Page 101 - Wire Bonding in Microelectronics

80 Cha pte r F o u r

are addressed below. While the pull test is valid for wedge bonds, it
is necessary to use a shear test, or, in some cases, a thermal stress test
to evaluate Au ball bonds (on Al pads) adequately. Therefore, the
ball-shear test is extensively described. The theory and applications
of these bond tests are fully understood, and the pull as well as the
ball-shear tests have become standard ASTM [4-1] test methods (see
the appropriate sections below for details). The pull tests and ther-
mal stress tests are currently described in U.S. military specifications
(Std 883-G/H). The shear test is called out in that updated document
later. It is based on the JEDEC commercial standard EIA/JESD22-
B116 [4-2] described in this chapter in Sec. 4.3.11. Various additional
bond evaluation tests are described. There are numerous warnings
about the differences or limitations of the various tests as they might
be applied to fine pitch bond testing, although the main technical
discussion of that subject is in Chap. 9 along with the details of fine
pitch bonding technology.
The nondestructive bond-pull test is currently in use only for some
very special purposes, mostly satellite and space probes, but also occa-
sionally for human-implanted devices. It is included in this edition as
an appendix of the present chapter, rather than dropping it, since its
data and statistics are otherwise unavailable. However, remember that
this test is never used today in any high volume production process,
but all such do use some form of statistical process control.

4.2 The Destructive Bond Pull Test
The wire-bond pull test is the most universally accepted method for con-
trolling the quality of the wire bonding operation. It was introduced to
evaluate the strength of wire bonds to semiconductor devices in the
1960s. Similar tests have been used for welded, soldered, and other wire
connections for many years. The objective of this section is to examine
the variables of the bond pull test, both theoretically and experimentally,
and from them determine the most likely sources of problems and errors
inherent in the test as it is typically performed. Where possible, sugges-
tions for better utilization of the test are included.

4.2.1 Variables of the Bond Pull Test
Numerous papers have been written on the subject of the wire-bond
pull test. Derivations of the equations, standard test methods, and, in
one case, their validation in round-robin tests have been given [4-1, 4-3
to 4-5]. To understand the intricacies of the pull test, it is necessary to
consider the geometrical configuration as well as the several equations
that define the resolution of forces. [Note that the introduction of the
pull angle, ϕ, complicates the calculation, see below Eq. (4-3). When the
hook is pulled straight up, it is better to simplify Eqs. (4-1) and (4-2) by