#homecmos IRC log for Thursday, 2011-10-20

azonenberg_workTesting lift-off patterning of Cr tomorrow, finally confirmed my lab time21:17
azonenberg_workfollowed by several deep wet etches21:17
mrdatacool21:19
azonenberg_workIf this works i'll be doing through-silicon vias/trenches this weekend21:21
azonenberg_workIDK if i can hit a 25:1 aspect ratio (20 um hole 500um deep) but i should be able to hit 5:1 at least21:21
azonenberg_workmeaning a 100um square hole 500um deep21:22
azonenberg_worki'll also be testing wafer thinning21:25
azonenberg_workdo a ~400um deep etch in the center of a die21:26
azonenberg_workthen flip over and do another etch on the front with my features21:27
azonenberg_workHmm, so here's my thought21:34
azonenberg_workTake half a wafer and spin coat in PR21:34
azonenberg_workExpose and develop a contact mask with die-separation lines on it21:34
azonenberg_workEvaporate Cr over the lines21:35
azonenberg_worklift off in acetone21:35
azonenberg_workNow i have a micron of Cr over my dies and nothing in the trenches21:35
azonenberg_workThe next step is to KOH etch the entire wafer until i'm around 100um deep21:35
azonenberg_work(the back side would also be Cr coated)21:36
azonenberg_workSo now i have hardmask over both sides of the wafer except in the kerf areas21:36
azonenberg_workwhich are etched ~100um deep21:36
mrdataneat21:37
azonenberg_workAt this point I spin coat PR and lay down a second contact mask21:37
azonenberg_workThis one is a rectangle in the center of each die21:37
azonenberg_workAlso coat the back side of the wafer in resist so i dont damage the Cr there21:38
azonenberg_workEtch in HCl until the Cr is removed from the center areas21:38
azonenberg_workThen KOH etch until the dies are separated21:38
azonenberg_workAt this point i have a bunch of ~3mm square dies with square outlines21:38
azonenberg_workin the middle of the top is a 400um deep rectangular hole21:39
azonenberg_workmaybe 1-2mm square21:39
azonenberg_workI then flip the dies over21:39
azonenberg_workone at a time, spin coat in PR21:40
azonenberg_workexpose projection mask with comb drive pattern over the center21:40
azonenberg_workdevelop, HCl etch21:40
azonenberg_workthen KOH a third time (leaving the 400um trench on the underside unmasked)21:40
azonenberg_workthe two etches will meet 50um into the die21:40
azonenberg_workleaving me with free-standing features in a 50um silicon plate21:41
azonenberg_workwith a nice thick 500um ring around it for mechanical support21:41
azonenberg_workIf successful, that's all of the bulk micromachining needed for my comb drive21:41
azonenberg_worki just have to grow thermal oxide, evaporate metal, and pattern the M1 mask21:42
mrdatastill, seems like a lot of steps21:43
azonenberg_worklol, it is21:43
mrdatahow many layers is the resulting circuit?21:44
azonenberg_workThis is for MEMS21:44
azonenberg_workit'd be one metal layer over 3D silicon geometry21:44
mrdataok21:44
azonenberg_worki'll do a CAD model later today and post a render21:44
azonenberg_workRight now, going afk as my battery is about dead:p21:45
uneducatedsavanthow do memristors work?21:48
uneducatedsavanthow do they actually change resistance?21:49
mrdataif you wnt them to be useful, you should be able to set their resistance many times21:50
mrdatabut, i dontknow much about memristors exactly21:50
uneducatedsavantyeah it has something to do with filament formation21:53
uneducatedsavantbut i'm not sure how those filaments form21:53
uneducatedsavantor what the filaments are made of21:53
uneducatedsavantis it metal?21:53
uneducatedsavanthmm i wish i knew someone who could help me understand. it's very interesting21:54
azonenbergBack22:39
azonenbergI dont understand memristance myself but there was someone here who wanted to do work on them22:39
azonenbergi forget who it was22:39
--- Fri Oct 21 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!