New FSE wiki article: Smoke toxicity test (written by Claude Opus 5.5) |
Remove bibliographic source lines from image captions and descriptions |
||
| (2 intermediate revisions by 2 users not shown) | |||
| Line 1: | Line 1: | ||
A '''smoke toxicity test''' is a laboratory method that burns a small specimen under controlled conditions and measures how poisonous the resulting smoke is. The result is a toxic potency, usually the mass of specimen per cubic metre of air that would kill half of a group of exposed rats in 30 minutes.<ref name="GB-19">Gann and Bryner (2008), p. 6-19.</ref> Early methods counted animal deaths. | A '''smoke toxicity test''' is a laboratory method that burns a small specimen under controlled conditions and measures how poisonous the resulting smoke is. The result is a toxic potency, usually the mass of specimen per cubic metre of air that would kill half of a group of exposed rats in 30 minutes.<ref name="GB-19">Gann and Bryner (2008), p. 6-19.</ref> Early methods counted animal deaths. By 2008, the radiant furnace method estimated that value from key gas concentrations, with a limited animal check.<ref name="GB-19" /> | ||
Such tests answer one narrow question within [[Fire toxicity]]: how potent is the effluent of this material, under this one set of decomposition conditions.<ref name="PU-145">Purser (2002), p. 2-145.</ref> They cannot say how hazardous the material is in a real fire, because that also depends on how fast it ignites, spreads flame and releases heat.<ref name="GB-30">Gann and Bryner (2008), p. 6-30.</ref><ref name="PU-145" /> Test results are therefore used as input to a fire hazard or risk assessment rather than as pass or fail criteria for products.<ref name="GB-19" /> | Such tests answer one narrow question within [[Fire toxicity]]: how potent is the effluent of this material, under this one set of decomposition conditions.<ref name="PU-145">Purser (2002), p. 2-145.</ref> They cannot say how hazardous the material is in a real fire, because that also depends on how fast it ignites, spreads flame and releases heat.<ref name="GB-30">Gann and Bryner (2008), p. 6-30.</ref><ref name="PU-145" /> Test results are therefore used as input to a fire hazard or risk assessment rather than as pass or fail criteria for products.<ref name="GB-19" /> | ||
The apparatus in | The apparatus reviewed in the 2008 handbook differ mainly in two ways. The specimen may be heated by a hot cup, by a radiant panel or by a travelling furnace, and the smoke may be held in a closed box or flow past the animals. The methods described include the cup furnace, the radiant furnace of NFPA 269 and ASTM E1678, and the University of Pittsburgh apparatus. The handbook also describes the closed cabinet of NFPA 270 and ISO 5659-2, and tube furnaces such as DIN 53436 and ISO/TS 19700.<ref name="GB-19" /><ref name="GB-22">Gann and Bryner (2008), p. 6-22.</ref> | ||
== What the test measures == | == What the test measures == | ||
| Line 16: | Line 16: | ||
The ventilation of the combustion is described by the [[Equivalence ratio]], the fuel to air ratio divided by its value for complete combustion.<ref name="GB-13">Gann and Bryner (2008), p. 6-13.</ref> Values below 1 mean a fuel-lean, well ventilated fire and values above 1 a fuel-rich, ventilation-controlled fire.<ref name="GB-13" /> The yield of carbon monoxide is very low below about 0.5, then rises steeply, and settles at roughly 0.2 kg per kg of fuel consumed once the ratio passes about 1.5.<ref name="GB-13" /> Because carbon monoxide dominates the lethality of most smoke, a test that burns the specimen too cleanly will understate the potency of a postflashover fire.<ref name="GB-13" /><ref name="PU-153">Purser (2002), p. 2-153.</ref> | The ventilation of the combustion is described by the [[Equivalence ratio]], the fuel to air ratio divided by its value for complete combustion.<ref name="GB-13">Gann and Bryner (2008), p. 6-13.</ref> Values below 1 mean a fuel-lean, well ventilated fire and values above 1 a fuel-rich, ventilation-controlled fire.<ref name="GB-13" /> The yield of carbon monoxide is very low below about 0.5, then rises steeply, and settles at roughly 0.2 kg per kg of fuel consumed once the ratio passes about 1.5.<ref name="GB-13" /> Because carbon monoxide dominates the lethality of most smoke, a test that burns the specimen too cleanly will understate the potency of a postflashover fire.<ref name="GB-13" /><ref name="PU-153">Purser (2002), p. 2-153.</ref> | ||
[[File:fse-smoke-toxicity-test-v3.svg|thumb|upright=1.5|alt=Line chart of carbon monoxide yield against equivalence ratio, low below 0.5 and constant near 0.2 kilograms per kilogram above 1.5|Schematic carbon monoxide yield against equivalence ratio, with ISO/TS 19700 temperatures listed separately.]] | |||
ISO 16312-1 lists what an ideal method should do. The combustor should replicate one or more fire stages, and the specimens should represent products as used. The apparatus itself should not affect the results, and the method should be easy to run. It should give gas and smoke yields together with specimen mass loss, and it should have demonstrated accuracy against real-scale fires. Finally, the results should be repeatable and reproducible.<ref name="GB-19" /> | ISO 16312-1, as described in the 2008 handbook, lists what an ideal method should do. The combustor should replicate one or more fire stages, and the specimens should represent products as used. The apparatus itself should not affect the results, and the method should be easy to run. It should give gas and smoke yields together with specimen mass loss, and it should have demonstrated accuracy against real-scale fires. Finally, the results should be repeatable and reproducible.<ref name="GB-19" /> | ||
The criterion on specimens is the hardest to meet. Several methods accept only a ground or crumpled specimen, which represents a layered product poorly, and the two forms of specimen can give different results.<ref name="GB-19" /> ISO/TR 16312-2 | The criterion on specimens is the hardest to meet. Several methods accept only a ground or crumpled specimen, which represents a layered product poorly, and the two forms of specimen can give different results.<ref name="GB-19" /> ISO/TR 16312-2 applied these criteria to twelve methods available when the handbook was written. It is a technical report rather than a standard, and it sets out the appraisal of the experts assigned to the responsible ISO subcommittee.<ref name="GB-19" /> | ||
Alongside the gases, a test should record several physical quantities. These are the temperature or radiant flux, the air to fuel ratio, the smoke [[Optical density]] and whether the specimen flamed. The minimum chemical set is carbon monoxide, carbon dioxide with their ratio, and oxygen. Hydrogen cyanide or the hydrogen halides are added where the material contains nitrogen or halogens.<ref name="PU-144" /> | Alongside the gases, a test should record several physical quantities. These are the temperature or radiant flux, the air to fuel ratio, the smoke [[Optical density]] and whether the specimen flamed. The minimum chemical set is carbon monoxide, carbon dioxide with their ratio, and oxygen. Hydrogen cyanide or the hydrogen halides are added where the material contains nitrogen or halogens.<ref name="PU-144" /> | ||
| Line 25: | Line 26: | ||
== Bench-scale apparatus == | == Bench-scale apparatus == | ||
The cup furnace burns the specimen in a quartz beaker set into the floor of a closed box of 0.2 m³. Six rats are held in tubular restrainers with their heads inside the box, inhale the smoke for 30 minutes and are then observed for 14 days.<ref name="GB-19" /> Flaming and nonflaming data come from tests just above and just below the autoignition temperature of the specimen.<ref name="GB-19" /> Repeating the test with different specimen masses varies the smoke concentration, and the deaths per test are plotted against that concentration to give the LC50.<ref name="GB-20">Gann and Bryner (2008), p. 6-20.</ref> The method models smouldering and early flaming reasonably well, but not a fully developed fire, because pyrolysis products escape from the cup before they decompose fully.<ref name="PU-147">Purser (2002), p. 2-147.</ref> | The cup furnace burns the specimen in a quartz beaker set into the floor of a closed box of 0.2 m³. Six rats are held in tubular restrainers with their heads inside the box, inhale the smoke for 30 minutes and are then observed for 14 days.<ref name="GB-19" /> Flaming and nonflaming data come from tests just above and just below the autoignition temperature of the specimen.<ref name="GB-19" /> Repeating the test with different specimen masses varies the smoke concentration, and the deaths per test are plotted against that concentration to give the LC50.<ref name="GB-20">Gann and Bryner (2008), p. 6-20.</ref> The method models smouldering and early flaming reasonably well, but not a fully developed fire, because pyrolysis products escape from the cup before they decompose fully.<ref name="PU-147">Purser (2002), p. 2-147.</ref> | ||
[[File:fse-smoke-toxicity-test-v1.svg|thumb|upright=1.5|alt=Two schematic cross sections, a closed box with a cup furnace and head-only animal ports, and a flow-through tube furnace with diluent air and an exposure chamber|Closed-box and flow-through smoke toxicity apparatus.]] | |||
The radiant furnace test replaced it. The same 0.2 m³ box is used, but the cup is replaced by a radiant heater, and a spark source ignites the vapours, which suits layered products such as furniture or wall coverings.<ref name="GB-20" /> A specimen of up to 10 cm by 15 cm by 5 cm is exposed to 50 kW/m² for 15 minutes. The smoke enters the box through the centre slot of a three-slot chimney, while box air recirculates to the combustion zone through the outer slots.<ref name="GB-20" /> Successive tests vary the exposed top surface area rather than the heating. The | The radiant furnace test had replaced it by 2008. The same 0.2 m³ box is used, but the cup is replaced by a radiant heater, and a spark source ignites the vapours, which suits layered products such as furniture or wall coverings.<ref name="GB-20" /> A specimen of up to 10 cm by 15 cm by 5 cm is exposed to 50 kW/m² for 15 minutes. The smoke enters the box through the centre slot of a three-slot chimney, while box air recirculates to the combustion zone through the outer slots.<ref name="GB-20" /> | ||
Successive tests vary the exposed top surface area rather than the heating. The 2008 handbook describes ASTM E1678 and NFPA 269 as identical standardisations of the apparatus and procedure. Both included a mathematical adjustment for oxygen-vitiated postflashover fires.<ref name="GB-20" /> | |||
The University of Pittsburgh apparatus is a flow-through system in which four mice breathe the effluent as it is produced. The specimen is heated at a rising temperature of 20 °C per minute. The 30 minute exposure starts when 1 per cent of the sample weight has been lost, and 10 minutes of observation follow.<ref name="GB-20" /><ref name="PU-147">Purser (2002), p. 2-147.</ref> Results are reported as LC50 values, but they are weights of specimen charged to the furnace rather than concentrations, so they cannot be compared directly with values from other methods.<ref name="GB-19" /><ref name="GB-20" /> There is no ignition source, so a sample may or may not autoignite, and it is unclear which stage of a fire the apparatus represents.<ref name="GB-22" /> | The University of Pittsburgh apparatus is a flow-through system in which four mice breathe the effluent as it is produced. The specimen is heated at a rising temperature of 20 °C per minute. The 30 minute exposure starts when 1 per cent of the sample weight has been lost, and 10 minutes of observation follow.<ref name="GB-20" /><ref name="PU-147">Purser (2002), p. 2-147.</ref> Results are reported as LC50 values, but they are weights of specimen charged to the furnace rather than concentrations, so they cannot be compared directly with values from other methods.<ref name="GB-19" /><ref name="GB-20" /> There is no ignition source, so a sample may or may not autoignite, and it is unclear which stage of a fire the apparatus represents.<ref name="GB-22" /> | ||
The closed cabinet test of NFPA 270 and ISO 5659-2 was designed for smoke optical density. With added gas measurements | The closed cabinet test of NFPA 270 and ISO 5659-2 was designed for smoke optical density. With added gas measurements, the 2008 handbook describes its use as an acceptance test based on gas concentrations.<ref name="GB-22" /> A planar specimen of 7.5 cm by 7.5 cm by up to 2.5 cm is exposed to 25 kW/m² or 50 kW/m² for 10 minutes. The products accumulate in a chamber of 0.5 m³ and are sampled at its geometric centre.<ref name="GB-22" /> If any specified gas exceeds a prescribed limit, the product fails.<ref name="GB-22" /> | ||
Tube furnaces decompose a strip of material continuously. In DIN 53436 a cylindrical furnace travels along a quartz tube holding about 40 mL of sample, at a constant temperature between 200 °C and 1000 °C. Decomposition takes place against an air stream, so effluent of constant composition flows to a chamber holding ten rats.<ref name="GB-22" /> Varying the air flow varies the carbon dioxide to carbon monoxide ratio, so both well ventilated and ventilation-limited fires can be simulated.<ref name="GB-22" /> | Tube furnaces decompose a strip of material continuously. In DIN 53436 a cylindrical furnace travels along a quartz tube holding about 40 mL of sample, at a constant temperature between 200 °C and 1000 °C. Decomposition takes place against an air stream, so effluent of constant composition flows to a chamber holding ten rats.<ref name="GB-22" /> Varying the air flow varies the carbon dioxide to carbon monoxide ratio, so both well ventilated and ventilation-limited fires can be simulated.<ref name="GB-22" /> | ||
The furnace of ISO/TS 19700 is wider, at 4.3 cm, and drives a sample boat at least 0.8 m long. | The furnace of ISO/TS 19700 is wider, at 4.3 cm, and drives a sample boat at least 0.8 m long. In the handbook's 2008 description, it is held at 350 °C for oxidative pyrolysis, 650 °C for well ventilated flaming or 825 °C for postflashover combustion.<ref name="GB-22" /> A fixed sample speed with variable air flow lets the equivalence ratio be set in advance. The handbook describes the apparatus as identical to the tube furnace of BS 7990.<ref name="GB-22" /><ref name="GB-23">Gann and Bryner (2008), p. 6-23.</ref> | ||
The cone calorimeter has also been tried, since it already measures heat release, ignitability and smoke. It has not proved suitable for toxic potency, because it reproduces only very well ventilated conditions. The products are also heavily diluted, which makes the toxic species hard to measure.<ref name="PU-148">Purser (2002), p. 2-148.</ref> | The cone calorimeter has also been tried, since it already measures heat release, ignitability and smoke. It has not proved suitable for toxic potency, because it reproduces only very well ventilated conditions. The products are also heavily diluted, which makes the toxic species hard to measure.<ref name="PU-148">Purser (2002), p. 2-148.</ref> | ||
| Line 57: | Line 61: | ||
The protocol of NFPA 269 and ASTM E1678 then limits the use of animals to a check. Rats are exposed at 30 per cent below and 30 per cent above the calculated value. If the mortality matches the prediction, the smoke is ordinary and the calculated value is used; if the smoke proves unusually toxic, further research or testing is needed.<ref name="GB-19" /> A comparison against room-scale postflashover fires was made after adjusting the carbon monoxide yield to about 0.2 kg per kg of fuel. Reproducing the low oxygen of such a fire would otherwise kill the animals through oxygen depletion alone.<ref name="GB-20" /> On five different criteria the bench-scale results were consistent with the room-scale data, and accuracy to within a factor of three was found to be both attainable and practical.<ref name="GB-20" /> | The protocol of NFPA 269 and ASTM E1678 then limits the use of animals to a check. Rats are exposed at 30 per cent below and 30 per cent above the calculated value. If the mortality matches the prediction, the smoke is ordinary and the calculated value is used; if the smoke proves unusually toxic, further research or testing is needed.<ref name="GB-19" /> A comparison against room-scale postflashover fires was made after adjusting the carbon monoxide yield to about 0.2 kg per kg of fuel. Reproducing the low oxygen of such a fire would otherwise kill the animals through oxygen depletion alone.<ref name="GB-20" /> On five different criteria the bench-scale results were consistent with the room-scale data, and accuracy to within a factor of three was found to be both attainable and practical.<ref name="GB-20" /> | ||
[[File:fse-smoke-toxicity-test-v2.svg|thumb|upright=1.5|alt=Flow chart from specimen mass loss and gas analysis through the dose fraction equation to a calculated LC50 with an animal confirmation step|The NFPA 269 and ASTM E1678 sequence: the LC50 is calculated from gas concentrations and only confirmed with animals.]] | |||
Analytical data from the DIN furnace have been used in the same way. Calculated and measured values for more than 30 materials correlated fairly well, which supports the view that lethal potency can largely be attributed to a small number of gases.<ref name="GB-22" /> What the calculation misses is sensory irritation. The effect of an [[Irritant gas]] on escape is concentration related rather than dose related, and a lethality test records neither.<ref name="PU-144" /> In nonflaming tests of wood and of flexible polyurethane foam, only 0.47 and 0.29 of the measured lethality could be explained by the measured gases, the remainder being attributed to organic irritants.<ref name="PU-152">Purser (2002), p. 2-152.</ref> | Analytical data from the DIN furnace have been used in the same way. Calculated and measured values for more than 30 materials correlated fairly well, which supports the view that lethal potency can largely be attributed to a small number of gases.<ref name="GB-22" /> What the calculation misses is sensory irritation. The effect of an [[Irritant gas]] on escape is concentration related rather than dose related, and a lethality test records neither.<ref name="PU-144" /> In nonflaming tests of wood and of flexible polyurethane foam, only 0.47 and 0.29 of the measured lethality could be explained by the measured gases, the remainder being attributed to organic irritants.<ref name="PU-152">Purser (2002), p. 2-152.</ref> | ||
=== Worked example: ranking two floor coverings === | === Worked example: ranking two floor coverings === | ||
Purser's theoretical example compares two hotel bedroom floor coverings. The room is closed and has a volume of 40 m³. Both coverings are 1 cm thick with a stated area density of 1 kg/m². Both burn through, so a front of complete combustion spreads out from the ignition point.<ref name="PU-98">Purser (2002), p. 2-98.</ref> Material A spreads flame at 10 cm/min and material B at 5 cm/min.<ref name="PU-98" /> | |||
Bench-scale tests under well ventilated flaming conditions gave material A a 30 minute LC50 of 20 g/m³ and material B a value of 10 g/m³.<ref name="PU-98" /> Multiplying by the 30 minute exposure converts these into lethal exposure doses of 600 and 300 g·m<sup>−3</sup>·min, so the smoke of material B is twice as potent.<ref name="PU-98" /> Burning area grows as the square of the radius, so the area of A consumed at any moment is (10/5)² = 4 times that of B.<ref name="PU-99">Purser (2002), p. 2-99.</ref> | Bench-scale tests under well ventilated flaming conditions gave material A a 30 minute LC50 of 20 g/m³ and material B a value of 10 g/m³.<ref name="PU-98" /> Multiplying by the 30 minute exposure converts these into lethal exposure doses of 600 and 300 g·m<sup>−3</sup>·min, so the smoke of material B is twice as potent.<ref name="PU-98" /> Burning area grows as the square of the radius, so the area of A consumed at any moment is (10/5)² = 4 times that of B.<ref name="PU-99">Purser (2002), p. 2-99.</ref> | ||
The published table gives material A 7,854 cm² burnt after five minutes, but also 7,854 g consumed.<ref name="PU-99" /> At the stated 1 kg/m², that area contains only about 785 g. The table's mass, smoke concentration and accumulated dose are therefore ten times too high unless the intended area density was 10 kg/m². Its five minute dose fractions of 0.548 for A and 0.274 for B inherit this discrepancy.<ref name="PU-99" /> | |||
At the stated area density and with even mixing, A's five minute exposure dose is about 32.7 g·m<sup>−3</sup>·min. B's is 8.18 g·m<sup>−3</sup>·min. Their dose fractions are 0.0545 and 0.0273, calculated from the stated mass density and lethal doses.<ref name="PU-98" /><ref name="PU-99" /> | |||
For a circular burn front at constant speed, the dose follows <math>D(t)=\sigma\pi v^2t^3/(3V)</math>, where <math>\sigma</math> is area density, <math>v</math> flame-spread speed and <math>V</math> room volume. On these assumptions the dose fraction reaches 1 at about 13.2 minutes for A and 16.6 minutes for B. Purser instead reports about 6 and 7.5 minutes, consistent with the higher masses in the table.<ref name="PU-99" /> The qualitative ranking is the same: B's smoke is twice as potent, but B burns over one fourth the area at any given time. | |||
[[File:fse-smoke-toxicity-test-v4.svg|thumb|upright=1.5|alt=Two calculated dose-fraction curves from the stated area density; material A crosses one at 13.2 minutes and material B at 16.6 minutes|Corrected dose-fraction calculation for two floor coverings; the source table has a factor-ten mass discrepancy.]] | |||
== Use in regulation and in hazard assessment == | == Use in regulation and in hazard assessment == | ||
Regulation by smoke toxicity has been limited. New York State required testing and filing of University of Pittsburgh data for certain construction and finish materials from 1987 to 1998. Values were filed for more than 15,000 products, but no acceptance level was ever set.<ref name="GB-19" /><ref name="GB-21">Gann and Bryner (2008), p. 6-21.</ref> The requirement was repealed when the results were widely recognised as often irrelevant to real fire hazards, and most laboratory testing stopped.<ref name="GB-19" /> By the 2008 edition of the ''Fire Protection Handbook'' no laboratory was running nonproprietary animal tests for smoke toxicity on a routine basis.<ref name="GB-19" /> | Regulation by smoke toxicity has been limited. New York State required testing and filing of University of Pittsburgh data for certain construction and finish materials from 1987 to 1998. Values were filed for more than 15,000 products, but no acceptance level was ever set.<ref name="GB-19" /><ref name="GB-21">Gann and Bryner (2008), p. 6-21.</ref> The requirement was repealed when the results were widely recognised as often irrelevant to real fire hazards, and most laboratory testing stopped.<ref name="GB-19" /> By the 2008 edition of the ''Fire Protection Handbook'' no laboratory was running nonproprietary animal tests for smoke toxicity on a routine basis.<ref name="GB-19" /> | ||
The 2008 handbook describes gas-concentration limits in transport. It says the closed cabinet method was used by the International Maritime Organization and the European Union as an acceptance test on measured gases.<ref name="GB-22" /> The IMO Fire Test Procedures Code was adopted in 1998 and became mandatory for international service vessels built after December 2003.<ref name="EB-161">Eberly (2008), p. 21-161.</ref><ref name="EB-171">Eberly (2008), p. 21-171.</ref> Its Part 2 is a smoke and toxicity test, and interior finishes had to pass testing to ISO 5659 under that code.<ref name="EB-174">Eberly (2008), p. 21-174.</ref> Hirschler describes the same apparatus under the designation ASTM E1995 in 2008 ship interior-finish rules of the United States Coast Guard and NFPA 301.<ref name="HI-33">Hirschler (2008b), p. 18-33.</ref> | |||
The 2008 handbook describes NFPA 269 as the United States method developed to generate data for hazard modelling, by the NFPA Technical Committee on Fire Tests.<ref name="BC-230">Benedetti and Colonna (2008), p. 6-230.</ref><ref name="HI-38">Hirschler (2008a), p. 6-38.</ref> Measurement of the gases themselves is covered by ASTM E800. Madrzykowski and Stroup described Fourier transform infrared spectroscopy as the preferred technique for continuous analysis of fire gases in 2008.<ref name="MS-45">Madrzykowski and Stroup (2008), p. 2-45.</ref> | |||
== Limits of the method == | == Limits of the method == | ||
| Line 79: | Line 89: | ||
The second limit is the fire model. In the closed cabinet the air becomes vitiated during the test and the generation rates change, so it is unclear which fire stage is replicated. No comparisons with real-scale tests of the same products have been reported.<ref name="GB-22" /> The smoke density chamber family has further known shortcomings, among them a static chamber while real smoke transfer is dynamic, oxygen-deficient atmospheres and soot deposits on walls and optics.<ref name="HI-33" /> Ramped heating in the University of Pittsburgh method decomposes the material fractionally, so products given off early leave the furnace before the higher temperatures are reached.<ref name="PU-147" /> | The second limit is the fire model. In the closed cabinet the air becomes vitiated during the test and the generation rates change, so it is unclear which fire stage is replicated. No comparisons with real-scale tests of the same products have been reported.<ref name="GB-22" /> The smoke density chamber family has further known shortcomings, among them a static chamber while real smoke transfer is dynamic, oxygen-deficient atmospheres and soot deposits on walls and optics.<ref name="HI-33" /> Ramped heating in the University of Pittsburgh method decomposes the material fractionally, so products given off early leave the furnace before the higher temperatures are reached.<ref name="PU-147" /> | ||
The third limit is relevance. | The third limit is relevance. Hirschler's 2008 review associates most fatalities with fires extending beyond the room of origin and with carbon monoxide production.<ref name="HI-50">Hirschler (2008a), p. 6-50.</ref> Gann and Bryner put the carbon monoxide yield at about 0.2 kg/kg fuel once the fuel-to-air equivalence ratio exceeds roughly 1.5; this is a specified ventilation condition, not a fixed yield for every fire.<ref name="GB-13" /> Heat release rate strongly affects the time available to escape.<ref name="HI-46">Hirschler (2008a), p. 6-46.</ref> In one small-room model, doubling the heat release rate of an upholstered item cut time to lethality by 70 per cent; doubling smoke toxic potency had little effect.<ref name="HI-50" /> | ||
Regulations that cap the concentration of one or two gases measured in a particular apparatus are therefore unlikely to deliver the intended level of safety. Gas yields are not singular properties of a product.<ref name="GB-19" /> The value of such a test lies in the potency and yield data it supplies to a hazard calculation. Those data sit next to measurements of ignition, flame spread and heat release from other tests.<ref name="GB-19" /><ref name="PU-145" /> | Regulations that cap the concentration of one or two gases measured in a particular apparatus are therefore unlikely to deliver the intended level of safety. Gas yields are not singular properties of a product.<ref name="GB-19" /> The value of such a test lies in the potency and yield data it supplies to a hazard calculation. Those data sit next to measurements of ignition, flame spread and heat release from other tests.<ref name="GB-19" /><ref name="PU-145" /> | ||
Latest revision as of 08:25, 8 November 2026
A smoke toxicity test is a laboratory method that burns a small specimen under controlled conditions and measures how poisonous the resulting smoke is. The result is a toxic potency, usually the mass of specimen per cubic metre of air that would kill half of a group of exposed rats in 30 minutes.[1] Early methods counted animal deaths. By 2008, the radiant furnace method estimated that value from key gas concentrations, with a limited animal check.[1]
Such tests answer one narrow question within Fire toxicity: how potent is the effluent of this material, under this one set of decomposition conditions.[2] They cannot say how hazardous the material is in a real fire, because that also depends on how fast it ignites, spreads flame and releases heat.[3][2] Test results are therefore used as input to a fire hazard or risk assessment rather than as pass or fail criteria for products.[1]
The apparatus reviewed in the 2008 handbook differ mainly in two ways. The specimen may be heated by a hot cup, by a radiant panel or by a travelling furnace, and the smoke may be held in a closed box or flow past the animals. The methods described include the cup furnace, the radiant furnace of NFPA 269 and ASTM E1678, and the University of Pittsburgh apparatus. The handbook also describes the closed cabinet of NFPA 270 and ISO 5659-2, and tube furnaces such as DIN 53436 and ISO/TS 19700.[1][4]
What the test measures
The measured quantity is the lethal toxic potency of the effluent. It is reported as the 30 minute LC50, in grams of test specimen consumed per cubic metre of chamber volume for a closed box, or per unit of total air flow for a flow-through apparatus.[1] Multiplying by the exposure time gives the lethal exposure dose LCt50 in g·m−3·min. That form can be compared between protocols and used with mass loss data from full-scale fires or models.[2]
The result is not a constant of the material. Toxic potency relates concentration and exposure time to a degree of adverse effect, and it depends on the composition of the smoke, which in turn depends on the conditions under which the smoke was generated.[5] Toxic hazard depends further on the fire scenario and on how the product is used.[5] A test value therefore has to be quoted together with the decomposition condition that produced it.[6]
Modern methods also report gas yields, that is the mass of each gas divided by the combusted mass of specimen. Less commonly the mass subjected to the test conditions is used instead, and the two agree only when the specimen burns completely without residue.[1] Yields are the quantities carried forward into hazard calculations, because they can be combined with a predicted mass loss rate for a full-scale fire.[1][2]
Fire conditions a test must reproduce
Thermal decomposition products depend on the temperature, on the oxygen supply and on whether the specimen flames. Three conditions matter: nonflaming oxidative or smouldering decomposition at mid-range temperatures, flaming decomposition at mid-range temperatures, and the high-temperature, low-oxygen state of a fully developed fire.[6] A method is useful only if it can be shown to model one or more of these.[6]
The ventilation of the combustion is described by the Equivalence ratio, the fuel to air ratio divided by its value for complete combustion.[7] Values below 1 mean a fuel-lean, well ventilated fire and values above 1 a fuel-rich, ventilation-controlled fire.[7] The yield of carbon monoxide is very low below about 0.5, then rises steeply, and settles at roughly 0.2 kg per kg of fuel consumed once the ratio passes about 1.5.[7] Because carbon monoxide dominates the lethality of most smoke, a test that burns the specimen too cleanly will understate the potency of a postflashover fire.[7][8]
ISO 16312-1, as described in the 2008 handbook, lists what an ideal method should do. The combustor should replicate one or more fire stages, and the specimens should represent products as used. The apparatus itself should not affect the results, and the method should be easy to run. It should give gas and smoke yields together with specimen mass loss, and it should have demonstrated accuracy against real-scale fires. Finally, the results should be repeatable and reproducible.[1]
The criterion on specimens is the hardest to meet. Several methods accept only a ground or crumpled specimen, which represents a layered product poorly, and the two forms of specimen can give different results.[1] ISO/TR 16312-2 applied these criteria to twelve methods available when the handbook was written. It is a technical report rather than a standard, and it sets out the appraisal of the experts assigned to the responsible ISO subcommittee.[1]
Alongside the gases, a test should record several physical quantities. These are the temperature or radiant flux, the air to fuel ratio, the smoke Optical density and whether the specimen flamed. The minimum chemical set is carbon monoxide, carbon dioxide with their ratio, and oxygen. Hydrogen cyanide or the hydrogen halides are added where the material contains nitrogen or halogens.[6]
Bench-scale apparatus
The cup furnace burns the specimen in a quartz beaker set into the floor of a closed box of 0.2 m³. Six rats are held in tubular restrainers with their heads inside the box, inhale the smoke for 30 minutes and are then observed for 14 days.[1] Flaming and nonflaming data come from tests just above and just below the autoignition temperature of the specimen.[1] Repeating the test with different specimen masses varies the smoke concentration, and the deaths per test are plotted against that concentration to give the LC50.[9] The method models smouldering and early flaming reasonably well, but not a fully developed fire, because pyrolysis products escape from the cup before they decompose fully.[10]
The radiant furnace test had replaced it by 2008. The same 0.2 m³ box is used, but the cup is replaced by a radiant heater, and a spark source ignites the vapours, which suits layered products such as furniture or wall coverings.[9] A specimen of up to 10 cm by 15 cm by 5 cm is exposed to 50 kW/m² for 15 minutes. The smoke enters the box through the centre slot of a three-slot chimney, while box air recirculates to the combustion zone through the outer slots.[9]
Successive tests vary the exposed top surface area rather than the heating. The 2008 handbook describes ASTM E1678 and NFPA 269 as identical standardisations of the apparatus and procedure. Both included a mathematical adjustment for oxygen-vitiated postflashover fires.[9]
The University of Pittsburgh apparatus is a flow-through system in which four mice breathe the effluent as it is produced. The specimen is heated at a rising temperature of 20 °C per minute. The 30 minute exposure starts when 1 per cent of the sample weight has been lost, and 10 minutes of observation follow.[9][10] Results are reported as LC50 values, but they are weights of specimen charged to the furnace rather than concentrations, so they cannot be compared directly with values from other methods.[1][9] There is no ignition source, so a sample may or may not autoignite, and it is unclear which stage of a fire the apparatus represents.[4]
The closed cabinet test of NFPA 270 and ISO 5659-2 was designed for smoke optical density. With added gas measurements, the 2008 handbook describes its use as an acceptance test based on gas concentrations.[4] A planar specimen of 7.5 cm by 7.5 cm by up to 2.5 cm is exposed to 25 kW/m² or 50 kW/m² for 10 minutes. The products accumulate in a chamber of 0.5 m³ and are sampled at its geometric centre.[4] If any specified gas exceeds a prescribed limit, the product fails.[4]
Tube furnaces decompose a strip of material continuously. In DIN 53436 a cylindrical furnace travels along a quartz tube holding about 40 mL of sample, at a constant temperature between 200 °C and 1000 °C. Decomposition takes place against an air stream, so effluent of constant composition flows to a chamber holding ten rats.[4] Varying the air flow varies the carbon dioxide to carbon monoxide ratio, so both well ventilated and ventilation-limited fires can be simulated.[4]
The furnace of ISO/TS 19700 is wider, at 4.3 cm, and drives a sample boat at least 0.8 m long. In the handbook's 2008 description, it is held at 350 °C for oxidative pyrolysis, 650 °C for well ventilated flaming or 825 °C for postflashover combustion.[4] A fixed sample speed with variable air flow lets the equivalence ratio be set in advance. The handbook describes the apparatus as identical to the tube furnace of BS 7990.[4][11]
The cone calorimeter has also been tried, since it already measures heat release, ignitability and smoke. It has not proved suitable for toxic potency, because it reproduces only very well ventilated conditions. The products are also heavily diluted, which makes the toxic species hard to measure.[12]
From animal exposure to chemical analysis
Since the toxic effects of the main fire gases are additive, the lethal concentration can be calculated from an analysis of the test atmosphere. The calculation uses the Fractional effective dose, the ratio of the dose of each gas present to the dose of that gas known to be lethal, summed over the gases and over time.[13] In the radiant furnace method the LC50 is the specimen mass loss that would give a dose fraction of 1 in one cubic metre:[9]
| Symbol | Description | Unit |
|---|---|---|
| mass lost by the specimen during the test | g | |
| fractional effective dose calculated from the measured gas concentrations | – | |
| volume of the exposure chamber, 0.2 m³ in this apparatus | m³ | |
| 30 minute lethal concentration of the smoke | g/m³ |
The protocol of NFPA 269 and ASTM E1678 then limits the use of animals to a check. Rats are exposed at 30 per cent below and 30 per cent above the calculated value. If the mortality matches the prediction, the smoke is ordinary and the calculated value is used; if the smoke proves unusually toxic, further research or testing is needed.[1] A comparison against room-scale postflashover fires was made after adjusting the carbon monoxide yield to about 0.2 kg per kg of fuel. Reproducing the low oxygen of such a fire would otherwise kill the animals through oxygen depletion alone.[9] On five different criteria the bench-scale results were consistent with the room-scale data, and accuracy to within a factor of three was found to be both attainable and practical.[9]
Analytical data from the DIN furnace have been used in the same way. Calculated and measured values for more than 30 materials correlated fairly well, which supports the view that lethal potency can largely be attributed to a small number of gases.[4] What the calculation misses is sensory irritation. The effect of an Irritant gas on escape is concentration related rather than dose related, and a lethality test records neither.[6] In nonflaming tests of wood and of flexible polyurethane foam, only 0.47 and 0.29 of the measured lethality could be explained by the measured gases, the remainder being attributed to organic irritants.[14]
Worked example: ranking two floor coverings
Purser's theoretical example compares two hotel bedroom floor coverings. The room is closed and has a volume of 40 m³. Both coverings are 1 cm thick with a stated area density of 1 kg/m². Both burn through, so a front of complete combustion spreads out from the ignition point.[15] Material A spreads flame at 10 cm/min and material B at 5 cm/min.[15]
Bench-scale tests under well ventilated flaming conditions gave material A a 30 minute LC50 of 20 g/m³ and material B a value of 10 g/m³.[15] Multiplying by the 30 minute exposure converts these into lethal exposure doses of 600 and 300 g·m−3·min, so the smoke of material B is twice as potent.[15] Burning area grows as the square of the radius, so the area of A consumed at any moment is (10/5)² = 4 times that of B.[16]
The published table gives material A 7,854 cm² burnt after five minutes, but also 7,854 g consumed.[16] At the stated 1 kg/m², that area contains only about 785 g. The table's mass, smoke concentration and accumulated dose are therefore ten times too high unless the intended area density was 10 kg/m². Its five minute dose fractions of 0.548 for A and 0.274 for B inherit this discrepancy.[16]
At the stated area density and with even mixing, A's five minute exposure dose is about 32.7 g·m−3·min. B's is 8.18 g·m−3·min. Their dose fractions are 0.0545 and 0.0273, calculated from the stated mass density and lethal doses.[15][16]
For a circular burn front at constant speed, the dose follows , where is area density, flame-spread speed and room volume. On these assumptions the dose fraction reaches 1 at about 13.2 minutes for A and 16.6 minutes for B. Purser instead reports about 6 and 7.5 minutes, consistent with the higher masses in the table.[16] The qualitative ranking is the same: B's smoke is twice as potent, but B burns over one fourth the area at any given time.
Use in regulation and in hazard assessment
Regulation by smoke toxicity has been limited. New York State required testing and filing of University of Pittsburgh data for certain construction and finish materials from 1987 to 1998. Values were filed for more than 15,000 products, but no acceptance level was ever set.[1][17] The requirement was repealed when the results were widely recognised as often irrelevant to real fire hazards, and most laboratory testing stopped.[1] By the 2008 edition of the Fire Protection Handbook no laboratory was running nonproprietary animal tests for smoke toxicity on a routine basis.[1]
The 2008 handbook describes gas-concentration limits in transport. It says the closed cabinet method was used by the International Maritime Organization and the European Union as an acceptance test on measured gases.[4] The IMO Fire Test Procedures Code was adopted in 1998 and became mandatory for international service vessels built after December 2003.[18][19] Its Part 2 is a smoke and toxicity test, and interior finishes had to pass testing to ISO 5659 under that code.[20] Hirschler describes the same apparatus under the designation ASTM E1995 in 2008 ship interior-finish rules of the United States Coast Guard and NFPA 301.[21]
The 2008 handbook describes NFPA 269 as the United States method developed to generate data for hazard modelling, by the NFPA Technical Committee on Fire Tests.[22][23] Measurement of the gases themselves is covered by ASTM E800. Madrzykowski and Stroup described Fourier transform infrared spectroscopy as the preferred technique for continuous analysis of fire gases in 2008.[24]
Limits of the method
The first limit is comparability. Values from different apparatus rest on different bases, and the University of Pittsburgh value is a charged mass rather than a concentration. Results cannot be placed in a single ranking without checking what was divided by what.[1] They also depend on specimen preparation, because a ground specimen and an as-used specimen of the same product can give different answers.[1]
The second limit is the fire model. In the closed cabinet the air becomes vitiated during the test and the generation rates change, so it is unclear which fire stage is replicated. No comparisons with real-scale tests of the same products have been reported.[4] The smoke density chamber family has further known shortcomings, among them a static chamber while real smoke transfer is dynamic, oxygen-deficient atmospheres and soot deposits on walls and optics.[21] Ramped heating in the University of Pittsburgh method decomposes the material fractionally, so products given off early leave the furnace before the higher temperatures are reached.[10]
The third limit is relevance. Hirschler's 2008 review associates most fatalities with fires extending beyond the room of origin and with carbon monoxide production.[25] Gann and Bryner put the carbon monoxide yield at about 0.2 kg/kg fuel once the fuel-to-air equivalence ratio exceeds roughly 1.5; this is a specified ventilation condition, not a fixed yield for every fire.[7] Heat release rate strongly affects the time available to escape.[26] In one small-room model, doubling the heat release rate of an upholstered item cut time to lethality by 70 per cent; doubling smoke toxic potency had little effect.[25]
Regulations that cap the concentration of one or two gases measured in a particular apparatus are therefore unlikely to deliver the intended level of safety. Gas yields are not singular properties of a product.[1] The value of such a test lies in the potency and yield data it supplies to a hazard calculation. Those data sit next to measurements of ignition, flame spread and heat release from other tests.[1][2]
See also
Notes
- ↑ 1.00 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 1.10 1.11 1.12 1.13 1.14 1.15 1.16 1.17 1.18 1.19 1.20 Gann and Bryner (2008), p. 6-19.
- ↑ 2.0 2.1 2.2 2.3 2.4 Purser (2002), p. 2-145.
- ↑ Gann and Bryner (2008), p. 6-30.
- ↑ 4.00 4.01 4.02 4.03 4.04 4.05 4.06 4.07 4.08 4.09 4.10 4.11 Gann and Bryner (2008), p. 6-22.
- ↑ 5.0 5.1 Iqbal and Salley (2004), p. B-66.
- ↑ 6.0 6.1 6.2 6.3 6.4 Purser (2002), p. 2-144.
- ↑ 7.0 7.1 7.2 7.3 7.4 Gann and Bryner (2008), p. 6-13.
- ↑ Purser (2002), p. 2-153.
- ↑ 9.0 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 Gann and Bryner (2008), p. 6-20.
- ↑ 10.0 10.1 10.2 Purser (2002), p. 2-147.
- ↑ Gann and Bryner (2008), p. 6-23.
- ↑ Purser (2002), p. 2-148.
- ↑ Gann and Bryner (2008), p. 6-17.
- ↑ Purser (2002), p. 2-152.
- ↑ 15.0 15.1 15.2 15.3 15.4 Purser (2002), p. 2-98.
- ↑ 16.0 16.1 16.2 16.3 16.4 Purser (2002), p. 2-99.
- ↑ Gann and Bryner (2008), p. 6-21.
- ↑ Eberly (2008), p. 21-161.
- ↑ Eberly (2008), p. 21-171.
- ↑ Eberly (2008), p. 21-174.
- ↑ 21.0 21.1 Hirschler (2008b), p. 18-33.
- ↑ Benedetti and Colonna (2008), p. 6-230.
- ↑ Hirschler (2008a), p. 6-38.
- ↑ Madrzykowski and Stroup (2008), p. 2-45.
- ↑ 25.0 25.1 Hirschler (2008a), p. 6-50.
- ↑ Hirschler (2008a), p. 6-46.
References
- Benedetti, R. P., & Colonna, G. R. (2008). Polymeric materials. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 6-209–6-231). National Fire Protection Association.
- Eberly, R. (2008). Marine vessels. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 21-151–21-185). 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.
- Hirschler, M. M. (2008b). Interior finish. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 18-23–18-42). 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.
- Madrzykowski, D., & Stroup, D. W. (2008). Flammability hazard of materials. In A. E. Cote, C. C. Grant, J. R. Hall Jr., & R. E. Solomon (Eds.), Fire protection handbook (20th ed., pp. 2-31–2-48). National Fire Protection Association.
- 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.