As an emerging electronic device,
a memristor facilitates the realization
of neuromorphic computing systems. However, high-performance neuromorphic
computing systems require not only robust analog memristors but also
digital memristors with complementary functions. Here, a digital–analog
integrated memristor based on Al/ZnO NPs/CuO NWs/Cu is proposed and
demonstrated. First, the electrical properties of the CuO NWs based
device are investigated, which exhibit a write-once-read-many-times
(WORM) performance. Then, typical bipolar resistive switching behaviors
are realized by spin-coating the ZnO NPs to form an Al/ZnO NPs/CuO
NWs/Cu based memristor. Through tuning of the applied voltage, an
abrupt-to-gradient transition of conductance is achieved in
situ. Several analog behaviors, such as long-term potentiation/depression
(LTP/LTD), and paired-pulse facilitation/depression (PPF/PPD) are
effectively emulated by applying a series of specified electrical
measurements to the memristor, which proves the achievement of a digital–analog
integrated memristor. Furthermore, on the basis of the LTD and LTP
behaviors of the memristor, neural network simulations for handwritten
recognition are implemented. The results reveal high recognition accuracies
of up to 93%, which are close to the performances for ideal memristors.