02-a Wet etching - Caltech Micromachining Laboratory

"...-~-: 1804' .' -J. Electrochem. SoC.: SOLID-STATE SCIENCE AND TECHNOLOGY December
,".:: , :." '- --.
~_£~~~~, : ~ -.The etching was done in a Teflon beaker 1 in. in -: : ~
."-:'~;'-.:.- -diameter and 2112 in. deep. Agitation was provided by
~ ::.' an electric stirrer equipped with a polyethylene paddle. ;j
,;;-=.~'~ The dice were etched one at a time in 10 cm3 of :! .6
..,1:':~: '. solution, and duplicate results were obtained by re- ~
~;:-~~' ,~'- peating the en~ire experiment. To quench the reaction ItJ
-:.
1:;'.:
at the proper tIme, a large volume of water was poured
into the beaker. The etching technique was later mod-
~ 12
:I:
'::','!':;
.;~.
ified when the ~igher concentration reagent~ we:e
used. Here the dIce were etched two at a tIme, In ~
"j':. -20 cm3 of solution, while encased in a small perforated ~
". ':. polyethylene basket. The basket and dice were agitated ~
for the required time in the solution, then immersed z:
in a beaker of cold water to quench the reaction. The I-
agitating action consisted of approximately 150 vertical
strokes per minute, with the etching jig rotated
". betv.:een th~ thumb and middle finger as rapidly as 15 20 25
possIble d~Ing the up and do":",n strokes.
TIME (SECONDS)
All etchIng was performed In a constant temperature
(A)
bath regulated at 25.00. =- 0.02.C. Unfortunately,
.: .the heat transmission of polyethylene and Teflon is
.~:
i:';:
so poor that the temperature of the bath could not be
realized in the .etch solution, and even after long
Vi
=
:';.:;: .periods of equilibration the maximum initial e~h ~ 12
:::; ~J temperature was only 24.70.C. .AS; a further Ca:n-
..,.;; ~ lequence of the poor heat transInlSSlOn of the etching
~
z
":":
~.~~
container, the more rapid etches were run under
essentially adiabatic conditions.
~
(.) 8
~~;;,::: It was realized early in the study that in some etch ~
..:0,: .,
':::',
compositions the rate of attack on the die would be so
high t~at if t~e specimen were left in the solution for ~ 4
~:,.,,;.,~. one mlnute, eIther the sample would be etched away x:
'.~:,,::.-i:Ompletely or it would be so thin that it would be. I-
,:_.:' :j:' '. was very made difficult to choose to handle. etch For time this for reason each composition an attempt
90 120 150
:",-:..such that only 4 to 6 mils of material would be re- TIME (SECONDS)
,4:::' moved. The etch rate was then linearly extrapolated (8)
of "', ,' c L - loft d f t. f d .- II .
:~;;.~';
'.:, to tobtain 1 t... the 1 11 min t value. ted b The Fi justification 1 h th for h the F... 1 C hang - ~ rn ~n~.s peas a unc Ion 0 ,.
~-.,: -:~.:, ..', .ex rapo a IO~ IS 1 us. ra y g. w ~re e.c ange posure time: (A) solution composed of 55% HF. 45% HNO3; (I)
::". .In two sam.ple different thickness etch 15 composItIons. plotte;~ as a It fu~ctIon.of 15 qwte tlme.for posslble solution composed of 30% HF . 70% HH03.
~:~-' .that for compositions where the etch rate is very high,
~~,--'.the linear relation between the amount of material -The solution composition field was subdivided into
-:,~:" removed and the time is not valid for the entire a number of evenly spaced points representing dif-
.f ."..:.- minute, since the reaction is known to be strongly ferent compositions. The amounts of each reagent
.~:j.:~:,: ~c exothermic and to proceed under adiabatic conditions. corresponding to each composition were then calc:-u-
~';~,'. Nevertheless, it is believed that by limiting the etch lated, and the solutions were prepared by puttIng
;~J,~;~ ~. time to minimize the effect of changes in experimental together the required amounts of reagents. The hydro-
7;.:.:;; -: conditions, one could obtain numbers by linear extrap- fluoric acid was weighed out directly to =- O.Olg. but
i",:~~ olation which are reasonably valid. The change in the weights of nitric acid and water were converted to
~:~' ".: concentration of the etch system during the experiment volumes and measured out in a burette to: 0.01 cm:J.
,~:..'{:~:.- was also considered an important factor. However,
-.~.;;:.. if one allows only about 0.005 in. of silicon to be etched
Experimental Results
~'::t'~':~' --away. during the operation, t~ would. .correspond to Figures 2 and 3 represent the iso-etch-rate contours
:",:~.'~:";.: reactIng only about ~.4 mequlV. of SIlIcon, and the determined for the normal and high concentratioD
'~;.'~;:'~' ;"";:-;c' '; change action would of concentration be most of 10;0, the which reagents is a due negligible to re- !-o