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)

AN EVALUATION OF FLAME SPREAD TEST METHODS FOR

FLOOR COVERING MATERIALS

James Quintiere and Clayton Huggett
National Bureau of Standards, Washington, D.C.

Flammability properties of materials have traditionally
been measured by small scale laboratory tests. The relationships
between test results and performance in real fires have been
largely inferred by intuition or subjective judgement.

spread test methods for floor covering materials are examined.
Through full-scale fire experiments and laboratory studies the
nature of the potential flame spread hazard of flooring materials
is presented. The factors promoting flame spread in each test
method are identified. Test method results are compared with
relevant full-scale fire experiments involving floor covering
materials in a corridor. An effort is made to relate test results,
where possible, to the potential flame spread hazard of floor
covering materials in building corridors and exitways.

Key words: Corridor fires; fire test methods; flame spread; flammability tests; floor covering flammability; floor coverings.

1. FIRE SPREAD OVER FLOOR COVERINGS IN BUILDINGS

In the last decade, concern has been expressed about the flammability hazard of floor covering materials. This concern has probably been induced by the widespread use of carpeting and by several fire incidents which have suggested that the floor covering was a significant factor in fire spread. Yuill [1] described the state of this issue in 1970. Previous to this, except for fire safety regulations for hospitals financed under the Hill-Burton Act (1965), fire officials had not considered floor coverings to be a significant fire hazard, although, several fires had indicated that carpeting could be significant in the early spread of a fire. Using a 5 lb wood crib as an ignition source, Yuill demonstrated that two high-cut pile (shag) carpets would propagate a flame over a 5 ft wide x 15 ft long corridor. (One carpet had a flame spread rating of 55 by ASTM E-84 [14], the other had a rating of 1100!) With the introduction of DOC FF-1-70 [2], a mandatory nationwide flammability standard for carpets incorporating the "Pill Test", one may assume that no carpets easily ignited by "small" ignition sources entered commerce after April 15, 1971. However, the flammability hazard associated with carpets which pass the Pill Test and with other floor covering materials not subject to regulation remains to be defined and dealt with.

1.1. Fire Incidents

In recent years several fires in nursing or retirement homes and hotels, apparently aggravated by floor covering flame spread, have resulted in deaths and extensive loss in property. Thus, one of the objectives of the evaluation reported here has been an analysis of the reports of these and other fires with the purpose of gaining some appreciation of the role played by the floor covering material in the spread of fire.

Robertson [3] has reviewed the available statistics. The following is taken from this report, where tables 1 and 2 of this paper are condensed versions of tables developed by Robertson. He states

"...the data available are not voluminous. They comprise reports published in the NFPA Fire Journal and fire injury reports in the NBS

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)

ADDITIONAL STUDIES OF THE TRANSFER OF FLAME

RETARDANT EFFECTS WITH CELLULOSIC FABRICS

Bernard Miller

Textile Research Institute, Princeton, New Jersey

Burning rate measurements on double layers of the same fabric
when one layer has been treated with a flame retardant have indi-
cated that certain effects of the retardant can be transferred to
the untreated layer. To learn more about the mode and chemistry
of this phenomenon, a study of non-flaming combustion of cellulosics
has been carried out on mixed systems using thermogravimetric analy-
sis. By arranging to have untreated cotton physically separated
from the flame retardant material during heating it was possible
to determine that the transfer depends on a chemical process and is
most likely the effect of a volatile product generated during
heating. Data are presented also showing that rayon containing an
alkoxy-phosphazene flame retardant does not transfer its flamma –
bility properties to untreated rayon.

Key words: Cellulosics; cotton; DAP; fabric flammability; flame
retardants; flammability; rayon; thermogravimetric analysis.

1. INTRODUCTION AND BACKGROUND

The customs of usage of textile materials make it highly probable that many fabrics will be used, either deliberately or inadvertently, as components of multilayer assemblies. When such structures have layers made up of different materials, it is necessary for us to know whether or not the flammability behavior of the combinations would be predictable from knowledge of the behavior of the individual fabrics involved. A large amount of experimental data is needed to answer this question; much of it is still to be obtained. We have been studying one aspect of the problem: the behavior of double layers of a single material when one of the layers has been treated with a flame retardant. In a previous publication [1] it was shown that the effect of a flame retardant, in terms of altering the burning rate, could be transferred to an adjacent untreated fabric. This was found to be the case for cotton treated with diammonium phosphate (DAP), for nylon treated with thiourea, and for polyester treated with tris (2,3-dibromopropyl) phosphate (T23P). With cotton plus DAP, transfer was most evident when the treated layer was placed underneath the untreated cotton (during horizontal burning). This has led to the conjecture that the transfer effect is the result of the crossover of a volatile product formed during the heating of the treated fabric (that is, physical contact between the layers may not be a necessary condition for this phenomenon).

It was thought that useful information relevant to this and other points might be obtained if potential transfer situations could be studied without the occurrence of flaming combustion. Consequently, an experimental program was devised to study the pyrolysis behavior of such systems with controlled degrees of contact between components.

2. PYROLYSIS WEIGHT LOSS STUDIES

Weight loss studies on heating in air were carried out with the aid of a thermogravimetric analyzer (TGA) which used a platinum cup as a macro sample holder (fig. 1). This holder, considerably larger and deeper than the sample holders of most commercial TGA units, allowed the stacking of materials one above the other. One material could be placed in the bottom of the cup below a removable metal mesh screen which served to support a piece of fabric as the upper component. It was thus possible to obtain weight loss data under programmed heating for a variety of stacked combinations.

2.1. TGA of Cotton and DAP-Cotton Fabric Systems

Figure 2 compares TGA thermograms for an untreated cotton fabric and the same fabric treated with flame retardant. A considerable change in the pattern of weight loss with increasing temperature is introduced by the addition of 7.8% DAP; both the temperature at which the maximum rate of weight loss occurs and the amount of weight loss are reduced. Figure 3 shows thermograms for two systems containing equal weights of the untreated and the treated cotton, in one case intimately mixed, and in the other separated, with the treated fabric underneath. Two weight loss peaks occur close together in the case of the intimate mixture, the first one at a temperature slightly higher than the maximum rate of loss temperature for the treated fabric. For the separated system, the two peaks are discrete. In both instances, the second drop in weight occurs at a temperature lower than the peak rate of loss temperature for plain cotton. This is an indication that the effect of DAP on the peak rate of loss temperature of cotton is in some measure transferred from treated to untreated layer even when the two are not in contact.

In table 1 are listed the amounts of residual char at 500°C (taken from the thermograms) for the four systems studied, and the directly measured final weights of residue in the two-layer experiment after heating to 560°C. It is seen that DAP, like most flame retardants for cellulose, increases char production. The results for the mixed systems substantiate that cotton is modified in the presence of DAP-treated fabric even without contact. If there were no interaction, the total char weight at 500°C for each of the mixed systems should be no more than (17.5 + 1)/2 = 9.3%, whereas the percentages are 13.8% and 17.28. Further, since plain cotton leaves virtually no char after heating to 560°C, the weight of residue in the cotton layer after the two-layer experiment, if there were no transfer, should be insignificant rather than the measured 5.0%. A control experiment in which two separated layers of untreated cotton were heated produced a 1% residue, with no significant difference between the amounts of solid remaining in each layer.

aFinal weights of each layer after heating to 560°C and cooling to room

temperature.

2.2. TGA of Cotton Fabric Pyrolyzed Over Solid DAP

A series of two-layer TGA experiments were run involving varying proportions of plain cotton fabric and solid DAP, with the latter in the bottom of the crucible. By itself DAP salt begins to lose weight slowly at 150°C; at 500°C under these heating conditions it has lost 29% of its original weight, a few percent more than the theoretical value if the only loss were two molecules of ammonia. Allowing for this, the weight of cotton residue can be calculated from the total weight loss of the system.