smoke. The time required to reach 80 percent optical attenuation as measured across the corridor was chosen as a level of untenable smoke from the point of view of escape and inhalation. Based on data from Jin [9] for smoke of various types, with 20 percent transmission one would be able to see an illuminated sign at a distance of about 15 feet. However, studies using human subjects indicate that the limitation on visibility is not due to light obscuration of exit signs, but due to lachrymal effects of irritating products on the eyes. This limitation frequently occurred prior to serious obscuration [10]. An additional consideration is inhalation toxicity that is not defined well enough to put into quantitative terms. With these conditions in mind, table 3 was constructed. The second and third columns in this table give the percent of oxygen and carbon dioxide when the carbon monoxide level reached the 0.50 percent level. In column 6 of the table, the first number is the time (minutes) required after ignition for a 7foot meter to reach 80 percent optical attenuation and the second number is the time for 80 percent attenuation for a 4-foot meter. The time required to give 80 percent visual attenuation is comparable to the time to produce a CO concentration of 0.50 percent. The time to collapse from (0.50 percent) CO can be calculated from the relationship of Minchin [11]: For the 0.50 percent level one gets about nine minutes. In all the experiments except two, the O2 level (column 2) at the time for 0.50 percent CO is low enough that muscular coordination for skilled movement (12-15 percent) is lost and faulty judgment occurs at the 10 to 14 percent oxygen level [12]. The high carbon dioxide levels further complicate the picture by increasing the respiratory rate. In addition, the hydrogen cyanide is also present in many of the fires. Based only on the carbon monoxide concentrations, one has something less than 10 minutes to leave the corridor, however, the smoke level by the 5-minute mark is quite severe and perhaps beyond the survival limit except very close to the floor. This data would suggest that the smoke hazard is as serious as the carbon monoxide and combining this with the very rapid oxygen depletion suggests a survival time less than either hazard alone would indicate. The conclusions regarding oxygen depletion would tend to support the conclusions of Kingman et al [13]. The hazard that exists in another room connected to the corridor may not be oxygen depletion, but CO, which is the conclusion of Robinson et al [14] in rooms above the fire although the results from the corridor do not suggest that one hazard is predominant. Another hazard that contributes to the observed rapid flame spread rate in the corridor experiments, that has not been considered, is the "gas phase flash-over" phenomena. As indicated in table 2 there were significant concentrations of methane, ethylene, and acetylene in the grab samples. The table further indicates that the total hydrocarbon content of these samples was high at the time when "flash-over" occurred. The "flash-over" phenomena seen in the corridor experiments consists of the build-up of a combustible gas mixture, and is not the same phenomena that results when a material, such as furniture, bursts into flames when exposed to a large radiant energy source in another area of a room. While it is true that radiant energy contributed to the build-up of a combustible gas mixture, the rapid flame spread down the corridor appears to be a gas phase phenomenon. A further verification of this was obtained in experiments 348 and 349 using a combustible gas detector located at the exit window of the corridor. In experiment 348 the detector showed a rapid rise at 450 seconds and in experiment 349 a rapid rise was recorded at 640 seconds, correlating well with observed "flash-over" times of 440 and 630 seconds, respectively. Since the purpose of the full-scale floor covering fire experiments was to determine the nature of the hazard and the relevance of present test methods, the obvious question is: what insights and guidance can be obtained from these experiments for developing a laboratory scale or model for testing purposes that is realistic and assesses the hazard? If a laboratory measurement of the amount of smoke produced from a floor covering material such as in the NBS smoke box is to be realistic, then the dependence or independence of smoke quantities of oxygen availability must be determined in the light of the extensive oxygen depletion found in the corridor. Perhaps smoldering combustion and flaming combustion conditions will actually be the boundary extremes that represent the maximum and minimum smoke-producing conditions and it will not be necessary to control the atmosphere in the smoke box. However, this can only be determined by actual experimentation, as indicated earlier. In relation to the toxicological hazard, it may also be necessary to determine the relative toxicity of the type of products produced in an oxygen limited atmosphere in addition to the increased CO production. One might raise the question, is the production of HCN, HC1 and other toxic products enhanced in the limited oxygen atmosphere? And, is there a correlation between smoke production and carbon monoxide? The hazard associated with the rapid "gas phase flash-over" must also be investigated. It may well be that the rapid flame spread reported in fires involving floor covering materials is actually a gas phase phenomenon resulting from the build-up of products related to the incomplete combustion process due to the limited availability of oxygen as seen in the corridor experiments and not just due to radiant energy. The authors would like to acknowledge the contributions of Mr. F. D. Hileman, Mrs. E. C. Hare, and Mr. B. D. Buchanan. 4. REFERENCES [1] Gross, D., Loftus, J. J. and Robertson, A. F., Method for Measuring [2] [3] [4] [5] Hilado, C., Jr., The Effect of Chemical and Physical Factors on Smoke Chien, W. P., Seader, J. D. and Birky, M. M., Monitoring Weight Loss Gaskill, J. R., Smoke Development in Polymers During Pyrolysis or Combustion, J. Fire and Flammability, Vol. 1, 183-216 (July 1970). Birky, M. M. and Manka, M. J., The Use of Room and Corridor Tests in [6] Hueper, W. C., Kotin, P., Tabor, E. C., Payne, W. W., Falk, H. and Sawicki, E., Arch. Pathology, Vol. 74, 89 (1962). [7] Einhorn, I. N. and Seader, J. D., Methods for the Identification of Products of Combustion of Polymeric Materials A Computerized Analytical System, paper presented at the 1973 Polymer Conference Series, Flammability Characteristics of Materials, June 11-15, 1973, The University of Utah, Salt Lake City, unpublished. [8] [9] [10] [11] [12] [13] [14] Fung, F., Suchomel, M. and Oglesby, P., Fire Journal, Vol. 67, 41 (1973). Jin, T., Visibility Through Fire Smoke, Report of Fire Research Lopez, E. L., Study of Smoke Emission From Burning Cabin Materials and the Effect on Visibility in Wide-Bodied Jet Transports, Report No. DOC FA 72 NA-655 (May 1973). Minchin, L. T., Mild Carbon Monoxide Poisoning as an Industrial Hazard, Bulletin of Research No. 53, Underwriters' Laboratories, Inc., 49. Robinson, M. M., Wagner, P. E. and Fristrom, R. M., The Accumulation NATIONAL BUREAU OF STANDARDS SPECIAL PUBLICATION 411, Fire Safety Research, Proceedings of a Symposium Held at NBS, Gaithersburg, Md., August 22, 1973, (Issued November 1974) CONTRIBUTION OF INTERIOR FINISH MATERIALS J. B. Fang and D. Gross National Bureau of Standards, Washington, D.C. Characterization of the fire environment from the burning of the combustible contents of wastebaskets, upholstered furniture and interior finish materials is important for developing rational tests and establishing design criteria for reduction of fire hazard in buildings. Some experimental results on the burning characteristics of an upholstered chair, contents of waste receptacles and wood crib arrays in a well-ventilated room are presented. A procedure has been developed for evaluating the contribution to fire growth of wall and ceiling panels in a full-scale room corner with a standardized wood crib duplicating the conditions produced by an incidental fire. Results of full-scale and laboratory tests with selected interior finish materials on ease of ignition, surface flammability, flame penetration and smoke and heat generation measurements are presented and compared. Key words: Buildings; fire intensity; flame spread; flames; fur- Interior finish materials on walls, ceilings and floors represent large exposed areas over which flame may spread, and rapid flame spread is one of the most frequent contributing factors to residential fire fatalities. An obvious method of reducing fire losses in buildings is to reduce the possibility of ignition and spread of a fire. This can be achieved by limitations on the use of materials with a high contribution to fire growth, confinement of the fire to the room of its origin with the aid of fire resistant constructions, and installation of fire detection and/or sprinkler systems for early warning and control of developing fires. In recent years, the rapidly increasing use of new interior finish materials in building construction introduces building fire safety problems since some of these newer materials can contribute to the rapid spread of fire in its early states. Urgently needed is the evaluation of the contributions made by these materials to the development of a compartment fire and to establish more meaningful performance criteria of building materials and constructions. Small-scale laboratory tests can provide a comparative measure of surface flammability, flame penetration, heat release and smoke producing properties which are important from the standpoint of life hazard and property damage. However, there is little substantiation that the test conditions prevailing in most laboratory scale experiments simulate real fire environments. The corner formed by the intersection of a ceiling and two adjacent walls constitutes a critical configuration for the evaluation of the fire performance of interior finish materials since fires within a corner represent a severe fire exposure through an increase of heat and confinement of combustion products. In order to compare flame spread characteristics of various building materials, corner wall fire tests have been developed to serve as an evaluation tool at the Forest Products Laboratory [1], Factory Mutual, Underwriters' Laboratories and other laboratories. Most recently, corner fire tests initiated by small incidental fire exposure were used to demonstrate residential fire hazards [2]. This work was supported in part by the U.S. Department of Housing and Urban Development. |