Silicon Pr0n

This page is read only. You can view the source, but not change it. Ask your administrator if you think this is wrong.
Good video: https://www.youtube.com/watch?v=ZQeHHYJYWXo

On torch vs acid quality: https://twitter.com/ringoware/status/1314766880893136897?s=20


====== General idea ====== 

Epoxy becomes brittle at elevated temperatures. Specific temperatures are under research, but seems to very from epoxy to epoxy. 

====== Heat shock ====== 

When rapidly cooled, stresses form microfractures in the brittle epoxy. When repeated a few times, the epoxy can be easily broken apart at room temperature.

Procedure
  - Prepare coolant: mix salt, water, and ice
  - Heat epoxy until red hot
  - Drop into coolant as soon as possible after removing heat
  - Try to apply light pressure with your fingers. If the chip crumbles, you are done. If not, repeat. Early tests showed chips tended to take 3 - 5 times

Advantages
  * Quick
  * Readily available materials
  * Relatively safe compared to high temperature acids

Disadvantages
  * Potential to torch die (eg: unnoticed crack during repeat)
  * Potential fire hazard
  * Burnt chip fumes
  * Heat shock damage to die itself?

Notes
  * Some passivation layers seem to push epoxy up (TODO: add picture). Leads to very clean removal
  * Metal packages work best: large temperature coefficient difference
  * Large dies work better
  * Recommend putting heat resistant barrier over ice so that it does not float to top and get in the way of the chip (ex: fiberglass)
  * More iterations should continue to make the chip more brittle
  * Unknown the potential heat damage to the silicon. Early experiments show its negligible for many parts

Media
  * "Virtex heat shock decap": [[http://www.youtube.com/watch?v ======rHvKBStE_ek]]
    * Ideal package: metal layer, expanding passivation, and large die

====== Thermal ====== 

Technique needs more experimenting. Ideal for when you have low equipment and a large supply of small ICs.
Most of the experimentation with this has been done so far on PDIPs. In general, SOIC and similar small chips have actually shown better results, possibly due to the thinner (weaker) epoxy holding the die in place. The main issue is that the die tends to fall off of the carrier island from the glue melting before it releases from the epoxy. A future research topic is if a cold plate (peltier, flowing water, etc) could be used to encourage the die side breaking firs.
Good heat sources are a heat gun or directed lighter. A torch or directed lighter in some cases has shown better results if acting quickly. This may be because the die has not yet had time to heat up and melt the glue, allowing for a better chance of a clean break.
Often times one will be left with a die with a chunk of the epoxy die stuck to it. This can either be finished off in nitric acid if available or

Procedure: best chance for a clean break (under development)
  - Place a sturdy part of chip in a vice. If a clean break is highly desired, place the bulk part
  - Apply heat. Torch should be brief, heat gun should be continuous
  - Work the center part (above the die). No significant difference has been observed on which side is compressed/stretched.
  - With luck, the epoxy has shattered and the die has very cleanly been released
  - Remove heat (if applicable)

Procedure: best chance for most of a chip (under development)
  - Clamp legs into a vice
  - Apply heat. This will work best with a heat gun
  - Work the legs by curling to the bottom of the die
  - After removing all of the legs, you should be left with half of a chip or a full chip that can easily be broken in half
  - Chip away the epoxy on the remaining half until the die can be grabbed with tweezers
  - Gently pull the die out with the tweezers
  - Turn off heat

Advantages
  * Easy to do
  * Quick
  * Low start-up and recurring costs

Disadvantages
  * High damage, residues, and/or complete miss rate
  * Low repeatability
  * Higher chance of damage on larger dies
  * Removes some of the top metal (thin layer, nut chunks, does not seem to effect imaging)