Silicon metallographic etch

DETAILED DESCRIPTION OF THE INVENTION Referring to FIG.
1 there is shown a graph showing the change in etch rate at two temperatures.
One being within the range of 23.
degree.
to 25.
degree.
C and the other being substantially at 0.
degree.
C.
Line A shows an etch rate for the etchant having a ratio of HNO.
sub.
3 to HF of 100:1.
This shows a steadily declining etch rate for the two temperatures.
Line B shows the etch rate for a HNO.
sub.
3 to HF ratio of 250:1.
Line C shows the etch rate for a ratio of HNO.
sub.
3 to HF of 500:1.
Line D shows the etch rate for an etchant having a ratio of 750:1 of HNO.
sub.
3 to HF.
The temperature of 23.
degree.
C typically represents room temperature which lies with the range of 23.
degree.
to 25.
degree.
C.
Referring to FIG.
2 there is shown a composition dependence of the metallographic etch at room temperature which is well fitted by equation (1): ##EQU1## Three different measurement techniques were used to establish the data shown in FIG.
2 for etch rates calculated during shallow etches of around 600 Angstroms, as well as with deep etches of around 50 microns.
The circles represent data taken from deep etches of greater than 30 microns.
The X represents data taken from shallow etches using a porfilometer to determine the depth of the etch.
The triangles represent unpatterned etches using shallow etches using weight losses to calculate the depth of the etch.
It is evident from the data shown in FIG.
2 that small errors in the preparation of the HNO.
sub.
3 --HF mixtures will alter the etch rate measurably.
The etch depth is given as Rt, where D = Rt.
D is the etch depth in microns, R is the etch rate in microns per minute and t is the etch duration in minutes.
The error in etch depth .
DELTA.
D, due to a fluctuation in etch rate .
DELTA.
R is given by the following equation: ##EQU2## where R.
sub.
0 and D.
sub.
0 are the assumed etch rate and depth.
The term (.
DELTA.
R/R.
sub.
0) may be calculated as a function of ##EQU3## with Equation 1.
Referring to FIG.
3, there can be seen the results of such computations