Breaking reciprocity lies in the heart of modern electromagnetic devices such as circulators, isolators, filters, and antennas. Generally, breaking reciprocity can be realized by biasing the device with a physical quantity that is odd symmetric under time-reversal. Conventionally, such response has been achieved by magnetically biasing ferromagnetic compounds and garnets. In the past decade, magnetless approaches based on nonlinear responses, spatiotemporal modulation, drifting electrons, and opto-mechanical effects have been explored across the electromagnetic spectrum. Recently, circular dichroism has been demonstrated to be an efficient tool to break reciprocity in 2D materials due to the optically-driven non-degenerate valleys.In this thesis, I propose a novel method to break reciprocity in 2D materials based on the use of circularly polarized light and strain engineering. To this purpose, I review and calculate the optical conductivity of graphene under uniform strain and discussed the impact of non-uniform strain on graphene. The results show a close resemblance between pseudomagnetic field and the effect of non-uniform strain. The utilization of non-uniform strain gives rise to energy level quantization in graphene and thus provides easily addressable optical transition between discrete energy levels using optical pump. I review the formalism and algorithm of a rigorous theoretical framework able to deal with Bloch equations regarding the population of different energy levels. The incorporation of some scattering mechanisms also closely reproduces what happens to the carrier population in picosecond time scale after the presence of an optical pump. The results show the population imbalance in the two non-degenerate valleys in graphene which indicates the broken reciprocity. This approach of breaking reciprocity unleashes large non-reciprocal response in graphene using magnetless implementations. In a related context, the field of plasmonics has opened new possibilities to control and manipulate light beyond the diffraction limit and has enabled countless applications in areas such as sensing, spectroscopy, and healthcare. The emerge of ultrathin metasurfaces and 2D materials have provided new knobs to excite, process, and route SPPs, while also enabling unexpected possibilities to manipulate and enhance nonreciprocal responses. Unfortunately, the design of quasi-optimal nonreciprocal metasurfaces is usually quite challenging and require significant computational resources. This thesis unveils the fundamental limits of linear and nonreciprocal plasmonic metasurfaces in terms of isolation and loss. The proposed bounds are related to surface waves and only depend on the nonreciprocal material employed within the metasurface, thus being independent of geometrical considerations and the presence of other materials. We apply these fundamental limits to explore two different platforms, namely drift-biased and magnetically-biased graphene metasurfaces. For each platform, we first analytically derive the upper bounds in terms of graphene conductivity. Then, we explore devices proposed in the literature and benchmark their response against their upper bounds. Results highlight that drift-biased hyperbolic metasurfaces exhibit outstanding performance in the mid-infrared region, whereas magnetically-biased devices are better suited for the low terahertz band. More broadly, our bounds allow to quickly assess the performance of nonreciprocal plasmonic metasurfaces with respect to their fundamental limit, thus streamlining the device design process and preventing that significant efforts are dedicated to marginal performance improvements. The proposed bounds pave the way toward the development of quasi-optimal nonreciprocal metasurfaces, with important applications in sensing, imaging, communications, and nonlinear optics, among many others.