Advertisement

How Long Do Carbon Monoxide Detectors Last?

Take the alarm off its bracket and turn it over. Somewhere on the back is a manufacture date, usually printed rather than stickered, and most guides tell you to compare it against a service life and put the unit back. That comparison answers half the question. The same date tells you something else about your carbon monoxide detector, which is whether it belongs to the generation that announces its own death or the generation that does not, and the dividing line is a specific day in 2009. An alarm that has never chirped at you is not therefore a working alarm.

This is general safety information, not medical advice. If a CO alarm sounds or you suspect exposure, move to fresh air and contact emergency services.

Advertisement

In 60 Seconds

  • Residential CO alarms run 5 to 10 years, and the figure is set by the model rather than by the category.
  • Alarms produced after 1 August 2009 carry an end-of-life warning. The Seattle Fire Department publishes the date; older units were not required to have one.
  • UL states that the end-of-life signal may be the same signal as a low battery, so the obvious fix is the one that hides the problem.
  • Silence from an old alarm is not evidence of a working sensor. It may be evidence of nothing at all.
  • The date on the back is the only figure that applies to your unit. Every range on this page, including ours, is a range.

The Short Answer, and Why It Is Not the Useful One

Residential carbon monoxide alarms last somewhere between five and ten years. Kidde publishes a ten-year life for everything it has built since 2013, seven years for anything older and for three legacy models it names by number. First Alert's guidance runs to five to seven years. The Seattle Fire Department, writing for householders rather than for customers, puts the life expectancy at around seven years.

Those figures disagree because they are describing different products, not because anyone is wrong. Which is why the range is close to useless to you personally: your alarm has one service life, it is printed on the unit or in the manual, and no article can tell you what it is. What an article can tell you is what else that date on the back means, and there the whole field goes quiet.

The Second Thing the Date on the Back Tells You

Carbon monoxide alarms have not always been able to tell you when they stop working. The end-of-life warning is a requirement that arrived on a date, and the Seattle Fire Department states it plainly: "All CO alarms produced after August 1, 2009 have an end-of-life warning notification that alerts the resident that the alarm should be replaced."

Read that sentence for what it implies about the other side of the line. A unit produced before 1 August 2009 was under no such obligation. Its sensor still had a service life, and it still reached the end of it, and it was not required to make any sound about it. The alarm keeps its power light. The test button still produces a noise, because the test button checks the horn and the electronics rather than the sensor's ability to detect gas. Nothing about the unit changes in a way a person standing in a hallway can see.

So on an old alarm, silence carries no information. That is the reverse of what most people assume about a device whose whole job is to make a noise, and it is the reason the manufacture date is worth reading even when you are not counting years. If the date on the back of your unit begins with 2008 or earlier, you are looking at a device that may have finished years ago without saying so.

Why a Post-2009 Warning Still Gets Missed

The newer alarms do announce themselves, and a surprising number of those announcements get answered with a battery.

UL, which publishes the standard American residential CO alarms are listed to, describes the signal this way: it "is distinct from CO and smoke alarm signals, but may be the same as other trouble signals, such as a low-battery signal." The end-of-life warning has to be different from the alarm that means gas. It does not have to be different from the chirp that means the battery is flat.

That permission is the whole failure mode. A householder hears an intermittent chirp, recognises it as the low-battery noise they have heard a hundred times, fits a fresh cell, and the chirping stops for a while or does not stop at all. Either way the unit stays on the wall. Kidde's own end-of-life chirp runs at thirty-second intervals and a new battery will not silence it, which is the tell, but only for somebody who already knows to listen for the interval rather than for the sound. Our guide to what CO detector beep patterns mean sets out the intervals brand by brand.

Service Lives, by the Manufacturers Who Publish Them

Two of the three results that rank above this page are alarm manufacturers, and that is the one part of this topic where a manufacturer is the best source available: they are describing their own hardware.

Maker and vintagePublished service life
Kidde, built 2013 onward10 years
Kidde, built before 20137 years
Kidde models KN-COEG-3, KN-COPE-I, KN-COPP-37 years, named exceptions to the 10-year rule
First Alert, general guidance5 to 7 years

Use the table to sanity-check what you find on your own label, not to replace it. A combination smoke and CO unit, a sealed-battery model and a plug-in with backup can all carry different figures from the same maker in the same year, and the sealed battery is a particular trap: a ten-year battery is a claim about the power source, and the sensor beside it has a service life of its own.

Why the Sensor Wears Out at All

An electrochemical CO sensor is a small wet cell. Gas diffuses through a membrane into an electrolyte, reacts at an electrode, and the current that reaction produces is what the alarm reads as a concentration. Nothing in that arrangement is inert. The electrolyte loses water in dry heat and takes it up in a steamy bathroom, the membrane collects dust and cooking grease, and the electrode surface is slowly changed by every gas it meets, including the ones it is not looking for. Household aerosols, solvents and cleaning products all leave a mark.

The unit does not fail at the end of that process so much as drift through it. Manufacturers build in enough tolerance to absorb the first years of drift and then set a date beyond which they will not vouch for the reading. That is what a service life is: not a prediction that the sensor will stop, but the point past which nobody is claiming it still measures what it says it measures.

Europe Made the Same Warning Impossible to Confuse

The American rule leaves the end-of-life warning free to sound like a flat battery. The European standard for domestic CO alarms took the opposite decision, and the comparison is worth a paragraph because it shows the design problem was recognised and answered somewhere.

EN 50291-1:2018, which replaced the 2010 edition of the same standard, made an end-of-life indicator mandatory and required it to be audible and visible. Kidde's European arm describes what that produces in a finished product: a warning that "consists of both a beep and flashing lights to provide both audible and visual alerts", on a sensor its documentation puts at a ten-year design life. A householder in Manchester gets a light they have never seen before alongside the chirp. A householder in Milwaukee may get a chirp indistinguishable from the one a new battery fixes.

Neither standard covers the other's market and nothing here changes what you should buy in the United States. It changes what you should conclude from a chirp: in this country the sound alone does not tell you which of two things is wrong, and the interval and the label are what settle it.

Three Dates, Three Different Jobs

Everything above resolves into one action, which is to go and read the back of each alarm in the building and sort it into one of three cases.

  • Manufactured before August 2009. Replace it, whatever it is doing. It is past any published service life, and it was never required to be able to tell you so.
  • Manufactured after August 2009 and inside its service life. Write the replacement month on the front edge in marker, where you can read it without taking the unit down, and set a calendar reminder a month before it. This is the only case in which nothing needs buying today.
  • Manufactured after August 2009 and past its date. Replace it now, and do not wait for the chirp to confirm what the label already says. A warning you have not heard yet is not a reprieve.

If the unit chirps while you are handling it, the interval is the diagnostic and it is covered on the beeping guide. If you find you are short of alarms rather than short of working ones, the coverage checklist works out how many the layout needs, and placement decides where the replacements go.

If the Alarm Is Not Yours

A tenant reading an expired date off a landlord's alarm has a different problem from a homeowner, which is that the device, the decision and the money all belong to somebody else.

Across the 51 US jurisdictions in our state law records, 44 record a landlord duty that reaches a carbon monoxide alarm, and the recurring shape of it is an obligation to provide working alarms and to repair or replace them once the tenant gives notice in writing. Most say so in a provision about alarms; a few, Indiana among them, get there through a general habitability clause instead, which is worth knowing before you quote a section number at anyone. Written notice is the step tenants skip and the step that starts the clock, so send it through the maintenance portal or by email rather than mentioning it in a corridor, and say what the manufacture date is. Our rental CO detector laws by state pages set out the duty split and the notice period where you live.

Sources & References

Last updated: August 7, 2026

Advertisement