Contents
LC50 is the concentration of a gas, or of smoke, that kills half of a group of test animals. The value always belongs to a stated exposure time and a stated observation period afterwards. In combustion toxicology the usual form is the 30 minute LC50 for rats. For a single gas it is a volume fraction in µL/L, and for smoke a mass of burnt material per cubic metre of air.[1]
The value measures the lethal toxic potency of fire effluent. It says how little smoke is needed to kill, not how much smoke a fire makes.[1][2] A smoke with an LC50 of 10 g/m³ is twice as potent as one with an LC50 of 20 g/m³.
The quantity belongs to Fire toxicity, but it is a property of a test rather than of a material. The same product gives different values in a smouldering test, a well ventilated flaming test and a postflashover test, because the gases differ.[3][2] By the early 2000s, routine animal tests had largely given way to estimates from measured gas concentrations. The 2008 radiant furnace protocol still included a limited animal check.[4][5] Related measures are the IC50 for incapacitation and the LCt50, the product of lethal concentration and exposure time.[1][4]
Definition and units
The handbook defines the LC50 as a volume fraction of toxic gas in µL/L, or of smoke in g/m³. The value is calculated statistically from concentration-response data. It is the level that produces lethality in 50 per cent of the test animals within a specified exposure and postexposure time.[1] The IC50 is defined in the same way, but for incapacitation instead of death.[1]
A footnote to that definition fixes the two times. The specified exposure is the interval during which the animals can inhale the gas or smoke. The specified postexposure time is the period afterwards, when they breathe fresh air and are still observed.[1] A volume fraction in µL/L is numerically the same as a concentration in parts per million by volume, a unit the handbook discourages.[6]
For smoke the unit g/m³ means grams of test specimen consumed, and presumed to be present as smoke, per unit chamber volume. In a flow-through apparatus the mass is divided by the total air flow instead.[5] The SFPE Handbook prints the same quantity as mg/L of material mass loss. That is the same number, because one gram per cubic metre equals one milligram per litre.[4] Fire Dynamics Tools states the meaning plainly. The LC50 is the mass that must be dispersed into one cubic metre of air to give a 50 per cent probability of lethality, and the notation must always include the exposure time.[7]
Multiplying the LC50 by the exposure time gives the exposure dose LCt50. For smoke it is written in g·m−3·min, and for a gas in ppm·min.[4] Toxic potency is not an intrinsic characteristic of a material. It is determined per unit of specimen mass, under the conditions of one test.[7] One further convention matters when values from different papers are compared. Some studies count only animals that die during the exposure, while others include deaths up to 24 hours or up to 14 days afterwards.[8]
Exposure time, test animals and dose scaling
The standard in combustion toxicology is the LC50 for a 30 minute exposure of rats. The observation period is usually up to 14 days.[1][5] Expert consensus holds the rat to be a reasonable model for human exposure, but mainly where the effects come from asphyxiant gases such as carbon monoxide.[1][9] The endorsement is weaker for irritants, because the nose of a rodent differs greatly from that of a person.[10] Outside fire work the usual acute test is a single four hour exposure, so a four hour LC50 is not directly comparable.[11]
One approximate way to rescale between exposure times is Haber's rule. The rule assumes that the product of concentration and time is constant for a given effect.[11] It fits substances whose target organ is the lung. Carbonyl fluoride is a lung irritant released when polytetrafluoroethylene decomposes. Its one hour LC50 of 0.990 mg/L is about four times its four hour value of 0.248 mg/L.[11]
The rule fails for gases that are both taken up and excreted through the lungs. In Purser's model of a 70 kg person at rest, the carbon monoxide concentration lethal in four hours is 0.38 times the one hour value, not the 0.25 that the rule predicts.[12] Hydrogen cyanide deviates in the other direction. Incapacitation needs a dose of about 400 ppm·min at 180 µL/L, but about 2,000 ppm·min at 100 µL/L.[12]
The handbook accepts the constant-product assumption as the traditional extrapolation. For estimating a five minute incapacitating concentration for smoke-sensitive people, it prefers C²t scaling because the narcotic component dominates that estimate.[9][13] The difference is large: applying the two scaling rules to a 30 minute concentration of 30 g/m³ gives 180 g/m³ under the constant-product rule, but 73.5 g/m³ under C²t. These are illustrative extrapolations, not measured five minute LC50 values.[13][11] Incapacitation sets in well below the lethal level. Reviews of animal data put incapacitating doses at one third to one half of lethal doses, with a mean IC50/LC50 ratio of 0.50 and a standard deviation of 0.21.[14]
Calculation from measured gas concentrations
The toxic effects of the main fire gases are roughly additive. An LC50 can therefore be computed from a chemical analysis of the test atmosphere, instead of from animal deaths. The idea goes back to a 1972 proposal by Tsuchiya and Sumi for a toxicity index. Each gas contributed the ratio of its actual concentration to the concentration fatal to a person in 30 minutes.[15]
The modern form is the Fractional effective dose. It is the ratio of the Ct product for a toxicant in a test to the Ct product known to kill half of the animals.[1] Summed over gases and over time, it reads:
When the accumulated dose fraction reaches 1, the mixture is expected to be lethal to half of the exposed animals. That level is the LC50.[1] Researchers at the National Institute of Standards and Technology built the N-gas model on this result. The model proposes that the potency of many commercial products can be estimated from a small number N of gases. Tests on rats gave a mean dose fraction of 1.07 at the measured LC50, with 95 per cent confidence limits of 0.20.[14]
Two fitted equations are in use, both derived from rodent data. The first was developed empirically by Levin and co-workers, from exposures of laboratory rats to single and mixed gases.[14] It adds terms for carbon monoxide, oxygen depletion, hydrogen cyanide, hydrogen chloride and hydrogen bromide:
Here m and b are the slope and intercept of the interaction curve between carbon monoxide and carbon dioxide. That curve describes how carbon dioxide raises the toxicity of carbon monoxide. The handbook prints a branch at , with m = −18 and b = 122,000 below it, and m = 23 and b = −38,600 above it. Its statement that CO₂ is expressed in volume per cent conflicts with this branch notation and with the scale of b; the printed equation should not be evaluated without confirming the original unit convention.[14] The second equation, due to Purser, divides each gas by its own LC50 and then applies two corrections:
The factor allows for the faster breathing caused by carbon dioxide. The acidosis term is .[14] The two equations use slightly different LC50 values for hydrogen cyanide, hydrogen chloride and hydrogen bromide. They also treat carbon dioxide and oxygen differently, and they agree to about ±30 per cent near a dose fraction of 1.[14]
| Symbol | Description | Unit |
|---|---|---|
| concentration of toxic component i | µL/L or g/m³ | |
| specific dose of component i required to produce lethality | µL/L·min or g·m−3·min | |
| volume fraction of species x | µL/L for most gases; per cent for O₂ and CO₂ as described in the handbook, subject to the Eq. 5 ambiguity above | |
| volume fraction of HCN corrected for other nitriles and for the protective effect of NO₂ | µL/L | |
| 30 minute LC50 of acid gas irritant X | µL/L | |
| 30 minute LC50 of organic irritant Y | µL/L | |
| multiplication factor for carbon dioxide driven hyperventilation | – | |
| acidosis factor | – | |
| , | slope and intercept of the CO/CO₂ interaction curve | – |
The radiant furnace method of NFPA 269 and ASTM E1678 applies this directly. The 30 minute LC50 is the specimen mass loss that would give a dose fraction of 1 in one cubic metre. It follows from the mass lost in the 0.2 m³ box:[16]
The predicted value is then confirmed with a small number of rats. They are exposed 30 per cent below and 30 per cent above the estimate. Agreement means the smoke is ordinary and the calculated value stands, while disagreement signals an unusual effluent that needs further study.[5][16]
Worked example: Douglas fir in a cup furnace
Douglas fir was decomposed under flaming conditions in the National Bureau of Standards cup furnace. The measured 30 minute LCt50 for rats was 1,194 g·m−3·min. The test atmosphere contained 3,400 µL/L carbon monoxide and 3.71 per cent carbon dioxide, a carbon dioxide to carbon monoxide ratio of 11 to 1.[17] Dividing the exposure dose by the 30 minute exposure gives 1,194/30 = 39.8 g/m³ of wood consumed per cubic metre of air.
Purser's equation is then applied with a 30 minute LC50 of 5,700 µL/L for carbon monoxide.[14] The carbon monoxide term is 3,400/5,700 = 0.597. The hyperventilation factor is 1 + e(0.14 × 3.71) − 1/2 = 1.309, and the acidosis term is 3.71 × 0.05 = 0.186. The dose fraction is therefore 0.597 × 1.309 + 0.186 = 0.97, which rounds to the value of 1.0 reported for this test.[17] The measured lethality is thus explained by carbon monoxide and carbon dioxide alone.[17]
The same wood behaves differently without flames. Under nonflaming oxidative decomposition at 440 °C the measured LCt50 was 684 g·m−3·min. The measured gases accounted for a dose fraction of only 0.47.[18] About half of the lethality therefore came from products that were not measured, almost certainly organic irritants.[18] For a flexible polyurethane foam at 400 °C the unexplained share was larger still, at 0.71.[18]
Typical values
The 30 minute LC50 values for rats used in the handbook's N-gas calculations come from the International Organization for Standardization.[14] They span almost two orders of magnitude. Acrolein, hydrogen cyanide and nitrogen dioxide have the lowest values in this table. The tabulated carbon monoxide value divided by the hydrogen cyanide value is about 35; the handbook's separate prose estimate of 25 does not match these particular tabulated values.[14][19] The SFPE Handbook lists the same values, but gives a wider range of 400 to 1,400 µL/L for sulphur dioxide.[17]
| Combustion product | 30 minute LC50 (µL/L) |
|---|---|
| Carbon monoxide | 5,700 |
| Hydrogen cyanide | 165 |
| Hydrogen chloride | 3,800 |
| Hydrogen bromide | 3,800 |
| Hydrogen fluoride | 2,900 |
| Sulphur dioxide | 1,400 |
| Nitrogen dioxide | 170 |
| Acrolein | 150 |
| Formaldehyde | 750 |
For whole materials the compiled laboratory values have a mean of about 30 g/m³ for well ventilated combustion. Measured values run from about 5 to 60 g/m³, so some smokes are more than five times as potent as the average.[9] A survey of three older test methods gives a mean 30 minute LC50 of 23 mg/L of mass loss, with a range of 5 to 61 mg/L and a mean for wood of 42 mg/L.[20] These apparatus results require care in comparison: the University of Pittsburgh test reports the mass charged to the furnace rather than the true concentration of smoke in the chamber.[5][16] Treated and untreated wood products tested by that method ranged from 7 to 72 g of material charged. In Shaw's account of these tests, treated wood smoke was no more toxic than that of untreated Douglas fir.[21]
Potency depends strongly on the fire condition. Cellulosic materials have an approximate LCt50 of 730 g·m−3·min in nonflaming decomposition and 3,120 g·m−3·min in early flaming. After flashover the value falls again, to 750 g·m−3·min.[20] Rigid polyurethane foam stays between 54 and 100 g·m−3·min in all three conditions.[20]
Where nothing is known about the fuel, the handbook allows generic exposure doses. They are 900 g·m−3·min for well ventilated preflashover fires and 450 g·m−3·min for vitiated postflashover fires.[9] Over 30 minutes these correspond to concentrations of 30 and 15 g/m³. Both values are approximations that call for toxicological and engineering judgement.[9]
Use in classification and in hazard analysis
Animal LC50 values underpin several legal classifications of gases. In the United States, NFPA 55 as described in the 2008 edition of the Fire Protection Handbook calls a gas highly toxic when its LC50 in air is 200 ppm by volume or less. A gas is toxic when the value lies above 200 ppm but not above 2,000 ppm. Both limits apply to one hour of continuous inhalation by albino rats of 200 to 300 g.[22]
The 2008 handbook describes internationally accepted transport definitions that also use this measure. Division 2.3 first requires the material to be gaseous at 20 °C and 1 atm, with a boiling point no higher than 20 °C at 1 atm. It must also be known to pose a transport health hazard to humans. Where adequate human data are lacking, an animal LC50 of at most 5,000 ppm can instead establish its presumed toxicity.[23] Its poison-inhalation gas zones A to D have upper LC50 bounds of 200, 1,000, 3,000 and 5,000 ppm, respectively.[24] Hazardous materials responders use the LC50 next to the lethal dose, the permissible exposure limit, the threshold limit values and the concentration Immediately dangerous to life or health.[25]
Clean agent design uses the measure from the other end. For two halocarbon extinguishing agents in the handbook's 2008 table, the LC50 or approximate lethal concentration exceeds 80 per cent by volume: FC-3-1-10 and HFC-227ea. Their no-observed-adverse-effect levels are 40 and 9.0 per cent, respectively.[26] The design discussion uses no-observed- and lowest-observed-adverse-effect levels rather than the LC50 for normally occupied spaces. It also describes conditions under which a concentration above the no-effect level may be accepted for a limited exposure time.[26]
Regulation of building products by smoke toxicity has not followed the same path. New York State required filing of University of Pittsburgh test data for certain construction and finish materials between 1987 and 1998. Values were filed for more than 15,000 products, but no acceptance levels were ever set.[5][27] The requirement was repealed once such results were widely recognised as often irrelevant to the toxic hazard of real fires.[5] The 2008 handbook describes using toxic potency data as input to a fire hazard or risk assessment, with gas yields as the quantities carried forward.[5]
Limits and uncertainty
The strongest limitation is that smoke toxicity is not a single property of a product. What is produced depends on how much is burning, on whether the item flames or pyrolyses, on the ventilation and on the distance from the fire.[3] Testing a material in the laboratory therefore does not by itself address its toxic hazard in a fire. Heat release, mass burning rate, the other combustibles present, the enclosure, smoke transport and the occupants all matter as well.[28] In one small-room upholstered-furniture model, doubling the heat release rate cut the time to lethality by 70 per cent, while doubling the toxic potency of its smoke had a negligible effect.[29]
The second limitation is the step from rats to people. Lethality data say nothing directly about the sensory irritation that slows escape. The Irritant gas component of smoke is therefore treated separately from the asphyxiants.[30]
The handbook suggests using one fourth of the 30 minute rat LC50 as a five minute incapacitating concentration for smoke-sensitive people. Where strong acid gases are generated, one fifth is suggested instead.[13] Applied to the generic value of 30 g/m³ these rules give 7.5 and 6 g/m³, with an estimated uncertainty of about a factor of two.[13] Excitement and physical activity raise the intake of smoke-laden air, and no quantitative method exists to allow for that across a population.[13]
Human responses are also distributed rather than fixed. The very young, the elderly and people with compromised respiratory systems are more sensitive. Safe design levels must therefore lie well below levels that would incapacitate a healthy young adult.[9] At a dose fraction criterion of 0.3, about 11 per cent of the population would still be subject to incapacitation, and no criterion is statistically safe for everyone.[9]
New Zealand's Verification Method C/VM2, second edition 2025, ends available safe egress time when the first of three criteria is exceeded. They are fractional effective dose of carbon monoxide, fractional effective dose of thermal effects and a visibility limit. The two dose calculations follow ISO 13571; the carbon monoxide dose includes contributions from CO, CO₂ and O₂.[31] Van Rossum describes a Dutch design assumption of less than 30 to 60 seconds in smoke, under which toxicity is usually not the governing escape criterion. It becomes important when the route through smoke is long, or when poor visibility lengthens the exposure.[32] An LC50 is therefore best read as one input to a Smoke toxicity test based hazard calculation, not as a ranking of materials.[5][33]
See also
Notes
- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 Gann and Bryner (2008), p. 6-17.
- ↑ 2.0 2.1 Purser (2002), p. 2-91.
- ↑ 3.0 3.1 Gann and Bryner (2008), p. 6-11.
- ↑ 4.0 4.1 4.2 4.3 Purser (2002), p. 2-92.
- ↑ 5.0 5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 Gann and Bryner (2008), p. 6-19.
- ↑ Gann and Bryner (2008), p. 6-14.
- ↑ 7.0 7.1 Iqbal and Salley (2004), p. B-66.
- ↑ Purser (2002), p. 2-115.
- ↑ 9.0 9.1 9.2 9.3 9.4 9.5 9.6 Gann and Bryner (2008), p. 6-25.
- ↑ Purser (2002), p. 2-113.
- ↑ 11.0 11.1 11.2 11.3 Purser (2002), p. 2-87.
- ↑ 12.0 12.1 Purser (2002), p. 2-88.
- ↑ 13.0 13.1 13.2 13.3 13.4 Gann and Bryner (2008), p. 6-26.
- ↑ 14.00 14.01 14.02 14.03 14.04 14.05 14.06 14.07 14.08 14.09 Gann and Bryner (2008), p. 6-18.
- ↑ Gann and Bryner (2008), p. 6-17.
- ↑ 16.0 16.1 16.2 Gann and Bryner (2008), p. 6-20.
- ↑ 17.0 17.1 17.2 17.3 Purser (2002), p. 2-151.
- ↑ 18.0 18.1 18.2 Purser (2002), p. 2-152.
- ↑ Gann and Bryner (2008), p. 6-15.
- ↑ 20.0 20.1 20.2 Purser (2002), p. 2-95.
- ↑ Shaw (2008), p. 6-135.
- ↑ Lemoff (2008), p. 6-167.
- ↑ C. J. Wright (2008a), p. 13-120.
- ↑ C. J. Wright (2008a), p. 13-121.
- ↑ C. J. Wright (2008a), p. 13-134.
- ↑ 26.0 26.1 DiNenno and Taylor (2008), p. 17-102.
- ↑ Gann and Bryner (2008), p. 6-21.
- ↑ Gann and Bryner (2008), p. 6-30.
- ↑ Hirschler (2008a), p. 6-50.
- ↑ Gann and Bryner (2008), p. 6-24.
- ↑ Ministry of Business, Innovation and Employment (2025b), p. 17.
- ↑ Van Rossum (2027), §7.7.
- ↑ Purser (2002), p. 2-93.
References
- DiNenno, P. J., & Taylor, G. M. (2008). Halon and halon replacement agents and systems. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 17-93–17-121). National Fire Protection Association.
- Gann, R. G., & Bryner, N. P. (2008). Combustion products and their effects on life safety. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-11–6-34). National Fire Protection Association.
- Hirschler, M. M. (2008a). Concepts and protocols of fire testing. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-35–6-60). National Fire Protection Association.
- Iqbal, N., & Salley, M. H. (2004). Fire Dynamics Tools (FDTs): Quantitative fire hazard analysis methods for the U.S. Nuclear Regulatory Commission Fire Protection Inspection Program (NUREG-1805). U.S. Nuclear Regulatory Commission, Office of Nuclear Reactor Regulation.
- Lemoff, T. C. (2008). Gases. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-165–6-186). National Fire Protection Association.
- Ministry of Business, Innovation and Employment. (2025b). Verification Method C/VM2: Framework for fire safety design (2nd ed.).
- Purser, D. A. (2002). Toxicity assessment of combustion products. In P. J. DiNenno, D. Drysdale, C. L. Beyler, W. D. Walton, R. L. P. Custer, J. R. Hall Jr., & J. M. Watts Jr. (Eds.), SFPE handbook of fire protection engineering (3rd ed., pp. 2-83–2-171). National Fire Protection Association; Society of Fire Protection Engineers.
- Shaw, J. R. (2008). Fire-Retardant and flame-resistant treatment of cellulosic materials. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-129–6-139). National Fire Protection Association.
- Van Rossum, E. (2027). Introduction to Fire Safety Engineering. Fire Safety Academy.
- Wright, C. J. (2008a). Managing the response to hazardous material incidents. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 13-117–13-144). National Fire Protection Association.