Raman spectroscopy, high pressure phosphine anneal, and the electrical properties of the DRAM capacitor

Summary form only given. In order to grow good HSG (hemispherical silicon grains), up-front heavy doping of the storage electrode (S-poly) must be avoided (doping blocks grain growth). Therefore, HSG is grown first on the storage poly electrode, then more phosphine is added to the film during the PH₃ anneal process. The effects of film crystallization on the doping level and on the capacitor electrical parameters were studied using Raman spectroscopy. In this paper, we show that at the different steps which the storage poly film goes through, the levels of crystallization and grain sizes change. Specifically, we show that in the PH₃ anneal process, most of the film had crystallized by the time the anneal step proper is reached. In the traditional recipe, the phosphine doping gas (PH₃) is turned on only during the anneal step. Our experiments show that during the initial stages of the PH₃ anneal temperature ramp up (to the anneal step), a broad Raman peak is present, indicating the presence of a large amorphous phase. Consequently, flowing the phosphine gas during the ramp up step while the S-poly film continues to be amorphous, distributes the dopants more evenly in the film and possibly lowers the sheet resistance further. Our results show superior electrical properties such as refresh, dielectric breakdown, and capacitance. Also, the ratio of lots with yields above 100% is higher