A silica etch process conducted using a CHF3/CF4 inductively coupled plasma
is characterized. This was accomplished by employing a statistical experim
ental design in conjunction with neural network process models. As a mask l
ayer, patterned AlSi (1%) metal was used. Parameters varied in the design i
ncludes source power, bias power, and gas ratio. Besides those conventional
etch responses including etch rate, selectivity, and profile, sidewall rou
ghness of the etched pattern is first modeled. Etch rate and sidewall rough
ness were found to be predominantly influenced by source power with a trade
-off between them. Bias power significantly affected selectivity while cont
rolling a trade-off against etch rate. A decrease in profile angle with inc
reasing bias power was attributed to AlSi (1%) film expansion induced by io
n bombardment effects. As gas ratio was varied, profile angle, remained alm
ost constant due to nearly same chemical reaction of its plasma on the Sili
ca surface. The roughness was little affected by bias power at its low leve
ls, thereby providing an increased degree of freedom to optimizing the proc
ess. The gas ratio exerted no noticeable impact on the etch responses. (C)
1999 American Vacuum Society. [S0734-2101(99)08705-9].