Page 44 - Wire Bonding in Microelectronics
P. 44

Ultrasonic Bonding Systems and Technologies       23


              the pad with the normal clamping weight of 25 g, but with no ultra-
              sonic energy applied. Figure 2-10(B) shows the lift-off pattern made
              by applying ultrasonic energy for a 4 ms period. The wire-to-pad
              microwelds have formed at points near the perimeter. The lift-off pat-
              tern in Fig. 2-10(C) resulted from a 7 ms bonding time. The welded
              area has spread part of the way around the perimeter. Figure 2-10(D)
              shows the bond formation at 10 ms. The welding has increased con-
              siderably, but is still primarily confined to the perimeter. At longer
              bond times, the wire could not be lifted up without tearing the pad or
              breaking the wire. Examination of many such patterns shows that
              weld formation begins around the perimeter, but that no two time-
              equivalent patterns are exactly the same. The amount and location of
              the welding around the perimeter may show considerable variation.
              However, the examples given in Fig. 2-10 were chosen to be typical of
              those observed for each indicated bonding time and power setting. A
              further verification that ultrasonic wedge welding progresses around
              the perimeter is shown in Fig. 2-11. These are photographs taken
              of disrupted metallization, as seen through the back side of an Al-
              metallized fused quartz substrate [2-11 in SP]. In each photograph,
              the bond parameters were held constant except for the power. Note
              that the third bond (c), made at the highest US power, has cracked the
              quartz. This is an example of cratering caused by excessive US energy
              (see Chap. 8 for a discussion on cratering). Studies of evolution of the
              bond interface on Au/Ni/Cu bond pads, obtained by etching the Al,
              1% Si wedge bonds off, have revealed growth patterns similar to





















              FIGURE 2-11  “Through the underside of a bond pad.” The disruption of the
              bond pad observed from the underside of a thin (~0.2 µm) evaporated Al
              pad, which was deposited on a clear fused quartz substrate. These three
              patterns were made by increasing the ultrasonic power for each Al wedge
              bond, while holding force and time constant. Power-supply dial settings
              were, from left to right, (a) 4.5; (b) 5.5; (c) 9.5. The third bond, made at
              the highest US power, has cracked the quartz, an example of cratering
              discussed in Chap. 8 [2-11].
   39   40   41   42   43   44   45   46   47   48   49