Data-oriented deterministic networks with pointing error and intelligent reflecting surface

To ensure the security and reliability of emerging data-oriented deterministic networks operating in the Terahertz (THz) band, this paper investigates the secrecy performance of an intelligent reflecting surface (IRS)-assisted wireless communication system. We propose a comprehensive single-reflection signal model that explicitly accounts for THz-specific propagation impairments, including path loss with molecular absorption and stochastic pointing errors denoted by the power efficiency coefficient, while also incorporating practical hardware limitations imposed by discrete phase shift quantization. We derive closed-form expressions for the ergodic secrecy capacity using Gaussian hypergeometric and Meijer-G functions, rigorously analyzing the system’s robustness against eavesdroppers in both colluding and non-colluding scenarios, where the eavesdroppers work in maximal ratio combining and selection combining mode, respectively. Extensive Monte Carlo simulations validate the theoretical analysis, demonstrating that increasing the number of IRS reflecting elements significantly enhances security. Specifically, doubling from 100 to 200 yields a substantial improvement in the secrecy capacity that outweighs the gains from increasing transmit power. Moreover, the low-resolution phase quantization is practically sufficient, achieving secrecy rates that closely approach the ideal continuous phase upper bound.