Search

Ferroelectric (FE) hafnium oxide is a promising candidate for memory applications. In this paper, endurance, imprint, and retention tests are carried out on FE aluminum-doped hafnium oxide thin films with different electrodes: 1) metal–insulator–metal (MIM) and 2) silicon–insulator–silicon(SIS). MIM devices exhibit higher endurance than SIS devices. However, SIS devices show superior imprint and retention, with an extrapolated retention time-to-failure larger than 10 years at 85 °C. These results are very promising to integrate FE-HfO2 in 3-D structures, either for dynamic random access memory or for NAND applications.

Phase-change memory (PCM) relies not only on phase transitions between the two structures of the chalcogenide materials but also on electronic switching in the amorphous material between off and on conductive states and vice versa. Transient effects associated to both transitions, like the delay time for threshold switching, the switching time, and the recovery time for the on to off transition, play a fundamental role in the operation of the PCM device, namely program, erase, and readout processes. This paper focuses on the transient modeling and simulation of the recovery effect in PCM cells. After providing physical insight into the recovery process, we present a numerical model able to account for the transient recovery of threshold voltage VT and of resistance R after the on -off transition. The impact of recovery effect on program-verify loops used for accurate positioning of R is discussed. The combination of switching and recovery effects is shown to give rise to relaxation oscillations under suitable bias conditions, and the transient model for switching and recovery is applied to describe and predict oscillation behavior of PCM.

Threshold switching, i.e., the electrical transition from high to low resistivity in amorphous semiconductors, plays a major role in the program/erase processes of phase-change memory (PCM) devices. Understanding and designing materials and cell structures for optimum memory performance require that accurate models for threshold switching be developed, in both the steady-state and the transient regime. This work presents an experimental and modeling study of threshold switching in amorphous chalcogenide materials for PCM applications. The delay time for the onset of threshold switching from the application of an electrical pulse is explained based on 1/f fluctuations of the subthreshold current. A Monte Carlo model is developed for calculating the statistical distribution of delay times. The model can account for 1) the voltage dependence of the average delay time and 2) the statistical spread of the delay time for both rectangular and triangular voltage waveforms. The delay model is used for a reliability assessment of read disturb in PCM devices.

The phase-change memory (PCM) relies on the electrical properties of the chalcogenide materials to represent the stored bit of information. As a result, data stability depends on structural relaxation (SR) in the amorphous chalcogenide phase, which results in a temperature-accelerated time evolution of the electrical properties of the active material. Here, we address the time, temperature, and bias dependence of SR effects on the amorphous Ge2Sb2Te5 (GST) material used in PCM cells. Electrical measurements for increasing annealing time and temperature indicate that SR can be described by a defect annihilation process in the amorphous chalcogenide material. Finally, the stability of chalcogenide resistance as a function of the read conditions is discussed, for the purpose of reducing the impact of SR on the reliability of PCM devices.

Phase-change memory (PCM) cells are affected by structural relaxation (SR), which is the atomic-scale rearrangement of the amorphous phase of the chalcogenide material. Since SR affects the stability of electrical parameters of the PCM cell, such as resistance and threshold voltage, physics-based models for SR are necessary to analyze and predict the device reliability. This paper presents a physical model for SR in amorphous chalcogenide materials, linking the defect annihilation dynamics to the conduction behavior of the cell, hence to the electrical characteristics of the PCM cell. The model is able to predict the PCM cell characteristics as a function of annealing time, temperature, and read conditions.