is the absolute energy minimum, either the Wenzel regime or a metastable Cassie state is observed. The existence of these metastable states is explained theoretically by the energy barrier that the system has to overcome to reach the Wenzel state when a drop is gently deposited on the surface, and good agreement with experimental data is demonstrated. Water and n-hexadecane dynamic contact angle measurements performed in two different configurations, i.e. with a "negligible" and a "NON-negligible" extra-pressure exerted on the drop-substrate system, permitted the validity of the theoretical model proposed to be confirmed. They clearly evidenced the region of stability of both Wenzel and Cassie regimes, and particularly the region where the Cassie regime is metastable, which is the really interesting one for all the potential applications of rough superhydrophobic surfaces as self-cleaning, and non-sticking surfaces, since only the metastable composite states are those characterized by a very high contact angle and a very low contact angle hysteresis, the two main requirements for a good superhydrophobic surface. With systematic reduction of the absolute size of the obstacles in the micrometer range, an increase in the extent and robustness of the composite states was observed. Drop contact line fragmentation seems to be the most important factor in determining the extent and robustness of the observed composite states. Results obtained by varying the obstacle top-surface shape demonstrate that contact line length and corrugations are only secondary factors. Studies on the influence of asperity absolute size performed in the micrometer range, together with preliminary but fundamental results obtained on nanostructured substrates, allow the proposal of interesting conclusions concerning determination of roughness size where macroscopic superhydrophobicity appears and vanishes. A typical size of ∼ 50 µm (where "typical size" refers to the width of the obstacle top-surface) can be considered the upper limit of the superhydrophobicity scale range, with a length scale of some tens of nanometers as the lower limit. In this range, the length scale comprised between several hundred nanometers and one micrometer can be assumed the most suitable size for obtaining superhydrophobic properties of larger extent and robustness. The related appropriate asperity height must be sufficiently great to prevent the drop meniscus from touching the bottom of the asperities by vibration, inducing a transition to the wetted Wenzel state.