Security in wireless networks has traditionally been addressed above the physical layer. With the expected proliferation of applications in 5G, the mm-wave spectrum and new network architectures, traditional methods of data encryption may not be scalable for energy-constrained applications. Thus, there has been a surge of interest in physical layer security that aims to impart confidentiality by exploiting the physics of the wireless communication channel without the need for exchanging secret cryptographic keys [1], [2]. The idea of a secure directional wireless link between a TX/RX pair is to preserve the signal information within a secure cone where the intended receiver is located while scrambling signals everywhere else to prevent eavesdropping. In a phased array, the same temporal digital information is fed to all the TX elements, transmitting the same information to all directions albeit at different power levels. This information can be recovered (especially at the side lobes) with a sensitive enough receiver. Spatial modulation with I, Q radiated out through separate antennas, and modulation of parasitic elements can distort the constellation in other directions [3], [4]. However, this one-to-one (bijective) mapping allows potential decoding by the eavesdropper, particularly using various signal processing and machine-learning-based classification techniques. Time modulation in an antenna array can incorporate such physical layer security through careful mapping of symbols to antennas in a time-modulated fashion.