For the self-energy recycling (SER) untrusted relay network, a two-phase destination-based jamming (DBJ) protocol is proposed, in which the destination transmits the jamming signal to reduce the received SNR of the untrusted relay node in the first phase and the relay node operated in full-duplex (FD) mode for simultaneous energy transfer and information relaying in the second phase. In addition, the loopback interference (LI) generated by FD relay node can be reused as part of energy. To satisfy some practical application scenarios, we consider the imperfect channel estimation error at the destination. Considering the total power constraint and the quality of service (QoS) requirement at the destination, the secrecy rate maximization problem is formulated by optimizing the transmission power of source and destination. However, the formulated problem is non-convex and difficult to solve directly. To cope with this difficulty, an iterative power allocation algorithm is proposed. The key idea of the proposed algorithm first integrates the non-convex constraint into the objective function by the exact penalty method. Then the difference of convex functions (DC) programming can be utilized to convert the non-convex objective function into the approximate convex function and the sub-optimal transmission power of source and destination can be obtained by the traditional convex programming. Simulation results show the superiority security performance of the proposed scheme with the traditional schemes.INDEX TERMS Physical layer security, self-energy recycling, power allocation, imperfect channel estimation error, secrecy rate.