*Mechanical structure to avoid in-process stiction [#t82d0687]
#ref(sam3.png,left,around,wrap,50%)
Mechanical contact is sometimes unavoidable in designing MEMS devices. In the SEM photo (left) of a MEMS [[VOA>[[Research/MEMS Variable Optical Attenuator with Parallel-Plate Tilt Mirror]]]], a round nail is attached at the edge of the electrostatic torsion plate to minimize the contact area. The VOA chip used in a commercial system is sealed in the dry nitrogen package to keep out moisture. For those MEMS devices used in the normal air environment, moisture is a source of causing failure called in-use stiction, where the mechanical contact would not separate.
#clear
*SAM coating process [#t2cc99eb]
#ref(sam1.png,left,around,wrap,22%)
#ref(sam2.png,left,around,wrap,48%)
From our previous experience in sacrificially releasing silicon micromechanical structures, we discovered that the [[vapor HF (hydrolfluoric acid)>[[Research/Vapor HF Sacrificial Oxide Release with Lab Utensils]]]] greatly helps to avoid process-stiction. Additional two vapor processes described here give further improvement to avoid in-use stiction. The silicon surface after hydrofluoric acid is terminated with hydrogen atoms (-H) and it is soon replaced with hydroxide (-OH). This surface is hydrophilic and is susceptible to moisture. However, furnishing the surface with materials of small surface tension force (surface free energy) reduced the risk of surface stiction. We used ozone gas from the UV-ozonizer to enhance the chemical modification from -H to -OH and appied the vapor of HMDS (hexamethyl disilazane) to make a SAM (self assembled monolayer). The HMDS-oriented SAM surface has small surface free energy and has lower risk of in-use stiction.
#clear
*Quantitative analysis of stiction [#pf82bbe5]
#ref(sam4.png,left,around,wrap,30%)
#ref(sam5.png,left,around,wrap,36%)
As a quantitative measurement of surface stiction force, we used the hysteresis curve of the electrostatic displacement as a function of drive voltage. After the electrostatic pull-in, the contact nail is left in touch with the surface until the drive voltage is lowered to some extent, which is called release voltage in our definition. The release voltage is a good index to show the strength of the contact force at the tip. The lower the release voltage, the stronger the force. The release voltage of the starting material (as HF-released) was high, implying that the surface is less sticky. After processing ozone vapor, the release voltage dropped to a lower value; this is a sign that the surface has become sticky. After applying HMDS vapor, however, the surface stiction became very small, as indicated by the large release voltage. This effect is maintained for more than three hours of mechanical contact repetition. A small degradation was observed, and it would be due to the wear out of the SAM coating after long use. Therefore, the HMDS SAM layer is useful to lower the risk of in-use stiction.
Neverthless, the best design is a design without mechanical contact.
#clear
*Reference [#z9daa714]
+Detail information about vapor HF, ozone and HMDS process is presented in our journal paper published in Japanese:
++泰井祐輔，角嶋邦之，横川隆司，小野志亜之，高橋琢二，諫本圭史，鄭 昌鎬，藤田博之，年吉 洋，「フッ酸，オゾン，ＨＭＤＳを用いたＭＥＭＳデバイスの全気相処理によるスティクション力の低減」 電気学会Ｅ準部門誌, IEEJ Trans. SM, vol. 127, No. 4, 2007, pp. 221-227．
+MSDS of HMDS can be found at, for instance: http://www.jtbaker.com/msds/englishhtml/h2066.htm
![[PukiWiki]](image/XYstage.jpg "[PukiWiki]")
