This letter investigates the performance of multi-tier heterogeneous cellular networks (HCNs) with macro- and pico-base stations employing non-orthogonal multiple access (NOMA) to improve spectral efficiency and achieve higher data rates for the sixth-generation (6G) wireless networks. Considering the limitations of Gaussian noise models in impulsive interference environments like smart homes and industrial IoT, stochastic geometry is used to compare HCN-NOMA performance under additive white Gaussian noise and impulsive noise. A comparative analysis of OMA and NOMA under these noise conditions with closed form expressions for outage probability (OP) and asymptotic OP, throughput and energy efficiency using the transmit antenna selection is presented, with performance measured by varying base station densities.