Semiconductors Depreciation

[raedbenz]

Hi...
Mechanical systems after long usage their effieciency drops, becasue of wear in mechanical parts, thermal fatigue, rust, etc...
Does this thing happen in semiconductors also?
For example does microcontroller's efficiency drop because of changes in semiconductor material or electrons, because of usage ?
And how important is this in semiconductors efficiency and functionality?

Thanks
Raed

 

[Papabravo]

Generally speaking the answer is no, but there are always exceptions.

1. Flash and EEPROM store charge on the floating gates of MOS transistors. The charge does leak off very slowly. That is why these memories have an endurance and a limit to the number of times they can be programmed.
2. The use of non lead solder (Thanks Europe!) and the development of tin whiskers may have a more dramatic effect on circuits in the future.
3. Electrolytic capacitors deteriorate with age, and often must be replaced to keep a unit functioning.

I'm sure there are other electronic gotchas, but those were the first three that came to mind.

 

[Hero999]

The value of resistors, frequency of crystals and offset of op-amps can drift slighty over time.

It also wouldn't surprise me if the gain of transistors also reduces over time.

 

[Papabravo]

Exceeding the absolute maximum ratings, thermal stress, and static discharge will all compromise the operation of most IC's

 

[Roff]

Here is a presentation from the U.S. Department of Defense that discusses some semiconductor wearout mechanisms.

 

[Speakerguy]

I don't know about silicon, but lots of passive components spec lifetime tolerances or performance specs if you look for them. Even good caps like ceramics and film will have specs for lifetime variation. The only passives I would worry about it in are electrolytics and carbon comp resistors (both dry out).