had simultaneously been exposed to binary mixtures plus NO2, plus CO, CO2, or O2.

The new seven-gas model including NO2 is presented in Equation (2).

where terms are defined as for Equation (1). The LC50 for NO2 is 200 ppm following 30-minute exposures. Either the six-gas (Equation 1) or the seven-gas (Equation 2) model can be used to predict deaths that will occur only during the smoke exposure or those that will occur during and following the exposure. The seven-gas model is used if NO2 predict the lethal toxicity of atmospheres that do not include NO2, Equation (1) is used.

The N-gas model has been developed into an N-gas method. This method reduces the time necessary to evaluate a material and the number of test animals needed for the toxic potency determination. It also indicates whether the toxicity is usual (i.e., the toxicity can be explained by the measured gases) or is unusual (additional gases are needed to explain the toxicity). To measure the toxic potency of a given material with this N-gas method, a sample is combusted under the conditions of concern, and the concentration of gases considered in the model are measured. Based on the results of the chemical analytical tests and the knowledge of the interactions of the measured gases, an approximate LC50 value is predicted. In just two additional tests, six rats are exposed to the smoke from a material sample size estimated to produce an atmosphere equivalent to the approximate LC50 level (this can be for exposure or exposure plus post-exposure). Since the concentration-response curves for animal lethalities from smoke are very steep, it is assumed that if some percentage (not 0 or 100 percent) of animals die, the experimental loading based on the predicted LC50 value is close to the actual LC50. No deaths may indicate an antagonistic interaction of the combustion gases. The deaths of all of the animals may indicate the presence of unknown toxicants or other adverse factors. If more accuracy is needed, then a detailed LC50 can be determined.

Results using the N-gas method have shown the good predictability of this approach,1 Validation studies looking at a series of materials and products under conditions ranging from laboratory bench-scale to full-scale room burns indicated that, in all cases, the six-gas model was able to predict deaths correctly (Braun et al., 1988, 1990. 1991; Babrauskas et al., 1990, 1991a, b). The seven-gas model works when the animals are exposed to various concentrations and combinations of the tested gases (Levin et al., 1995). Studies need to be done to ensure that the seven-gas model predicts the outcome when nitrogen-containing materials and products are thermally decomposed.

TOXICITY TESTING

Toxicity screening tests for both acute and delayed effects are needed to evaluate the combustion products, including irritant gases of any newly proposed aircraft interior materials and products. It is imperative that the materials and products be tested under experimental conditions that simulate the realistic fire scenarios of concern in aircraft interiors as described earlier in this appendix. Tests should be simple, rapid, inexpensive, use the least amount of sample possible (since, in many cases, only small amounts of the developmental material may be available), use a minimum number of test animals, and have a definitive toxicological endpoint for comparison with other material candidates. While faulty judgment and incapacitation of passengers in an aircraft fire are significant causes of worry since they can prevent escape and cause death, they are complex endpoints that cannot be directly measured. Death of experimental animals (e.g., rats) is a more definitive and easily determined endpoint and can be used to compare the relative toxicities of alternative materials. Using lethality as the sole endpoint assumes that materials with greater toxicity based on a lethality endpoint will also cause more severe incapacitation and impairment. The number of experimental animals can be significantly reduced by utilizing one of the predictive mathematical models developed for combustion toxicology (Hartzell, 1994; Levin et al., 1995; Purser, 1995).

Many test methods for the determination of the acute toxicity of combustion products from materials and products have been developed over the last two decades and continue to be developed and improved (Kaplan et al., 1983; Norris, 1988; Caldwell and Alarie, 1991; Levin, 1992a, b). Two methods that are well known and have been used to generate much toxicity data are University of Pittsburgh I, a flow-through smoke toxicity method (Alarie and Anderson. 1979, 1981) and the closed-system cup furnace smoke toxicity method developed at the National Institute of Standards and Technology (NIST) (Levin et al., 1982, 1991). More recently,

1  

The LC50 values given for use in equations (1) and (2) are dependent on the test protocol, on the source of test animals, and on the rat strain. It is important to verify the above values whenever different conditions prevail and if necessary, to determine the values that would be applicable under the new conditions.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement