TechInsights discusses which wafer bonding technology hints at a possible future for stacked dies.
Samsung’s S7 smartphones are offered with several unique builds depending on the country of use. For example, the US models sport the Qualcomm Snapdragon 820 processor and the European versions that we at TechInsights bought contained the Samsung Exynos 8 Octa processor. Both of these processors were fabbed by Samsung using its 14 nm Low Power Plus (LPP) process. Other chips in the phones were dual sourced as well, including their 12 megapixel CMOS image sensors. We knew that Sony was one vendor and Samsung the other, but how was the split done and what technologies were used?
Figure 1 shows the US (SM-G930A) and European (SM-G930F) smartphones with their cover plates removed to expose their circuit boards and camera modules. The layout of the two circuit boards is quite similar but we note small differences in the metal housings used by the two image sensors. We originally thought this to be a marker for the Sony and Samsung variants, but this turned out not to be the case.
Figure 1 Samsung S7 with cover removed SM-G930A US model left, SM-G930F European model right (Source: TechInsights)
Figure 2 shows the camera modules from the US sourced Smartphone (left) and the European model (right). The US model has ‘SONY’ printed on its flex ribbon and our examination of the die confirms it to be the Sony IMX260 12 megapixel backside illuminated CMOS image sensor. Our first sample of the European phone’s module revealed it to be the Samsung S5L2L1 backside CMOS image sensor, but our remaining six European S7 phones housed the Sony image sensors. This was a surprise as we had assumed that Samsung would do a geographic split for the image sensors, much like they did for the Qualcomm Snapdragon and Exynos processors.
Figure 2: CMOS Image Sensor Modules (Source: TechInsights) SM-G930A US model left, SM-G930F European model right
Figures 3 and 4 are die photographs of the Sony IMX260 and Samsung S5k2L1SX 12 Mp backside illuminated (BSI) CMOS image sensors (CIS), respectively that were removed from the two phones. We have removed the organic microlenses and color filters that cover the two dies so that we can get a better view of the pixel array size, and in the case of the Samsung die, the layout of through silicon vias (TSVs) that are used to connect the CIS die to an underlying control ASIC.
The two dies are the same size and their array sizes are essentially the same size as well. No surprise here as the two dies use what appear to be the same optical housings in the European versions of the phones.
Sony had used TSVs in their earlier CMOS image sensors and we had expected the same for the IMX260. But we don’t see them, as they have been replaced by a direct wafer bonding process that we will discuss later.
The Samsung CMOS image sensor has arrays of TSVs along its perimeter and these are used to make the electrical connections to the underlying ASIC.
Figure 3: Sony IMX260 CMOS Image Sensor (Source: TechInsights)
Figure 4: Samsung S5k2L1SX CMOS Image Sensor (Source: TechInsights)
|Table 1: Die measurements
|Sony die size
|Samsung die size
|Sony pixel array size
|Samsung pixel array size