| At present, the intelligent society is experiencing the increasing demand of portable devices, therefore, the demand for non-volatile memory is growing every day. With the rapid development of the electronics industry, device processing dimension is shrinking, the storage capacity is increasing, the floating-gate memory is facing serious challenges. To this end, the priority is to develop a new memory with low operating voltage, low power consumption and high stability to replace the traditional floating-gate memory. Currently, the concentrated researches are being focused on two types of memories, one is charge trapping memory(CTM) improved based on floating-gate memory, the other is a non-volatile memory with a brand-new structure. The design ideas of charge trapping memory inherited the traditional floating-gate memory. The peripheral circuits, storage array and processing are basically similar to the floating-gate memory, which means that it can be compatible with conventional CMOS semiconductor technology. With the advantages of low operating voltage, low power consumption and excellent resistance to fatigue, CTM is attracting much attention and shows a good prospect.Charge trapping memory consists of blocking layer, trapping layer and tunneling layer. It uses defects inside the trapping layer to store charges. TAHOS (TaN/Al2O3/ HfO2/SiO2/Si) is a typical structure being studied presently. On the basis of this structure, memory performance could be improved by studying properties of materials. Due to the fact that the first-principle can avoid the shortcomings of long preparation period and high platform requirement, the first-principle calculation was used to analyze the physical links between the microscopic parameters of materials and the macroscopic properties of memory.Firstly, we study two defect systems by first-principles calculations, one is four-coordinated oxygen vacancy (V04) defects system, and the other is doping Al substitutional impurity to HfO2 trapping layer with an intrinsic defect of Vo4 to form the co-doped composited defect system. The results show that the incorporation of impurity Al can effectively improve the data retention and endurance of device. The calculated oxygen vacancy formation energy shows it is easier to form the co-doped composited defect system. The program/erase(P/E) operation was simulated by the addition and removal of charge. The calculated results show the co-doped composited defect system can trap both electrons and holes, and the capture capacity of carriers is increasing, the energy of erase the carrier becomes large, so, it is an important role to hold carriers. Bader Charge Analysis showed that the co-doped composited defect system is more conducive to maintain the data retention, because the difference number of holes and electrons can be trapped is very small in this system. The analyses of Density of states and band structure demonstrate that the co-doped composited defect system has a strong effect on the trapping energy of holes and holes located on the defect level are excited into the valence band required larger energy. It indicates that this system has better data retention characteristics. And structural analysis and endurance calculation reveal that the doping impurity Al can improve the endurance of the device.Secondly, we research the effect of the distance variation between doping impurity Al and intrinsic defect Vo3 (three-coordinated oxygen vacancy) on data retention characteristics. The results indicate that the system presents the strongest data retention when the distance of the defects is 2.107A. A study for the incorporation of A1 concentration problems and the results show that device has fastest write speed and best data retention characteristics when the mole ratio of Hf/A1 is 1:1. The calculation results indicate that the number of quantum states is most when the distance of the defects is 2.107A. The quantum states of defects level is largest and the number of carriers can be trapped are most for this system, so this system has strongest data retention of carrier. Population number and bond length analysis shows the smallest population number and the longest Al-0 bond length when the distance of the defects is 2.107A. In this case, A1-0 bond is more prone to fracture, which will create more carriers in the system. By studying the bond length changes of the three systems after writing a hole, we obtain a result that the change of Al-0 bond length is the smallest after writing a hole when the distance of the defects is 2.107A. At the same time, in order to study the relation of P/E speed and doping concentration to obtain the optimal doping ratio, five kinds of concentration structures were simulated by calculating the tapping energy and band offset. The results show that the easiest electron tunneling into the trapping layer and the shortened tunneling time of carrier were obtained with the doping ratio of Hf/Al=1:1, realizing the fastest programed speed in the device. Different concentrations of A1 impurities were doped into the system with intrinsic defects. By calculating the formation energy of oxygen vacancy and the charge trapping energy, the system was found to show a good data retention with the doping ratio of Hf/Al=1:1.Finally, the HfO2/SiO2 interface composed by trapping and tunneling layers was explored in this work. The results show that the interface gap states are mainly generated by the unsaturated bond Hf atoms and Si atoms near interface and the incorporation of interstitial oxygen atom at the interface can reduce gap states. The method of establishing interface model and the reason of producing interface gap states were mainly studied. With the calculations of DOS, PDOS and bader charge, the results show that the interface gap states are mainly generated by the unsaturated bond Hf atoms and Si atoms near interface and the incorporation of interstitial oxygen atom at the interface can reduce gap states. It can improve the data retention characteristics of the device.
|
PDF Full Text Request