Page 51 - ChipScale_Mar-Apr_2021-digital

using a reactive ion etch (RIE) plasma clean is not beneficial for bond
quality. The process is used for a microfluidic system.
In [6], several photostructurable materials for wafer-to-wafer
bonding are compared. The underlying chemistry of the materials is
either epoxy-based, siloxane-based, phenol-based or acrylic-based.
The authors needed materials with a curing temperature below 250°C,
so using BCB or a polyimide was not an option. In this study, a glass
wafer and a silicon wafer were bonded to each other. The result was
analyzed using optical microscopy and scanning electron microscopy
(SEM) imaging. Thermogravimetric analysis was used to assess the
thermal stability of the materials and Fourier transform infrared
(FTIR) spectroscopy was applied to analyze chemical crosslinking of Table 1: Variation of parameters under analysis.
the materials. The bond strength was determined using tensile testing
for all the materials that depend on the bonding temperature and
bonding pressure. Cross-sectional SEM images conclude the paper. In
[7], a similar comparison between BCB and SU-8 was done, and in [8],
BCB was used for building a microfluidic device.
In the application analyzed in this article, the silicon chips are
placed face down onto the adhesive layer. This process step is done
using a standard die bonding machine. Both the substrate and the chip
can be heated during this process step. A force needs to be applied
on the chip in order to achieve sufficient bonding strength. The shape
of the adhesive material at the edges is expected to be influenced by
the process parameters. Furthermore, the resulting strength of the
bond and the resulting shape of the adhesive with respect to their
dependence on the process parameters have not yet been referenced in
literature.
One of the main differences between the application under analysis
and the microfluidic application is the resulting pressure field within Figure 3: Shear testing results before curing with respect to
the adhesive. Pressurizing the sandwich of two wafers and the temperature and bonding force variation.
adhesive layer results in a homogeneous distribution, whereas the
single placement of the silicon die results in a local deformation and
a local temperature change only. Therefore, the process parameters
referenced in the given literature (like temperature and bonding force)
cannot be used for this application and are analyzed in detail in the
following sections.

Methodology
The material under analysis is expected to show the desired bonding
behavior after the lithographical structuring process steps. Therefore,
bonding temperatures above 150°C are needed. This process window
will be used to place the silicon chip into the adhesive material.
Furthermore, the reflowing of the material is expected to assist the
covering of the edges. The build-up of wafer, adhesive layer and
silicon dies can be cured accordingly after bonding. Because of the Figure 4: Shear testing results after curing with respect to
underlying chemistry, the material is expected to show high chemical temperature and bonding force variation.
and mechanical durability.
Figure 2 shows the applied process chain. First, the processed
wafer is heated to a standby temperature of 50°C. This ensures a
homogeneous activation of the adhesive material, but is still low
enough for the material not to show any curing or outgassing effects.
Second, the silicon die is picked up from a waffle pack and heated
up to the required chip temperature, which will be analyzed later.
A delay is implemented to ensure sufficient heat flow from the tool
tip to the silicon component. Next, the component is placed onto the
adhesive and pressure is applied. The influence of the applied pressure
will be analyzed in detail in the results section.
The bond quality needs to be tested, therefore, standard die
shear testing is used to quantify the bond strength. Standard light
microscopy is used to analyze the resulting fracture after die shear Figure 5: Direct comparison of shear values before and after curing at
a 220°C tool temperature.

49
Chip Scale Review March • April • 2021 [ChipScaleReview.com] 49