74
P. Kumar et al.
more parameters including phase morphology and distribution have been proposed
for quantification of microstructure [11, 12].
Figure 4.2 shows that the grain size of the Cu via-filler increases significantly
upon annealing at elevated temperatures [6]; for example, annealing at 450 °C for
30 min leads to an increase in the grain size from 0.84 to 1.26 µm [14]. However,
as shown in Fig. 4.2b, the grain growth during annealing is often inhomogeneous
[13]. Furthermore, impurities in the filler material slow down the grain growth during
annealing [15]. Electroplating at lower current density may lead to fillers with large
grain size; for instance, plating current densities of 1 and 9 A/cm
2 resulted in Cu
fillers with average grain sizes of 2.57 and 0.88 µm, respectively [7]. Annealing has
also been noted to induce twinning in Cu fillers [6, 16].
The micro-texture of TSV fillers is often measured by electron back-scattered
diffraction (EBSD) conducted in SEM (see Fig. 4.2a). However, if the filler grains
are very small, as found in the seed layer, diffraction-scanning transmission electron microscope (D-STEM) can also be used to map the micro-texture [17]. It has
been noted that although annealing increases the grain size, it does not have any
other noticeable effect on the texture of the filler material [6]. The micro-texture of
filler materials appears to depend on the orientation of the seed layer, electroplating
conditions, e.g., current density, bath chemistry, bath temperature, etc., impurity
concentration, etc. [17]. However, to date, the dependence of the crystallographic
texture of the via-filler on various process and geometrical parameters has not been
unambiguously determined.
The general microstructural characterization techniques, such as STEM, energy
dispersive X-ray spectroscopy (EDXS), electron probe microanalysis (EPMA), etc.,
can be used to identify the chemical species in TSVs and map their concentration
profile. Time-of-flight secondary ion mass spectrometry (TOF-SIMS) may also be
used to study the chemical species in TSVs [15]. X-ray diffraction (XRD) can also
be used to identify the phases present in the TSV structures [18].
4.2.2 Measurement of Stress State
Substantial stresses are produced in the TSV assemblies, as a result of processing and
during thermal cycling. The sources and implications are discussed in Sect. 4.3. These
stresses affect the overall reliability of microelectronic packages comprising these
TSV structures, and therefore various approaches have been utilized to understand
the nature of these stresses. Three main techniques that been used to measure the
stress state in TSVs are briefly discussed below.
4.2.2.1 Wafer Curvature Method
If a layered assembly of two or more materials having different coefficient of thermal
expansion (CTE) are joined together and then heated, the assembly bends with a
P. Kumar et al.
more parameters including phase morphology and distribution have been proposed
for quantification of microstructure [11, 12].
Figure 4.2 shows that the grain size of the Cu via-filler increases significantly
upon annealing at elevated temperatures [6]; for example, annealing at 450 °C for
30 min leads to an increase in the grain size from 0.84 to 1.26 µm [14]. However,
as shown in Fig. 4.2b, the grain growth during annealing is often inhomogeneous
[13]. Furthermore, impurities in the filler material slow down the grain growth during
annealing [15]. Electroplating at lower current density may lead to fillers with large
grain size; for instance, plating current densities of 1 and 9 A/cm
2 resulted in Cu
fillers with average grain sizes of 2.57 and 0.88 µm, respectively [7]. Annealing has
also been noted to induce twinning in Cu fillers [6, 16].
The micro-texture of TSV fillers is often measured by electron back-scattered
diffraction (EBSD) conducted in SEM (see Fig. 4.2a). However, if the filler grains
are very small, as found in the seed layer, diffraction-scanning transmission electron microscope (D-STEM) can also be used to map the micro-texture [17]. It has
been noted that although annealing increases the grain size, it does not have any
other noticeable effect on the texture of the filler material [6]. The micro-texture of
filler materials appears to depend on the orientation of the seed layer, electroplating
conditions, e.g., current density, bath chemistry, bath temperature, etc., impurity
concentration, etc. [17]. However, to date, the dependence of the crystallographic
texture of the via-filler on various process and geometrical parameters has not been
unambiguously determined.
The general microstructural characterization techniques, such as STEM, energy
dispersive X-ray spectroscopy (EDXS), electron probe microanalysis (EPMA), etc.,
can be used to identify the chemical species in TSVs and map their concentration
profile. Time-of-flight secondary ion mass spectrometry (TOF-SIMS) may also be
used to study the chemical species in TSVs [15]. X-ray diffraction (XRD) can also
be used to identify the phases present in the TSV structures [18].
4.2.2 Measurement of Stress State
Substantial stresses are produced in the TSV assemblies, as a result of processing and
during thermal cycling. The sources and implications are discussed in Sect. 4.3. These
stresses affect the overall reliability of microelectronic packages comprising these
TSV structures, and therefore various approaches have been utilized to understand
the nature of these stresses. Three main techniques that been used to measure the
stress state in TSVs are briefly discussed below.
4.2.2.1 Wafer Curvature Method
If a layered assembly of two or more materials having different coefficient of thermal
expansion (CTE) are joined together and then heated, the assembly bends with a