IC Decapping With Sulphuric Acid
by theglassman in Workshop > Science
119 Views, 0 Favorites, 0 Comments
IC Decapping With Sulphuric Acid
.png)
This is a guide on how to destructively decap ICs with sulphuric acid. Firstly a disclaimer:
DISCLAIMER
I am not a qualified chemist nor chemical safety expert. This experiment will generate toxic fumes and require the use of dangerous chemicals; if you enjoy having lungs/skin/eyes etc go and find another hobby :)
Supplies
Sulphuric Acid - You will need a high concentration, 98% is common for this application. I was forced to buy 25% and increase the concentration by boiling off all the water. This is quite a dangerous process, if you can get high concentration acid (even by using off-the-shelf cleaning products), I would recommend it.
https://apcpure.com/product/sulphuric-acid-25
Hotplate/Burner - Anything to heat up the acid solution will do. If you have no other option, a pan of hot water will probably be sufficient. Do not buy a PCB hotplate just for this purpose, a dedicated chemistry hotplate will be much more suitable (I only include this link for completeness, this is exactly what I used).
Gloves - Most kinds will do, check if safe with high conc sulphuric acid.
Acetone and Distilled Water - These will be used as cleaning solvents. Normal water is fine if you don't have distilled. Acetone is optional.
https://apcpure.com/product/acetone-995?gbraid=0AAAAADk1sy4OXpyfp2vAwl3vFyEXklfq5
Glassware - A conical flask is best suited, beakers are also useful.
Other Bits - Tweezers (get a bag of plastic ones) will be useful, squeezy bottles for the acetone and distilled water are really handy. I wore a respirator any time I was inside near the fuming acid. A temperature probe (immersion or infrared) will probably be useful if you have one.
Setup
I began by fashioning a budget fume hood from a fan, some plastic ducting and zip ties. Setup your hotplate and conical flask, I used tinfoil to ensure any spillage did not stain my workspace/hotplate. I would recommend having a large quantity of tap water available, just in case. I would highly recommend doing all of this outside, if not, proper ventilation/extraction is necessary.
Experiment
If you have a high concentration acid, you can skip straight to the next step. To increase the concentration of dilute acid, begin by heating it gently on the hotplate. I usually begin with a quantity of 125mL 25% sulphuric acid. As the acid comes up to temperature, you will notice the water starting to boil off, once this has begun keep the temperature constant. Do not heat the solution above the boiling point of sulphuric acid (~337°C). You will know when the acid has reached a high concentration as the boiling will stop.
Prepare your ICs by removing as much material as you can (e.g. removing pins on DIP packages, cutting away the sides of long ICs, removing PCB substrates etc). Drop them into the conical flask with the hot acid (you should immediately see bubbling) and watch the solution darken as the acid carbonises the plastic. Leave the solution to react for at least 10 minutes, check how much of the packages remain periodically. Once the ICs have disintegrated a suitable amount, prepare a beaker in which to pour the acid (which you should rinse with distilled water and dry beforehand). Turn off the hotplate and wait for the solution to cool, then pour off the acid into the beaker. Rinse the conical flask carefully with distilled water (the residual acid will react with the water so do this in very small increments). Extract what is left of the ICs and rinse them again with distilled water and acetone.
Pick away any remaining pieces of lead-frame/plastic and find the dies. If there is lots of plastic residue remaining on the dies you can scratch it off with plastic tweezers (some acetone may help). Rinse the dies a final time to remove any large debris.
Results
.png)
I arranged some dies on a microscope slide, these are (L→R): Xbox 360 GPU - Revised Y1 (90nm, 2005), NVIDIA GeForce 7650 GS GPU - G73B (80 nm, 2006), Intel GM45 Mobile Chipset (65nm, 2008), two unknown devices (!), and an ATMEGA16U2 Microcontroller (350nm, 2010). Interestingly, you can see the die size gradually decreasing in the first three devices as the process technology evolves.
I have also included a die-shot of an ATMEGA8L-8PU from an Arduino Uno through my ebay microscope at 10x. Unfortunately I was unable to get a nicer photo for the article as I lost the die on the carpet while trying to clean my microscope slide :'(