Page 10 - Chip Scale Review_Jul Aug_2022-digital

Detecting the temperature in real time and with accuracy
are essential first steps to controlling it. ATT-Systems’ low-
thermal-resistance (LTR) wafer chucks include a MultiSense
option, which places multiple sensors across the chuck to
accurately measure the temperature. The number of sensors
used varies depending on the application, and is mainly
driven by the contact area size. The smaller the contact
Figure 2: Number of sensors scaled to contact area size, ordered from low- to area, the higher the quantity and array density of sensors.
high-sensor array density. While full-wafer contact applications such as DRAM testing
only require one sensor, multi-touchdown devices like
automotive microcontrollers or graphics processing units
(GPUs) can use up to 21 sensors (Figure 2).

Using multiple wafer chuck sensors vs. a single sensor
Let’s look at an example of testing with and without our
multiple sensor solution. Figure 3 shows a wafer chuck test
setup of a 21 x 22mm contact area with a liquid cooler. In
this test, 200W was applied for two minutes with a desired
temperature of 70°C. The solution with multiple sensors
(Figure 3a) shows that the temperature initially jumped
to 74.1°C when the power was switched on. After 28s, the
temperature was brought back down to +/-1°C of the desired
number. At the 33s mark, the 200W of power was completely
dissipated. The test using only a single sensor (Figure 3b)
presents different results. When the power was turned on, the
temperature rose past 96°C in the two-minute time frame. The
system did not detect a thermal change or implement an active
temperature control solution.
With the flexibility of additional sensors, this new LTR wafer
chuck technology can be used for a wide variety of high-power
applications, including full-wafer contact DRAM testing,
automotive microcontrollers, and 5G devices. These active
thermal control solutions have proven to be valuable in a series
of case studies.
Case studies
The first case study is a full-wafer contact example
that is used for high-bandwidth memory (HBM) testing.
The setup includes a wafer chuck and 300mm heating
Figure 3: a) Multiple sensors vs. b) single-sensor temperature response and plate to simulate the 2,000W wafer test conditions. With
control results. a 300 x 300mm contact area, the power will be uniformly
distributed throughout the entire chuck area, prompting
the need for only one sensor. For this test, the three
temperature points were -40°C, 25°C and 125°C. Figure
4 shows that once the 2,000W of heat was applied, it
dissipated within seconds and was brought back down to
within +/-2°C of the desired temperature.
The automotive microcontroller case study used a 100
x 100mm contact area with medium parallelism. For these
devices, heat transfer can be up to 1000W, with a moderate
2
power density of 10W/cm . Nine sensors are recommended for
this contact area size to cover all areas of the chuck. During
the case study, 400W was applied across the wafer at three
temperature points: 0°C, 85°C and 105°C. As seen in Figure
5, the correct sensor detected the temperature increase and the
chuck system was able to dissipate the 400W within a +/-2°C
Figure 4: Full-wafer contact test (2000W) with -40°C test temperature.

8 8 Chip Scale Review July • August • 2022 [ChipScaleReview.com]