[3], which also discussed the general relationship between the measured specific optical density and the level of smoke through which a light (or lighted exit sign) may be seen.

The tests involved a thermal irradiation exposure of 2.5 W/cm2 (2.2 Btu/ft2 s)2 normal to the exterior surface of a 3 x 3 in specimen and were performed under both flaming and nonflaming (smoldering) exposure. To induce open flaming in the former case, a small pilot (0.35 SCFH natural gas diffusion flame in a 6 in i.d. tube) was applied at the base of the specimen. These conditions were selected to provide a wide range of smoke levels for different types of materials. The size of the specimen and the volume of the chamber were such that complete oxidation of practically all materials could occur without appreciable decrease in oxygen content. Materials were furnished by FAA and were tested using a typical section in the thickness supplied.

100 T

Optical density, defined as D = log (where T = percent light transmission), is the single most characteristic measure of the obscuring quality of a smoke. Specific optical density, D., is a property of a specimen of given thickness,

FIGURE 1. Smoke test chamber.

36°

by infrared techniques because they have weak absorption bands and because pigments, fillers and polymer components with which they are mixed generally have overlapping spectral bands. As much as 20 to 40 percent of PVC or poly (vinylidene chloride) could go undetected.

Generic names are given in all cases, even though the spectra for some materials were so similar to reference spectra identified by trade name from the literature that very little doubt existed as to source.

2.2. Smoke Measurements

The smoke level was determined by measuring the progressive attenuation of a light beam passed through the smoke aerosol within an enclosed smoke chamber (see figs. 1 and 2). Smoke is reported in terms of specific optical density, a dimensionless attenuation coefficient which defines the amount of smoke accumulated from a specimen of unit surface area in terms of its photometric obscuration over unit path length within a chamber of unit volume. For the typical application in which the material is to be used as an interior finish (e.g. on walls, ceilings, floors), the fire-exposed surface area of the specimen governs its smoke-production behavior. Specimen thickness (unit weight) correspond to the materials as supplied and used. The basis and limitation of the method were described in detail in a recent paper

=

=

For the test chamber, V = 18 ft3 (0.510 m3), A 0.0456 ft2 (0.00424 m2), and L = 3 ft (0.914 m). Ideally, the change in D, with time during the smoke accumulation process will depend only upon the thickness of the specimen, its chemical and physical properties, and the exposure conditions. The results are reported in terms of (a) maximum (total) smoke accumulation, Dm, (b) maximum rate of smoke accumulation (over a 2-min period), Rm, and (c) the time period, te, to reach a "critical" specific optical density of 16, under the test conditions.

However, there are definite limitations to the use of specific optical density for extrapolation and comparison with other box volumes, specimen areas and photometric systems, and for extension to human visibility. The degree to which such extensions are valid depend upon a number of major assumptions: the smoke generated is uniformly distributed and is independent of the amount of excess air available and of any specimen edge effects; coagulation and deposition of smoke is similar regardless of the specimen size, or the size and shape of the chamber; for any given smoke the optical density is linearly related to concentration; and human and photometric vision through light-scattering smoke aerosols, expressed in terms of optical density, are similar.

2.3. Gas Analysis

Indications of the concentrations of gaseous products were obtained by drawing a sample of the gas mixture in the smoke test chamber through commercial colorimetric gas detector tubes and

reporting results on the basis of the manufacturers calibrations for the selected gases [4]. Essentially, a colorimetric tube is a small-bore glass tube containing a chemical packing which changes color when exposed to a specific component of a gas mixture, and the length of color stain is related to the concentration of that component for a given quantity and rate of flow of gas. Layers of precleaning granules and a plug to absorb interfering gases and to control the sample flow rate are generally provided. Sampling was done several times during each smoke test using a small syringe or bellows pump designed to aspirate a measured volume of gas each stroke. The gas detector tube was inserted into the smoke chamber from the top, and was situated 3 in below the top surface of the chamber (approximately 25 in above the level of the specimen). In some instances an attempt was made to extend the range of these indicators by drawing less than the recommended gas volume through them and reporting results on the basis of individual laboratory calibrations, as reported in a later section. More detailed discussion of product gas analysis by colorimetric detector tubes and by specific ion electrode are presented in appendix 1.

Indicator tubes were used to detect CO, HCN, HCI, HF, SO2, NO + NO2, NH3, Cl2 and COCl2, since these gases have generally been considered toxicologically hazardous compared with other possible components. However, these are not necessarily the only potentially toxic components released. No attempts were made to determine high concentrations of CO2 or low concentrations of O2, or to consider the type, size, or concentration of smoke particles in toxicological terms. Information on the analytical limits for the tubes used, and references to the toxic hazard limits of these gases are discussed in appendix 1. Where HCl was one of the products, in many cases the gas was also absorbed in water and analyzed by a chlorine ion electrode to provide a more accurate indication at high concentrations.

and wool), a few were composed of a mixture of natural and artificial fibers, but the bulk of the fabrics were made from 100 percent artificial fibers, including acrylics, modacrylics, polyesters, polyamides (nylon-type), vinyl, and glass.

Of the sheet and laminate materials, approximately one-half were composed entirely or predominantly of poly vinyl chloride (PVC), and the remaining sheet and laminate materials were composed of acrylonitrile-butadiene-styrene (ABS), methyl methacrylate, and other copolymers, blends, and varieties of polymers. The rugs tested included wool, modacrylics, polyamide (nylon and aromatic types), and polypropylene. Of the pads used for seats, there were several urethane foam materials and one rubber (chloroprene). The materials used as ceiling or bulkhead insulation in

Smoke measurements are summarized in appendix 3 in terms of the maximum smoke accumulation (Dm), the maximum rate of smoke accumulation (R) and the time (t.) to reach a specific optical density of 16 for both flaming and smoldering exposure. These results represent averages of duplicate tests (with few exceptions). For Dm values up to 200, the standard deviation was 11.8 for flaming and 9.2 for nonflaming tests. Smoke buildup curves for typical flaming and smoldering tests on selected types of materials are shown in appendix 4.

=

A wide range of Dm values was measured. Slightly more than 15 percent of the materials produced smoke corresponding to a Dm 16 or less, for both flaming and smoldering exposures. These included materials composed of glass, asbestos, aromatic polyamide, polyimide plus others, but many of these materials were very thin (lightweight). Dm values in excess of 200 were recorded for flaming and smoldering exposures on approximately 20 percent of the materials.

For flaming exposure of 140 materials, frequency distribution histograms of the maximum smoke values are shown in figure 3 for all materials, and in figure 4 within the classification groups: (a) fabrics, (b) rugs, (c) sheets, films, and laminates, and (d) pads, insulations, and assemblies. Of the materials in the Dm 16 category, 16 were fabrics, 6 were sheets or films, and 4 were glass or asbestos fiber insulations.

With one exception, all materials in the Dm ≤ 16

category under flaming conditions were also Dm < 16 under nonflaming conditions.

Figures 5, 6, and 7 comprise a complete histogram showing smoke and toxic gas concentrations for flaming and nonflaming exposures on each material based on the data in appendix 3. Materials have been arranged according to classification by groups, by composition, and by generally increasing weight within each subgroup.

It should be noted that only the "front" side of a material was exposed, and that specimens exhibited a very wide range in their physical and thermal behavior during flaming and nonflaming exposure. Materials which melted at fairly low temperatures, including nylon, polysulfone, and polyethylene, flowed to the bottom or dripped off the sample holder in varying degrees, resulting in less smoke. Some materials evaporated fairly rapidly before extensive decomposition or combustion took place. All urethane foam materials produced more smoke under smoldering exposure than with flaming exposure, except in one instance where the material was noted to shrink into a corner of the holder and was, therefore, subjected to less radiation. Rubber (chloroprene), ABS, methacrylate, and PVC materials nearly always produced more smoke under flaming exposure. Under thermal radiation exposure alone, elastomers generally formed a bell-shaped protrusion at their center through which gaseous products streamed out rapidly. The maximum smoke level depends upon the thickness (and density) of the specimen, and for some materials Dm may be expected to increase with thickness but not always in direct proportion [3].

FIGURE 5. Smoke and gas concentrations for individual materials-fabrics, rugs.