Nitric Oxide Rapid Thermal Nitridation of Thin Gate Oxides
- PDF / 412,142 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 87 Downloads / 224 Views
current stress [9]. The N 20 process is limited in two areas. First, the oxide growth rate is significantly lower than for 02, resulting in a higher thermal-budget process [10]. Second, the incorporation of nitrogen is limited to moderately low levels, approximately 1.7 atomic percent for a 40 A film [11]. In contrast, this paper will show that an NO-nitridation process can incorporate significant levels of nitrogen with moderate thermal exposure, and the bulk of the film thickness can be grown in 02 with high growth rate. EXPERIMENT Experimental Setup This work was performed in an Applied Materials RTP Centura system. The system incorporates dual vacuum load-locks, a controlled-atmosphere transfer chamber, and multiple isolated process chambers. This architecture provides full isolation, even in continuous
production operation, from the contamination of atmospheric moisture and oxygen. This isolation also prevents exposure to toxic gases which may be used in the process chambers. The RTP chamber provides precise process control through a multi-zone lamp source controlled by multiple, emi ssi vity-in dependent temperature measurement probes. Wafer rotation at 90 rpm enhances uniformity as well as controllability. Chamber pressure can be regulated from I to 760 Torr during the process. This process control results in highly uniform and repeatable oxide growth. Typical oxide thickness uniformity is between 0.5 % and 0.75 %. An integrated oxidation-nitridation process sequence was used in this work. First, an initial rapid thermal oxidation in 02 is used to grow the bulk of the thickness. Second, with the temperature stabilized at 600 'C, the 02 is pumped out and the chamber is backfilled with NO. Third, the oxide is nitrided with a rapid thermal anneal. Finally, the NO is pumped out and the chamber is backfilled with N2 before the wafer is transferred out. The thickness uniformity for this integrated process is unchanged from the uniformity of 02 oxidation, as seen in Figure 1.
0
Process:
1000 'C, 37 sec., 02 1000 'C, 60 sec., NO
Mean:
43.12 A
Std. Dev.:
0.28 A (0.65 %)
Contour:
0.09 A
Edge:
6 mm exclusion
Figure 1. Thickness uniformity for NO-nitrided oxide Nitrogen Incorporation Initial experimentation focused on characterizing nitrogen incorporation as well as oxide thickness growth as a function of temperature, time, NO flow rate, and initial oxide thickness. The experiment was a linear design using ECHIPTM software and is summarized in Table I. 382
Table I. Nitrogen incorporation experiment design Variable Temperature Time Initial Thickness NO Flow
Low 900 °C 15 sec. 25 A 1 Sim
Mid-Range 1000 °C 67 sec. 37 A 5 slm
High 1100 °C 120 sec. 55 A 9 slm
The variables in order of their effect on nitrogen concentration were: temperature, time, initial oxide thickness, NO flow rate. NO flow rate had a statistically insignificant effect, therefore 1 slm was used for subsequent processing. Using ECHIPTM, the nitrogen concentration and oxide growth data were fitted to a linear model. Because no replicates
Data Loading...
