Page 203 - Fiber Fracture

I88 H.U. Kiinzi
It is also known that too large die opening angles favor the formation of these defects.
The drawing force necessary to give a certain cross-section reduction varies strongly
with the die opening angle a!. At small angles the length (fixed reduction) of the conical
part of the die is long. This favors friction and radial compressive stress. In this range of
a! the drawing force decreases when a! increases. At larger angles the excessive plastic
work starts to increase and becomes the dominant contribution in the drawing force. The
conditions prevailing at the minimum in between these two regimes are usually chosen
for drawing wires. Beyond this minimum experimental observations indicate that the
state of stress starts to change. Above a critical angle a! the wire loses contact with the
conical part of the die and the so-called dead zone forms between wire and die. Under
these conditions the wire is just kept back by the smallest opening of the die. At a
slightly larger angle the wire is even rather shaved than thinned. The increase of the
drawing force and the dead zone formation beyond the minimum drawing force indicate
that the zone of negative hydrostatic pressure extends with respect to the compressive
zone and thus creates the conditions for the central burst phenomenon (voids along the
wire axis, see Fig. 3) to become possible.
Since in subsequent traction of the final wire these voids produce stress concentra-
tions, the wire starts to yield prematurely near an internal void until the final rupture
separates the two parts in a cup fracture (Murr and Flores, 1998).
Defect-free Cu wires, as most other ductile wires, show usually necking with a
rough final fracture surface orthogonal to the wire (Fig. 4a,c). With very ductile wires
(recrystallized), necking may go up to the center of the wire before the final failure. In
micro-wires recrystallization may give rise to grain sizes that become (comparable or)
equal to the wire diameter. In this limiting case of a bamboo structure a single grain
having a well-oriented glide system may produce a wedge-shaped neck (Fig. 4b,d).
The nonhomogeneous deformation during drawing not only creates occasional
problems with defects but it gives also rise to microstructural differences which
influence the mechanical properties and the recrystallization behavior. In heavily drawn
wires the grains are usually too small to be observable in an optical microscope (Fig. 5).
In fact, TEM observations of as-drawn wires (prior to annealing) reveal a mixture of
very small microstructural elements (Busch-Lauper, 1988). Fig. 6 shows a longitudinal
TEM image of a 38 pm thick Cu wire (purity 99.99%). Strongly elongated dislocation
cells or subgrains appear in the form of micro-bands that are arranged along the drawing
direction. Some of them have a thickness of only about 0.01 pm whereas others are
clearly thicker (0.1-0.3 pm). The overall dislocation density in these regions is very
large. Local diffraction patterns indicate that either their [ 1001 or their [ 1 1 I] crystal axis
points in the drawing direction. Sometimes also spontaneously recrystallized regions
having extensions of 0.1 to 1 bm or bigger can be observed. They are almost free of
dislocations. These regions appear usually in the core of the wire and indicate that the
stored deformation energy, which acts as driving force for the recrystallization, is larger
here than elsewhere.
Similar observations on as-drawn micro-wires of Cu where made by Murr et al.
(1997) and Murr and Flores (1998). In contrast to our results their samples appear to
show even bigger microstructural differences between the core and the surface near
regions. This may probably be attributed to different drawing conditions and techniques