Our national fire statistics system, the National Fire Incident Reporting System (NFIRS), does not collect information concerning the proportion of structure fires that start as smoldering fires. Since it is hard to imagine that there are any fire service officers who would think this is an unimportant question, the reason for this omission has to be in feasibility. Presumably, it was judged that firefighters respond predominantly to flaming fires, and once they encounter a flaming fire, they do not have a basis for assessing whether this fire started as a smoldering fire or not. This may be partly true, but it is also true that even a modest effort at investigating the circumstances of a particular fire might well provide the answer to this question. Be that as it may, statistics are not collected.

This does not, however, mean that useful estimates cannot be made. In 19891, The International Fire Chief reported that “most home fires start from a smoldering source.” In 1963, the National Fire Protection Association (NFPA) reported that 75% of dwelling fires originated as smoldering fires.2 Much as is the case for NFIRS, the NFPA does not currently provide data on the fraction of structure fires starting in the smoldering mode. Queries to current NFPA staffers failed to elicit any history of these old—but very important—findings.

Smoking materials (cigarettes, primarily) are a major source of smoldering fires. Here, there have been two notable changes. Cigarette sales have dropped drastically, by about two-thirds since 1980. Between 2004 and 2011, all states adopted legislation requiring that only reduced ignition potential (RIP) cigarettes, also called fire-safe cigarettes, be sold. The implications of this will be considered below. Thus, we do not expect that between 50% and 75% of structure fires today begin in the smoldering mode. The fraction necessarily is lower, but it still is significant.

All of combustion can classified as either flaming or smoldering. Technically, the former is referred to as a homogeneous reaction—that is, combustion reactions occur throughout a volume. Smoldering, however, is heterogeneous combustion, meaning that the reactions only occur at a surface, not throughout a volume of gas. Both types of combustion cause high temperatures. However, only flaming fires exhibit visible flames. Smoldering fires are likely to manifest a glow, but this will not necessarily be visible in cases where a layer of ash or char obscures the surface.

It is important to appreciate that the scientist uses terms more narrowly than does the fire chief or the layman. To a scientist, smoldering combustion means an absence of flames. Others, however, are likely to view “smoldering” fires as ones that show only small or limited flames, as distinguished from no flames.

A smoldering fire can start either from an external or internal source of ignition. External sources of ignition can be just about anything, ranging from cigarette lighters to red-hot rivets. An internal source of ignition, on the other hand, means that runaway self-heating led to spontaneous combustion (photo 1). Thus, all spontaneous combustion fires start as smoldering, but not all smoldering fires start because of spontaneous combustion.

(1) Self-heating of bales of hops ended up producing spontaneous combustion, which manifested as smoldering. Note the circular smoldering pattern at the surface. [Photo courtesy of Yakima (WA) Fire Department.]

Certain substances’ tendency to undergo spontaneous combustion is one of the few areas within ignition amenable to both quantitative testing and theoretical analysis. The Ignition Handbook3 covers these topics at length.

Fire investigators or fire training officers are unlikely to undertake testing or calculations pertinent to self-heating or to smoldering, but one question is crucially important to them: What materials are susceptible to smoldering? Smoldering Fires4 provides extensive lists and documentation. Here, we can take a simpler look.

For smoldering to be possible, the fuel in question has to be capable of producing a noncollapsing charcoal-like appearance. Thus, it must be capable of charring, its melting tendency—if any—has to be limited, and the char has to exhibit a certain amount of mechanical strength. In practice, most materials capable of smoldering are granular materials, which are woody or are agricultural products.

Very few plastics can smolder, notably some grades of polyurethane foam. These grades of polyurethane foam, however, are unlikely to be found outside a combustion researcher’s laboratory, since safety tests serve to restrict commercial usage of such materials. A few other types of rarely found materials can also smolder—for example, dusts of certain metals. Primarily, though, the overwhelming fraction of smolderable materials are wood-based or agricultural products.

It is also known that whole wood (e.g., 2 × 4s) can smolder. Most people are aware that granular wood products, such as sawdust or wood chips, will smolder readily. But how does whole wood smolder? The answer is that prolonged exposure to heat creates porosity and a charcoal-like appearance. Thus, for example, if there is a bad electrical connection at an outlet, the first fuel to start burning is likely to be wire insulation, but wire insulation burning, by itself, may not lead to a house fire. What tends to happen is that the outlet is likely to be nailed to a 2 × 4. This wood member starts slowly heating and eventually begins to smolder. Much later, a flaming fire may break out there.

Cigarettes are a unique source of smoldering (they smolder, they do not flame) and a potential ignition source to other smolderable materials. There are no “equivalent” products. Thus, understanding the role of cigarettes in smoldering fires is important. Without the necessary information, a person might conclude that RIP cigarettes will not ignite target fuels, since that is the sole reason for their existence, but the roles of reduction in smoking and progressively increasing dominance of RIP cigarettes cannot be separated out with respect to how effective RIP cigarettes have been in reducing fires. What is disturbing, however, is that in numerous studies, laboratory research has shown that RIP cigarettes readily ignite ignitable materials.

In study, the experimenters used commonly found smolder-prone products, such as mattresses or upholstered furniture, and found that RIP cigarettes successfully ignited these fuels. For example, California6 reported that the “majority of [RIP] cigarettes did cause smoldering ignition of furniture mockups, certainly with smolder-prone cover fabrics. Also, many RIP cigarettes burned their full length when laid down in an ashtray. It seems like the ASTM standard for testing RIP cigarettes using filter papers does not adequately address the smoldering propensity of those cigarettes in real life.” A number of other researchers obtained similar or identical results.7-14

To consider quantitatively the rate of burning of smolderable materials requires considering the details of the geometry associated with the smoldering process. In laboratory tests, commonly geometries are set up where the smolder front moves linearly through the fuel material. This is easier to study, but most fires do not smolder in a 1-d progression through the fuel. Instead, the smolder front movement is likely to be more complex; a 1-d progression through a smolderable material means the smolder front moves like a flat sheet through the material. This is sometimes set up in specialized rigs at universities, but this is not how smolder tends to progress through most real materials.

So you can learn some things by studying smolder in a rig set up to provide a smolder front of this sort, but then how can this be applied to real life? The rigs for doing this are usually a wind tunnel, where if you ignite the fuel bed in a certain way, you can elicit a 1-d progression.

Nonetheless, an important understanding can be obtained from simplified experiments. Perhaps the most important experimental result is the rate of smolder front propagation—i.e., how many millimeters (mm) per minute does the front move at? It turns out that much of the experimental data is in the range of 1 to 3 mm/minute. This allows you to make certain fire development estimates.

What may be the most important factor to appreciate, however, is the low rate of mass loss, or the low heat release rate (HRR) when a fire burns in the smoldering mode. To visualize the order of magnitude involved, if a typical fuel package might be considered to burn at 1 megawatt (MW) in the flaming mode, it is likely to burn at 1 kilowatt if it is smoldering instead. This is an enormous difference. The toxic gas of most significance in fires is carbon monoxide (CO), and the CO to carbon dioxide (CO2) ratio is likely to be higher in a smoldering fire than in a flaming one, but since the mass loss rate is so low for smoldering, people do not die from short-term exposures to smoldering fires. Instead, lethalities occur either because of a cumulative exposure many hours long or when smoldering fires transition to flaming and mass loss rate and HRR increase vastly.

Fire investigators have long wondered about cases where an unreasonably small amount of material susceptible to self-heating ended up doing exactly that, resulting in smolder and possibly flaming later on. The materials can be flowerpots, single bags of charcoal briquettes, bags of woody mulch, and so forth.3 Some such reports are likely to involve a bad investigation, but enough incidents of this type have been reported so that all such incidents cannot simply be dismissed as constituting bad investigation efforts. What research has indicated, instead, is that some materials exist which can either promote or inhibit smoldering combustion. These are usually unintended contaminants. Smoldering Fires4 features the first systematic effort to tabulate and discuss these contaminants.

Smoldering Fires4 is the first-ever book devoted solely to this topic. It is focused wholly on the needs of the fire investigator and the fire service training officer. The information is all practical, and unlike other books of mine, I don’t include theory sections but I do give references to readers who would be interested in this material. I omitted theoretical treatments not featured because the state of the art is such that useful techniques do not yet exist for using theory to solve practical problems.

Note that this refers solely to the smoldering process. Self-heating may cause smoldering, and self-heating does indeed offer theories that can be helpful in the analysis of practical problems. For readers needing this information, the Ignition Handbook3 contains extensive materials on both the theory and practical aspects of self-heating and spontaneous combustion.

1. Residential Smoke Alarm Report, The International Fire Chief 46:9, 62-67 (Sep. 1980).

2. Los Angeles Fire Department Tests—Fire Detection Systems in Dwellings, NFPA Q. 56:3, 201-215 (Jan. 1963).

3. Babrauskas, V., Ignition Handbook, Fire Science Publishers/Society of Fire Protection Engineers, Issaquah WA (2003).

4. Babrauskas, V., Smoldering Fires, Fire Science Publishers, New York (2021).

5. Richter, F., Has Smoking Lost its Cool? Statistica (2020).

6. Nurbakhsh, S., private communication, California Bureau of Electronic Repair, Home Furnishings and Thermal Insulation (7 Feb. 2013).

7. Mehta, S., Fansler, L., Miller, D., and Garland, S., Reduced Ignition Propensity Cigarettes: Is There a Change in Smoldering Ignition Hazard? 297-307 in Fire & Materials 2013, Interscience Communications Ltd., London (2013).

8. Henriksen, T. L., Warren, C., and Lewis, K. H., Relative Humidity and Wildland Fire Ignition by Cigarettes, 860-866 in Fire & Materials 2015, Interscience Communications Ltd., London (2015).

9. Sasaki, F., Uyama, K., Uetake, M., Matsufuji, T., Matsuyama, K., Tanaike, Y., Sekizawa, A., Nagawa, Y., and Ogino, K., Experimental Study on the Effectiveness of RIP Cigarettes to Fire Ignition, 839-850 in Interflam 2013, Interscience Communications Ltd., London (2013).

10. Albers, J. C., and DeHaan, J., “Fire Safe” Cigarettes, Aren’t! California Fire Service Magazine 8-9 (Jan./Feb. 2014).

11. Larsson, I., and Bergstrand, A., Study: Reduced Ignition Propensity (RIP) Cigarettes—Theory and Reality, 235-245 in Interflam 2016, Interscience Communications Ltd, London (2016).

12. Crombie, P. E. jr, “Fire Safe” Cigarettes: Can They Still Cause Fires? 141-150 in ISFI 2006—2nd Intl. Symp. on Fire Investigations Science and Technology, NAFI, Sarasota FL (2006).

13. Frazier, P., Schaenman, P., and Jones, E., Initial Evaluation of the Effectiveness of Reduced Ignition Propensity Cigarette-Ignited Fires: Case Studies of the North American Experience, Tridata Div., System Planning Corp., Arlington VA (2011).

14. Matsuyama, K., Uyama, K., Sasaki, F., Ogino, K., Nagawa, Y., and Sekizawa, A., Experimental Study on the Effectiveness of RIP Cigarettes to Fire Situation in Japan. Part 2. Verification of Effectiveness of RIP Cigarettes in Futon and Quasi-Crack Setting, Bull. Jap. Assn. for Fire Science & Engineering 64:1, 1-7 (2014).

Vyto Babrauskas, Ph.D., is known for his research contributing to the field of fire investigation. He is the author of Ignition Handbook, Electrical Fires and Explosions, and Smoldering Fires. His firm Fire Science and Technology Inc. is located in Clarkdale, Arizona.