Reconfigurable intelligent surface (RIS) is a promising technology for wireless communications applications. A RIS is a thin planar array that consists of multiple reflecting elements, each of which is connected to a tunable chip and can induce a controllable reflection coefficient to the incident signal. The RIS is a nearly passive unit as it only leverages the reflection on its elements and it does not consume any power for transmission. Furthermore, RISs can actively control the propagation channel, by accurately tuning the reflection coefficients of their elements to fit a specific need. The advantage of the RIS arises in creating energy-efficient smart radio environments wherein the wireless channel becomes an optimization variable. In this thesis, we study some of the potential applications and deployments for the RIS in wireless communications. First, the multiple-input multiple-output channel matrix in line-of-sight (LoS) environments turns out to be rank deficient such that spatial multiplexing becomes unattainable. Thus, we exploit the RISs to create additional degrees of freedom by synthesizing a sort of multi-path propagation. Then, we optimize the transmit covariance matrix and the reflection coefficients of the RISs using an alternating optimization algorithm to maximize the achievable rate. Alternatively, we propose different schemes to enhance the composite channel power which would result in an improvement to the achievable rate. Second, we characterize the efficient regions of RIS deployments with single and multiple antenna systems in Rician fading channels. We show that in RIS-aided single antenna systems, near RIS deployments relative to the transmitter or receiver are always better than far deployments. Moreover, we show that in RIS-aided multiple antenna systems, the efficient regions of operations are highly dependent on the propagation environment itself. In LoS environments, both the near and far deployments can result in substantial achievable rate gains. However, as the channel becomes richer with multipath, near deployments gradually become more efficient than the far deployments. Third, we propose the RIS to act as an access point for information transfer by exploiting the polarization control ability of the RIS. In particular, the RIS alternates the polarization state of the reflected waves to perform conventional as well as differential polarization shift keying (PolSK) modulation schemes. In RIS-aided conventional PolSK, two different schemes are proposed. In the first scheme, the receiver corrects for the polarization mismatch loss that occurs in the wireless channel. In the second scheme, the RIS additionally pre-codes the reflected wave to compensate for the polarization mismatch. In RIS-aided differential PolSK, the detection process is independent of the polarization mismatch. Thus, there is no need for a polarization mismatch compensation process by either the receiver or the RIS.