How well do IC's survive fires ?

[lordvader88]

Like if a stereo or computer is a house that burns down, what's left of the transistors and chips after ? Silicon melts +1400C, so it shouldn't melt at least, but IDK about the bond wires, or leads.

[ogden]

Like if a stereo or computer is a house that burns down, what's left of the transistors and chips after ? Silicon melts +1400C, so it shouldn't melt at least, but IDK about the bond wires, or leads.

Zero chances. You can damage IC's even with your hot air gun, not to mention house fire.

[T3sl4co1l]

Chips are often processed in the 400C+ range, so expect fine-pitch devices to fail due to uncontrolled diffusion around there.  Not to mention oxidation of metals (RIP aluminum film?), silicon oxide growth, and other metallurgical problems (say for copper interconnect devices?).

Probably, a lot of chips subject to modest pyrolysis (enough to char the package) can still be used if rebonded, and possibly if re-metallized as well (which, mind, would be a pretty heroic effort, if the top metal layer is just trashed, and you don't know what the chip was originally).

Chips are so cheap that it's very much worthwhile to consider even mere loss of legs/pins/pads a total loss of the chip.  It's still perfectly fine inside, it's just not nearly worth getting at.

It can be justified if the chip contains data worth the while: Flash chips, or stacked CPUs for instance.  For which, you can find videos of repairs done by grinding into the package, locating a particular trace or pad in the interposer, and running a bodgewire to the PCB.

Tim

[daqq]

How well do IC's survive fires ?

They don't.

[T3sl4co1l]

Like if a stereo or computer is a house that burns down, what's left of the transistors and chips after ? Silicon melts +1400C, so it shouldn't melt at least, but IDK about the bond wires, or leads.

Zero chances. You can damage IC's even with your hot air gun, not to mention house fire.

This may be delamination (package lifts off pads/substrate/chip), or cracking (chip is bent, and ultimately cracked, by internal pressure).  In the former case, the bond wires would break, but the chip would still be intact.

At least, I assume that's what he's getting at.

For full devices to still be viable in a rework context, you of course need the package neither charred, nor stressed.  The peak temperature could still be in the 300C range, but not for very long and probably not for very much rate of survival (reflow is normally 260C, so this isn't really that horrible).  Lesser temperatures are very ordinary (150-175C max storage/operating temperature, for example), and could still cause thermal stress, but are unlikely to damage parts alone.

Contamination from smoke and fumes may be a problem I suppose.  Acid fumes from PVC, for example, will cause recurring corrosion problems.

Tim

[CJay]

The answer is, as always, depends.

I've recovered data from hard drives that have been in fires and the ICs had desoldered from the heat of the fire, reattaching them after cleaning up the smoke residue and the drive worked.

The machines in question hadn't been *in* the fire but were in the room where the fire broke out.

Chances of an IC surviving a fire that engulfs it, who knows, the package probably wouldn't survive but might if it was ceramic?

There are just too many variables including temperature of the fire, length of exposure, even the methods used to extinguish the fire

[Stray Electron]

   I have an HP 67 calculator that survived a fully involved house fire. The corner of it's factory plastic box was charred and one small corner (about 3/16") of the calculator melted but it still works fine.

[ogden]

I have an HP 67 calculator that survived a fully involved house fire. The corner of it's factory plastic box was charred and one small corner (about 3/16") of the calculator melted but it still works fine.

Seems, we shall define "house fire" first My suggestion: take lowest fire safe specs (Class E): 30 minutes in 1550 degrees F (843 ^oC).

[Stray Electron]

  House fire: Enough to burn the entire structure.  There, that wasn't too difficult was it?

[ogden]

  House fire: Enough to burn the entire structure.  There, that wasn't too difficult was it?

Dumbest definition of house fire ever. What if structure of the house is made out of material that does not burn (concrete) - then what?

[T3sl4co1l]

Houses don't burn very evenly, especially once the firefighters get there and start spraying things with water.  More common is, there's a lot of burned stuff in the ignition room, melted stuff at the bottom, or upstairs or in neighboring rooms, smoke damage most everywhere, etc.

Tim

[Stray Electron]

  House fire: Enough to burn the entire structure.  There, that wasn't too difficult was it?

Dumbest definition of house fire ever. What if structure of the house is made out of material that does not burn (concrete) - then what?

   "House Fire". It isn't that difficult of a concept.

    If the house had been made out of nothing but concrete then it wouldn't have burned would it? But since it did burn then it's safe to assume that it wasn't made of concrete.

[ogden]

Houses don't burn very evenly, especially once the firefighters get there and start spraying things with water.  More common is, there's a lot of burned stuff in the ignition room, melted stuff at the bottom, or upstairs or in neighboring rooms, smoke damage most everywhere, etc.

That was exactly my point - better we define conditions IC's are exposed-to, not just plain fact that house is on fire. House could burn down completely yet some lucky stuff buried under rubble can survive never experiencing high temperatures. When we find such "lucky calculator", only conclusion we can draw is that we can't draw any conclusions.

[ogden]

   "House Fire". It isn't that difficult of a concept.

Seems you deliberately misinterpreted my suggestion just to do pointless tongue gymnastics and waste precious ink of the internet. I did not ask you to define how exactly your house burned down, I suggested to define temperature and time IC is subjected to which would be much better *measure* of "house fire".

[Stray Electron]

    Go back and read it again. I never drew any conclusion I simply stated that I have a calculator that had survived while inside of it's box.  That's one case, it's not a generality or a conclusion.  If you want to rate IC survival is different fire categories then you need to go some controlled tests and not rely on anecdotal evidence.

[Stray Electron]

   "House Fire". It isn't that difficult of a concept.

Seems you deliberately misinterpreted  my suggestion just to do pointless tongue gymnastics and waste precious ink of the internet. I did not ask you to define how exactly your house burned down, I suggested to define temperature and time IC is subjected to which would be much better *measure* of "house fire".

  No, no one "misinterpreted" anything until you stepped in to argue about weather the case I described was really a fire that meet your set of arbitrary conditions. The OP, not you, asked a general question but then you started disputing the conditions of a particular fire that you know absolutely nothing about.

[ogden]

  No, no one "misinterpreted" anything until you stepped in to argue about weather the case I described was really a fire that meet your set of arbitrary conditions. The OP, not you, asked a general question but then you started disputing the conditions of a particular fire that you know absolutely nothing about.

Where exactly did I argue about anything or disputed conditions of particular fire? 

Seeing your anecdotal example of calculator that survived "house fire", I suggested to define conditions of "house fire", suggested temperature and time. You conveniently ignored suggestion just to make big thing out of "house fire" with your sarcastic, pointless comment.

[Gyro]

As the OP has departed and it was a pointless question in the first place, is it really worth getting heated about?