Page 37 - Wire Bonding in Microelectronics
P. 37

16    Cha pte r  T w o



                         0.400               T

                       Distance from tool tip (in)  0.240
                         0.320




                         0.160


                         0.080        3
                                        1
                            0
                               60    40   20   0
                                     Vibration amplitude (µ in)
              FIGURE 2-3  Wedge bonding–tool vibration modes for 60 kHz excitation. The
              vibration amplitude (peak to peak) was measured with a laser interferometer
              (as shown in Figs. 2-3 to 2-5). The dark portion with the T in it represents the
              position and the vibration amplitude of the transducer horn. The transducer
              frequency was 60 kHz. Modal patterns for a short, 30° (angle of wire
              entering the tool) WC too (1) shown unloaded and during the actual bonding
              of 25-µm Al, 1% Si wire. The bonding force was 25 g: curve 1, (×), unloaded;
              curve 3, (o), near end of an actual bonding cycle. The symmetry of the
              vibration is indicated by the dotted curves on the back side of the tool; these
              mirror-image amplitude curves are omitted for subsequent fi gures for clarity,
              showing only the front portion of the vibration. The large dark area (T)
              represents the position and vibration amplitude of the transducer horn.
              (Note: 10 mil = 0.254 mm) [2-9 to 2-11].

              and during bonding, are shown in Figs. 2-3 to 2-5, in curves 1 to 3.
              These absolute vibration amplitude measurements were made with a
              laser interferometer. The US power supply used to drive the trans-
              ducer for these studies maintained a fixed frequency of ~60 kHz and
              constant amplitude during bonding. ∗  In these figures, “T” indicates
              the end of the transducer.
                 The loaded vibration nodes are seldom as sharp (usually rounded)
              or as complete as the unloaded ones. The nodal position during bond-
              ing is dependent upon the mechanical loading of the tool tip, which
              is primarily a function of the bond deformation (~1.6 wire diameters).
              A smaller deformation, such as 1.2, may result in a nodal rise of only
              0.50 to 0.64 mm (20–25 mils) from the unloaded position. This tool
              showed no significant change in tip amplitude during bonding, but
              such changes would be expected if the bonding force and/or the bond


              ∗ Modern U.S. power supplies are phase-locked and can change the frequency a few
              hundred hertz or so during bonding to keep the resonant system optimally tuned.
              This should not significantly alter the modal patterns or the tool-loading effects.
   32   33   34   35   36   37   38   39   40   41   42