Page 340 - Sami Franssila Introduction to Microfabrication

Tools for Hot Processes 319

Table 31.2 Comparison of furnace and RTP processes 5. What temperature error does emissivity change from
0.71 to 0.87 cause in rapid-thermal oxidation?
Furnace Rapid-thermal processing 6. What power rating does an RTP system for 300 mm
wafers need if its maximum operating temperature
Batch Single wafer ◦
Hot wall Cold wall is 1200 C?
Long time Short time 7. Anneal time and junction depth are connected
Small dT /dt Large dT /dt as follows: x j = k × (Dt) 1/3 . If junction depth is
Indirect Direct temperature measurement ca. 100 nm in 0.25 µm technology and the corre-
sponding anneal time is 10 s, what is the anneal
time for 0.1 µm technology? What is the junction
31.4 EXERCISES depth?
8S. Typical furnace anneal activation is 950 C/30 min,
◦
1. What should the oxygen flow be in a horizontal
but in RTA, a much higher temperature and a much
batch furnace to make sure that oxidation is not
shorter time are used. Compare junction depths
mass transfer–limited? Write out and justify the
that can be made by RTA and FA. Use implant
assumptions you need in your solution. 15 −2
conditions of 20 keV boron, 10 cm into a
2. If reproducibility and other uncertainties in a batch- 15 −3
phosphorous-doped wafer with 10 cm .
loading furnace limit the shortest practical oxi-
dation time to 15 min, what is the thinnest gate
◦
◦
oxide that can be grown at 1000 C, at 950 C, REFERENCES AND RELATED READINGS
◦
at 900 C and 850 C? What are the corresponding Bensahel, D. et al: Front-end, single wafer diffusion processing
◦
CMOS linewidths? for advanced 300 mm fabrication line, Microelectron. Eng.,
3. How rapid is RTP? Calculate how long the heat 56 (2001), 49.
pulses must be to result in thermal equilibrium Bratschun, A.: The application of rapid thermal processing
of the whole silicon wafer. Thermal diffusivity technology to the manufacture of integrated circuits – an
2
in silicon is 0.80 cm /s at room temperature and overview, J. Electron. Mater., 28(12) (1999), 1328 (special
2
0.1 cm /s at 1400 C. issue on RTP).
◦
4S. Rapid-thermal oxidation (RTO) data is given in the Deaton, R. & Massoud, Z.: Manufacturability of rapid-thermal
table below. How does RTO compare with furnace oxidation of silicon: oxide thickness, oxide thickness
variation and system dependency, IEEE TSM, 5 (1992),
oxidation? Data from Deaton, R. & Massoud, Z.: 347.
Manufacturability of rapid-thermal oxidation of Endoh, T. et al: Influence of silicon wafer loading ambient
silicon: oxide thickness, oxide thickness variation on chemical composition and thickness uniformity of sub-
and system dependency, IEEE TSM, 5 (1992), 347. 5 nm thickness oxides, Jpn. J. Appl. Phys., 40 (2001),
7023.
Fair, R.B., Conventional and rapid thermal processes, in
Constant time 30 s Constant temperature C.Y. Chang & S.M. Sze (eds.): ULSI Technology, McGraw-
1050 C Hill, 1996.
◦
Roozeboom, F. & Parekh, N. Rapid thermal processing sys-
Temp Thickness Time Thickness
tems: a review with emphasis on temperature control, J. Vac.
◦
950 C 44 ˚ A 30 s 75 ˚ A Sci. Technol., B, 8(6) (1990), 1249.
◦
1050 C 75 ˚ A 150 s 158 ˚ A Saga, K. et al: Influence of silicon-wafer loading ambients in
1150 C 145 ˚ A 270 s 240 ˚ A an oxidation furnace on the gate oxide degradation due to
◦
organic contamination, Appl. Phys. Lett., 71 (1997), 3670.