Imagine your car air bags deploying randomly when you hit a pothole or speed bump but failing to deploy over half the time in a collision. That is the stark reality with the smoke alarms that are most often found in North American homes. Like most, I had always assumed that a smoke alarm was a smoke alarm. What I now know is that there are two basic types of residential smoke alarms sold in the U.S.: ionization and photoelectric. In real-world fatal fires, these two types of alarms behave very differently. In this case – different is not good. Understanding the difference could very well save your life.
“A smoke detector that sounds approximately nineteen minutes after smoke reached its sensing chamber is like an airbag that does not deploy until nineteen minutes after
a car accident.”
- Judge David E. Schoenthaler, Mercer v. Pitway/BRK Brands (First Alert)
Over 90% of U.S. homes have ionization sensor smoke alarms installed. Around 5% of U.S. homes have photoelectric sensor alarms installed. Approximately 4% have no alarm of any kind installed. (footnote 1) Back in the 1960s, residential smoke alarms were almost unheard of and the fire death rate was about 7 to 8 fatalities per 1,000 U.S. home fires. Between the mid-70s and now, we have gone from about 10% of U.S. homes having smoke alarms to 96% of U.S. homes reporting having at least one smoke alarm. (footnote 2) Surprisingly, after installing smoke alarms in over 100 million U.S. homes over 30 years, the odds of dying in a fire remain about the same. Perhaps it’s just me, but that doesn’t make sense.
Between 1977 and 2009 the number of U.S. home fires and fire deaths have fallen by roughly 50%. However, the risk of dying when a fire occurs today is only slightly lower than in the 1970s. Over the period, the rate fluctuates considerably up and down between 6.5 and 10 deaths per 1,000 fires. This brings into question the value of installing hundreds of millions of ionization alarms.
Overall for the 1977-2011 period, the number of home fire deaths decreased from 5,865 in 1977 to 2,520 in 2011 for a decrease of 51%. The number of home fire incidents also declined steadily for an overall decrease of 49% for the same period. When the death rate per 1,000 home fire incidents is looked at (Figure 1), there is no steady decline, but rather the rate fluctuates considerably up and down. In fact, the death rate per 1,000 home fires was 8.1 in 1977 and 6.8 in 2011 for a decrease of 16%. These results suggest that even though the number of home fires and home fire deaths declined similarly during the period, the death rate did not, and when there is a home fire, the fire death rate risk has not changed much for the period.
The smoke alarm industry points out that all alarms and detectors must meet the standards developed by Underwriters Laboratories (UL). U.S. residential smoke detectors must meet the UL 217 standard. Alarms in Canada have a somewhat different UL-Canada (ULC) standard. For years the major smoke alarm manufacturers–UL, the National Fire Protection Association (NFPA), the Consumer Product Safety Commission (CPSC) and the National Institute for Standards and Testing (NIST)–maintained that either ionization or photoelectric alarms meeting the UL standard afforded adequate protection in most fires. Beginning circa 2006, their recommendations changed. After decades of saying that either alarm was adequate, they started to recommend that we have both types of alarms. The reason for this abrupt shift was never explained.
Most industry studies infer that all fires carry an equal risk of death. An analysis of the underlying data published by these same organizations does not support this position. In general, cooking/fast-flame fires account for a large percentage of fires and injuries but only about 15% of fire deaths. Bedroom and general living area fires are predominantly smoldering fires. This group accounts for only about 12% of fires but over half of all fire deaths and a third of injuries. (footnote 3)
Current UL alarm standards are essentially the same as those developed in the 1970s. Smoke alarm response requirements are defined in the UL 217 standard. The UL tests use a set of standard test scenarios and materials. One scenario is a “fast flame” fire, the other is for smoldering fires. A fast flame fire is the flaming/last stage of a smoldering fire or one based on accelerants, such as gasoline, cooking oils, grease, paper, etc. Fast-flame fires produce large quantities of sub-micron/small fast moving particles. A smoldering fire occurs in the early stages, before open flames develop. Smoldering fires produce slow moving particles. The smoke particles are much larger and tend to be fewer in number but more dense.
The UL smoldering fire test standards were developed when most home furnishings were natural materials such as cotton, wool, etc. To simulate a smoldering fire such as one in upholstered furniture, UL smolders Ponderosa Pine sticks on a hot plate at slightly over 700 degrees with a fan blowing the smoke. Today virtually all furnishings and a large percentage of building materials are synthetic or engineered materials. The behavior and type of smoke produced by burning natural materials is radically different than those produced by burning synthetic ones. Yet the UL standards have not substantially changed for decades.
Under UL test conditions, ionization alarms consistently respond about 30 to 90 seconds faster to open or “fast-flame” fires than photoelectric smoke alarms. However, the vast majority of residential fire fatalities are due to smoke inhalation, not the actual flames. Nearly 2/3’s of fire fatalities occur at night while you sleep. Ionization alarms respond on average between 15 to 50 minutes slower in a smoldering fire than photoelectric alarms. Studies by UL,4 NIST, (footnote 5) Texas A&M (footnote 6) and others found ionization alarms may completely fail to activate in 20 to 25 percent of fires.
Let’s look at testing conducted on smoke alarms by one of our top universities. During 1991-1994, a research team at Texas A&M University, Department of Construction Science, conducted extensive testing on residential fire detection devices. The research project was titled “Full-Scale Research and Testing on Fire Detection Systems in a Residential Structure.” (footnote 7) The Texas A&M study concluded that, during smoldering fires, the probability of a fatality was 55.8% with ionization alarms but only 4.06% with photoelectric alarms. The study also concluded that, in fast-flame fires, the probability of a fatality was 19.8% with ionization alarms but only 3.99% with photoelectric alarms. This testing was based on a fault-tree analysis design developed by Bell Labs for the U.S. Air Force. The Texas A&M research clearly demonstrates that when all factors are taken into account, such as how often each alarm gets disabled due to nuisance alarm problems, to how they respond in actual testing across the full spectrum of fires, photoelectric alarms have a clear advantage.
In 2007, UL published the “Smoke Characterization Study”. This study tested both types of smoke alarms using current testing standards and materials; they also tested the alarms using UL test criteria integrating a variety of synthetic materials. The alarms were tested with burnt toast as well. The results are frightening. Ionization alarms failed the UL 217 test 20% of the time using required test materials. This is the test that the alarms must pass to even be offered for sale in the U.S. When tested using synthetic materials, ionization alarms Did Not Trigger (DNT) in 7 out of 8 synthetic test scenarios, for a 87.5% DNT rate. In the one test where the ionization alarm did trigger, it activated almost 43 minutes after the photoelectric alarm and at a level exceeding the maximum allowed under the UL 217 standard. Understand, this is UL running UL tests and showing that ionization alarms did not respond in 8 out of 8 smoldering test scenarios.
In the same UL tests, photoelectric alarms activated in 5 of 5 tests, or 100% of the time using the standard UL 217 test materials. Photoelectric alarms activated properly in 8 of 8 synthetic material tests, a 100% activation rate. The ionization alarm outperformed the photoelectric in only one case, the burnt toast test. There the ionization alarms responded 22% faster.
It should be noted that there were three test scenarios where neither alarm activated. The researchers determined that the sample size used was too small to generate sufficient smoke. (footnote 8) Those materials were retested using larger samples, and the results included in the above eight test scenarios.
The issue with ionization alarms is far more than just the slow response to deadly smoldering fires. Ionization alarms are notorious for nuisance tripping. They frequently go off when you cook, burn toast, shower, etc. When alarms nuisance trip, people become frustrated and intentionally disable them. This leaves their families completely unprotected. Several CPSC and NFPA studies report that 97% of all nuisance alarm activations are from ionization alarms.9 An Alaskan Public Housing Study shows that about 19% of ionization alarms were disabled within six months of installation; 10 other studies indicate that the percentage may be higher.
“Considering photoelectric smoke alarms are deter-mined by industry experts to be significantly less prone to nuisance alarm and potential disabling of the batteries by consumers, we support and encourage fire service administration and lawmakers that are moving toward the use of photoelectric smoke sensing technology.”
– BRK/First Alert Letter to Vermont fire departments,
July 17, 2008
Remember, about 96% of US homes have at least one smoke alarm. Nearly two-thirds of all residential fire deaths occur in homes that are unprotected. Roughly 50% of homeowners with nonfunctional alarms cite nuisance tripping as the reason for disabling their alarms. To complete the picture, many of the remaining 1/3 of residential fire deaths occur in homes where an alarm sounds, but it sounds too late for the occupants to escape. Over the years a number of government, university and manufacturer research studies, many going back to the mid-1970s, clearly show that ionization alarms are slow to react in deadly smoldering fires and account for the vast majority of nuisance trips.
It has taken decades, but there is finally a growing public awareness of this issue. On October 3, 2012, the NBC Today Show and NBC Nightly News aired a “Rossen Report” investigative segment on this issue. On July 7, 2012 with a follow-up report on August 1, 2012, Huntsville, Alabama TV station WHNT aired “A Taking Action” investigative report featuring ASHI President-Elect Bill Loden. On November 16, 2012, San Francisco CBS 5 “ConsumerWatch” aired a segment with Albany, California retired Fire Chief Marc McGinn and me demonstrating the poor performance of ionization alarms.
The International Association of Firefighters (IAFF) is the largest firefighters union in the U.S. and Canada, with nearly 300,000 members. During the IAFF 2008 conference, they adopted an official position recommending that only photoelectric smoke alarms be installed. The IAFF position also commits the organization to working for changes in the law and model codes to require photoelectric technology alarms. Further, the IAFF position specifically states that combination type alarms are not acceptable. In July, 2010, Albany, California became the first city in that state to require photoelectric smoke alarms in new construction and remodels. In California, the cities of Palo Alto, Orange and the Sebastopol have enacted ordinances requiring photoelectric technology alarms. Shaker Heights, Chagrin Falls and several other cities in Ohio have enacted similar ordinances.
In 2011, the California Real Estate Inspection Association (CREIA) became the first home inspection organization to take a stand when it adopted a position mirroring the IAFF position. At the January 2013 Board of Directors meeting, ASHI became the first national home inspection organization in the world to take a stand when it adopted a pro-photoelectric alarm position. At this time, Vermont, Massachusetts, Maine and Iowa have laws that require photoelectric technology alarms in residential construction. Similar action is under consideration in several states. Cincinnati, Ohio recently became the largest U.S. city to enact a photoelectric ordinance. The ordinance covers rental housing. Six smaller cities in Ohio have photoelectric ordinances. New York City is currently considering a photoelectric ordinance. The Northern Territory in Australia recently adopted a photoelectric technology law.
The New York Senate and Assembly are currently considering a photoelectric smoke alarm tax credit referred to as Averyana’s Law. (footnote 11) As part of the justification, the law states: “On March 11, 2012, two year old Averyana Dale tragically lost her life due to smoke inhalation in a fire in Auburn, NY. “ It was later determined that the fire was a smoldering fire, which produces a significant amount of smoke but very little actual flame. Averyana Dale and her godmother most likely lost their lives because the ionization smoke detector that was present in the home did not alert to the fire until it was too late. If a photoelectric detector had been in the home, it is considerably more likely they would have been alerted to the smoke significantly sooner and would have made it out safely.
“Nationally, the percentage of people dying when the smoke detector works, but works too late, is approximately 40 percent.”
– Jay Fleming, Boston Deputy Fire Chief,
CBS Boston Interview, 2007
WHICH TYPE OF TECHNOLOGY DO I HAVE? It is not always possible to know. In general, if the label says anything about radioactive material, Americium-241 or the model number has an “I,”—then it is almost certainly an ionization alarm. If you have any doubt, there is over a 90% chance that the alarm you have is an ionization unit. Photoelectric models often have the word “photoelectric” or the capital letter P printed or embossed on them. To be safe, simply replace any unknown units with photoelectric-only alarms. Any smoke alarm that is 10 years old or older should be replaced regardless of type.