I'm having a hard time understanding a requirement to replace smoke detectors every 10 years (or whatever the manufacturer suggests). With previous generations of ionization based detectors, I understand that the radioactive substance would decay over time, reducing its ability to reliably do what it's supposed to do, but the photoelectic ones I'm having a harder time understanding exactly why there is still a requirement to replace them.

For example, I own 2 Nest Protect 2nd generation smoke/CO detectors, which as I understand use a split-spectrum photoelectric sensor. The device will trigger all kinds of alerts and warnings when it comes time to replace them. However, as this is LED based, why does it do this? Is there an actual technical reason for replacing it, or does this basically come down to some law somewhere?

  • 2
    While modern smoke detectors may last longer, CO detectors will still degrade over time due to the different technologies involved.
    – Chenmunka
    Commented Oct 31, 2017 at 11:43
  • I had to replace all of mine two years ago. One started chirping even after replacing the battery (also hardwired). Checking the info on the unit it was rated for 15 years and was now 20. Commented Oct 31, 2017 at 13:36
  • Keep in mind that there are two versions of the Nest Protect, one of which has a "lifetime" non-replaceable battery. That's the key reason for the lifespan of that unit, I think.
    – isherwood
    Commented Oct 31, 2017 at 14:57
  • 2
    @DanielGriscom now this I did not know! I always thought it was due to the degredation of the material, becoming less effective as time goes on, but as you've demonstrated that's not the case. Interesting indeed!
    – dannosaur
    Commented Oct 31, 2017 at 17:16
  • 2
    @DanielGriscom: It's an exponential process, so after 100 years the activity of the Am-241 in your smoke detector will have reduced by ~16%.
    – caf
    Commented Nov 1, 2017 at 3:52

3 Answers 3


It likely has to do with the lifetime of the hardware itself. Remember, there's a circuit board and a light source, as well as a detector. Those don't last forever. So the manufacturer certifies the device will work for only 10 years, and then (in some modern units) sets a hard sunset by using an unreplacable battery.

In some regards, this solves the problem of people just putting new batteries into older detectors forever, not realizing the devices have stopped working (when was the last time you tested one?)

  • 1
    I see what you're saying. So it's more along the lines of the manufacturer covering themselves by guaranteeing that the hardware will work for 10 years (or whatever), and if you continue to use it beyond that, and it fails because the hardware has indeed degraded, and your house burns down because it didn't sound an alarm because of that, they're not liable? I test my smoke alarms every couple of months, primarily because Nest makes it so easy to do so. If they were normal ones, probably not so much!
    – dannosaur
    Commented Oct 31, 2017 at 13:15
  • 4
    I think you've got it. Limiting liability is the upshot here.
    – Machavity
    Commented Oct 31, 2017 at 13:16
  • 1
    @dannosaur Probably also reducing costs. Things naturally decay, especially electronics, especially if not specially sealed, and with something like a smoke detector that's placed anywhere and everywhere, it'd be very difficult to make it extremely long-lived. Setting a lifespan of 10 years is probably just as much to make it possible to cheaply manufacture as it is to limit liability.
    – anon
    Commented Oct 31, 2017 at 15:25
  • 1
    @dannosaur it's not as cut and dried as that. The more they do to deter you from using a smoke detector too long, the more it reduces their liability and the cost of litigation. It can make the difference between getting thrown out at preliminary hearings, or grinding all the way through to a jury and having risk of appeal. Better the case be so cut-n-dried the judge tells them "nope" and threatens to sanction the attorney if he brings another bowser of a case into his courtroom... Commented Oct 31, 2017 at 16:02
  • I've ended up with a smoke detector in my kitchen (which is really something I should probably take up with my landlord), so it gets a good test whenever I use the grill ;)
    – Muzer
    Commented Oct 31, 2017 at 16:06

Highly condensed version:

  1. studies have shown that the electronics within the detectors fail at a rate of 3% per year.
  2. The 10 year mark was chosen based on the cumulative fail rate being over 25%.


  • Council of Canadian Fire Marshals Fire Commissioners
  • National Fire Protection Agency
  • Underwriters Laboratories

There is little to gain for manufacturers to create something that will never need to be replaced.


Smoke alarms come in two categories (sometimes combined). Photoelectric and ionization.

The lifetime is directly related to the technologies and the environment.

The technology limitation of the ionization is that a small amount of a radioactive isotope is used to ionize the air/particle near the electrostatic detector plates. The particles attracted to the plate indicate particles (smoke) in the air.

The radioactive isotopes decay and the ionization grows weaker - thus impairing the ability of the alarm to count as many particles.

Environmental conditions can also impact lifetime. Smoke can build up on electrostatic plates (ionization type) and impair their ability. Similarly smoke can build up on photoelectric emitters and detectors.

The detectors do attempt to detect these issues and one (not guaranteed) response is to beep to annoy you into their replacement.

As it's difficult to vette these impairments over time it is truly best to replace them every 10 years or so and the batteries every year (put the partially used batteries in less critical items).

Here's a very approachable short summary I found to accompany this response:


Consider a smoke alarm in a kitchen. Boiling pasta water - particulate laden water vapor; stir fried foods, pan fried meats, burn off of spills on the range.

In particular the greasier smoke from cooking for example is particularly likely to impact both photoelectric components as well as ionization electrostatic plates.

So kitchen alarms are often the most critical to replace early and often.

Other places to watch are bathrooms (steamy water, particularly if not very soft), shops and garage (think sawdust), your porch near your smoker.

Please don't wait until they fail into constant beeping to replace them. That's their last ditch notice to you and they've likely been highly compromised long before then.

  • 4
    While the accepted answer properly addresses the life of the hardware itself (MTBF: "Mean Time Before Failure" figures in here), your "Environmental conditions" is also important. They just get dirty! And dirty sensors can't work properly.
    – Auspex
    Commented Oct 31, 2017 at 15:44
  • 1
    Hi! You have a couple of unregistered accounts. Please register one of them, then merge them together, which will allow you to edit and comment on any of your posts. Thanks, and welcome to the site!
    – Niall C.
    Commented Oct 31, 2017 at 17:41
  • 1
    One of the more common radioactive elements used in smoke detectors is Americium 241.It has a halflife of roughly 432 years, so even after 10 years it is active enough to work well in a detector. The reason for 10 years is really loss of sensitivity and an average mtbf of a lot of the parts being not that much higher.
    – PlasmaHH
    Commented Oct 31, 2017 at 20:03
  • Dirty sensors and depleted isotopes work too well. The sensitivity increases. Frequent false alarms are as good as no alarms at all. They could be built to expire like printer toner cartridges, or warn if a 6 month test has not been performed.....It's only a matter of time before they are all networked and self order their own replacement too.
    – mckenzm
    Commented Nov 1, 2017 at 19:37

Not the answer you're looking for? Browse other questions tagged or ask your own question.