Moore’s Law predicted that the number of transistors within the same area on an integrated circuit would double every 18 months, and chip performance would continue to improve. After nearly 60 years of development in the semiconductor industry, however, this law is reaching its limit.
Major manufacturers have been searching for new ways to continue to improve the overall performance of chips when transistors can no longer be made smaller. One approach is to stack semiconductor circuits in layers. The key to this technology lies in using “heterogeneous integration” of “packaging” to continue Moore’s Law. The biggest advantage of advanced packaging is that it greatly shortens the distance between the metal wires of different chips, thus greatly improving transmission speed as well as reducing the power lost during the transmission process.
Various wafer fabs and packaging and testing plants are currently developing their own advanced packaging technologies. In terms of technology and capital expenditure, Intel and TSMC are the most active. Their combined capital expenditure accounts for 55% of that of the industry, and their technology is the most advanced.
|
Demand for AI Chips is Driving Innovation in the Advanced Packaging and Testing Industries |
Using TSMC as an example, advanced packaging technologies can be divided into 2D InFO (fan-out) packaging, 2.5D CoWoS (Chip on Wafer on Substrate) and the 3D SoIC. Among them, InFO technology is the most mature as well as the cheapest, accounting for about 70~80% of the advanced packaging production capacity (80 to 100 thousand pieces a year). It has been widely used in Apple’s A and M series chips. As the demand for AI chips grows rapidly, the demand for CoWoS advanced packaging technology too is rising. This has led the often overlooked semiconductor “testing” industry to undergo a makeover. In accordance to MA-tek’s strategy of keeping pace with industry trends,we have accumulated rich experience in the failure analysis of advanced packaging. Below, the key tools and applications for this analysis will be deciphered one by one.
|
Characteristics of 2D X-ray Analysis |
X-rays generated through the use of high energy impacts on metal targets possess penetrative properties. This type of imaging can be used to determine whether there are empty solder joints or signs of the HIP or Hop phenomena inside advanced packaging. It can also be used to quickly check whether there are defects such as broken wires or severe burns in the packaging.
|
Characteristics of SAT (CSAM) Analysis |
SAT, also known as CSAM, refers to the formulation of images using the differences in reflection rates and return energies of ultrasound waves in materials with different densities. Different frequency probes are selected according to the ultrasonic penetration rates to detect abnormalities such as delayering, voids and cracks in advanced packaging.
Figure 3. In the packaging health inspection, the role of the SAT is to locate defects. |
3D X-Ray Applications: Nondestructive Inspection and High Resolution Imaging |
What kinds of situations call for the use of 3D X-rays? When there is only one failure sample and destructive analysis cannot be performed, we recommend using a nondestructive 3D X-ray machine to perform ultra-high resolution analysis of the abnormal area. At present, MA-tek has the ZEISS Xradia 520 and the 620 Versa High Resolution 3D X-ray Microscopes..
The principle behind the 3D x-ray is to hit a metal target (W) with high energy electrons to generate highly penetrative X-rays with short wavelengths and high energies. Diffraction waves are then generated by penetrating the object being tested. These waves are received by the detector and converted into visible images by the scintillator. The sample is rotated a full 360゚ on the stage so that tomographic 2D X-ray images can be obtained from different spatial locations. These images are then combined by the computer to produce the 3D X-ray tomographic image. This is also the principle behind computerized tomography. The resolution of the 3D X-ray depends on the size of the pixel. The smaller the pixel, the better. The current spatial resolution limit of this machine is 0.5