Joined: 26 Nov 2025
Posts: 3
Location: Korea
 Posted: Fri Nov 28, 2025 7:16 am Post subject: IMC crack at time zero.....(T_T) |
 |
| IMC cracks were occurred at time zero.(=EOL) and electrically failed.
The phenomenon is very unique. - Only 2pins are failed out of 1 thousand pin. - IMC growth was not observed....just 1um. What is the root cause any possible? Best RDS,...... |
|
  |
Guest
| Posted: Fri Nov 28, 2025 11:04 am Post subject: |
|
| Hi Nurung73,
You're referring to intermetallic compounds formed between solder bumps and underbump pads of your package, right? What package is this? Are you using Pb-Sn solder bumps? What about your underbump pads - what are their metal composition? What's the location of the affected pins? Rgds, SJR |
|
Joined: 20 Jun 2026
Posts: 43
| Posted: Fri Nov 28, 2025 11:51 am Post subject: |
|
| Hello Nurung73,
Did you do SEM inspection on the fracture surface and affected bumps already? Did the fractures occur through the IMC itself or not? Are there no signs of other mechanical damage nearby? Are the damaged bumps next to each other? FARel Engr |
|
| |
Joined: 26 Nov 2025
Posts: 3
Location: Korea
| Posted: Fri Nov 28, 2025 12:30 pm Post subject: |
|
| I'm very sorry and thank you for responses...
It's a wire bond ball IMC crack. Au5Al2......so so..... The package is BGAs. Any damage was not observed near the failed pins. failed area is random. |
|
| |
Joined: 20 Jun 2026
Posts: 43
| Posted: Fri Nov 28, 2025 2:24 pm Post subject: |
|
| Hi Nurung73,
It's good that you clarified we're talking about Au ball bond IMC failure. I would assume that the ball bonds lifted, causing the electrical failures. Is the intermetallic coverage area greater than 60% of the ball bond area? If not, how low is it? If the IMC is good, then things are a bit tougher. You might need to cross-section the ball bonds to determine the quality of the intermetallic growths up the ball bonds. What is the failure rate? Is it high enough to merit the resources needed to find the root cause of this issue? BTW, if the failure rate is high, this may also be a rel issue for units that pass etest. Good Luck! Thanks, FARel Engr |
|
| |