Study Of Low-K Dielectric Material And Local Storage Performance In Nano-scaled SONOS Devices

Polysilicon-oxide-nitride-oxide-silicon (SONOS) devices have been extensively investigated thanks to its multi-bits and multilevel storage technology. With size scaling down of SONOS devices, research on device reliability and metal interconnection technology become more and more important. Based on the two aspects, a series of studies have been carried out in this paper.For metal interconnection system, as the application of low-k dielectric material to reduce interconnection capacitance, study on low-K dielectric material has been an important aspect. In this paper, the deposition process and reliability of low-K SiOx dielectric in metal interconnection system was discussed, which provided experimental and theoretical basis for low-K dielectric material’s research. On the other hand, SONOS memory uses channel hot electron injection for programming and band-to-band tunneling hot-hole injection for erasing. While the injected electrons has a relatively wider distribution than injected holes due to spatial charge accumulation in the nitride layer after successive P/E cycling. This charge accumulation will reduce cycling endurance and accelerate device degradation. In this paper, we also proposed a new method to detecting local electrons’ distribution in storage layer, which promoted the further study of nano-scaled devices’ degradation. The main results are as follows:1. In metal interconnection system, low-K dielectric material and dielectric reliability study was discussed first. Then on basis of SiO2 dielectric, deposition of low-K SiOx dielectric was achieved by increase the dosage of reactant SiH4. While this dielectric’s lifetime and strength performance were both unqualified.2. Using low voltage CHEI programming method, the single electron’s injection in nano-scaled SONOS devices was achieved. By combination of theoretical calculation, TCAD simulation and experiments, single electron’s effect on VTH and lateral location was detected. Then on basis of single electron effect and data fitting, the distribution of trapped electrons induced by device degradation was investigated.