Page 538 - Cam Design Handbook
P. 538

THB15  9/19/03  8:03 PM  Page 526

          526                      CAM DESIGN HANDBOOK

          major failure mechanism for microscale devices is not fatigue failure but wear between
          contacting surfaces. A detailed study conducted by Sandia National Laboratories on poly-
          crystalline silicon micromechanical transmissions concluded that the major failure mech-
          anism for operating MEMS was the wear of the rubbing surfaces.
             Friction and wear can be most effectively minimized or avoided if elastically flexible
          structures are used. This is probably the reason for the widespread use of simple elastic
          structures  in  microtransducers.  However,  intermittent  surface  contacts  are  unavoidable
          when applications such as the ones described in the previous section involve sophisticated
          mechanical movement. Cams and gears suffer from the problem of friction and wear but
          provide the designer with numerous possibilities. As the tribological issues at the micro-
          and nanoscale are better understood, cams are likely to play an increasing role in MEMS
          devices. Some past and current studies on friction and wear pertinent to the successful
          operation of cams are described next.
             Two  early  studies  on  in  situ  friction  measurement  on  electrostatically  actuated
          micromotors reported friction coefficients in the range of 0.21 to 0.38 for polysilicon on
          silicon nitride and 0.25 to 0.35 for polysilicon on silicon. In the latter study, the micro-
          motors stopped running after 0.75 to 1 million revolutions and this was attributed to wear.
          Another study used a polysilicon wobble micromotor in which the wear between the rotor
          and the central hub changes the gear ratio of the motor. This was used to quantify wear.
          A very detailed three-year study (Sandia 2000) considered the effects of humidity, tem-
          perature, shock, vibration, and storage. The device on which this study focused was called
          a microengine, which as described earlier has many polysilicon parts that rubbed against
          each other. This study outlined the observed failure modes in operating and nonoperating
          conditions and recommended some design rules to avoid such failure. Even though the
          study was based entirely on a particular type of polysilicon microdevice, their findings
          could  be  generalized  to  any  MEMS  device  with  rubbing  surfaces  or  surfaces  in  close
          contact. Wear was observed to be the major failure mechanism. The debris and the asper-
          ities caused by wear lead to momentary and intermittent sticking and eventually perma-
          nent  adhesion  and  seizure.  The  three-body  wear  in  which  debris  gets  caught  between
          the rubbing surfaces was found to be a major contributor to wear. It was also found that
          humidity helps to mitigate wear by acting as a lubricant. Therefore, relative humidity of
          30 percent to 60 percent is recommended for operation at room temperature. Reduction
          of three-body wear by removing the initial debris is recommend but it may not always be
          practical. Minimizing rubbing surfaces by design and minimizing the impact forces at the
          rubbing surfaces by controlled actuation were recommended to reduce wear. All of these
          are relevant when cams are used in MEMS because rubbing surfaces are inevitable in
          cams.
             Several studies have been conducted to study friction at the microscale. In one such
          study (Bhushan, 1999), a silicon nitride (Si 3N 4) probe whose tip had a radius of 50nm (50
              -9
          ¥ 10 m) was scanned over a 1 ¥ 1mm sample area at a scanning speed of 5mm/s in the
          load  range  of  10  to  150nN.  Several  samples  of  single-crystalline,  polycrystalline,  and
          oxide-coated silicon were used. It was found that the coefficient of friction was between
          0.02 and 0.04 for all the samples. This is considerably smaller than the coefficient of fric-
          tion at the macroscale, which was found to be about 0.18 for the same samples but using
          a Si 3 N 4 ball of radius 3mm. Two reasons are cited for this. First, the indentation hardness
          and elastic modulus are higher at the microscale and this reduces wear. Second, there is a
          smaller apparent area of contact, which contributes less to plowing. Even though the coef-
          ficient of friction is an order of magnitude smaller than the macroscale value, it should be
          remembered that friction forces at the microscale are still dominant because of the scaling
          effect mentioned earlier.
   533   534   535   536   537   538   539   540   541   542   543