Combustible ceilings are typically known to enhance fire growth in a room fire. For wooden linings, impregnating fire-retardant chemicals is popular and simple solution, but there is still a concern that fire safety cannot be assured due to unevenness of the chemical amount inside wood. This unevenness is caused by efflorescence during manufacturing process and their service lives, with fire-retardant chemicals being excess near the surface and less inside the wood. This study evaluated those effect on its heat release and flame spread through cone calorimeter test (ISO5660) and model box test (ISO/TS 17431:2006).
In this study, three patterns of chemical distribution were considered. Group-S represents the wood after efflorescence; more chemicals near the front, and group-B is opposite. U is the group of uniformly treated wood. All fire-retardant-treated specimens were processed in a chemical pool under a pressure of 0.95MPa and then dried. The eventual thickness of the specimens was 18mm thick. The base wood was Japanese cedar (Cryptomeria Japonica D. Don) and the main component of fire-retardant was polyphosphate carbamate, both commonly used in Japan.
To figure out potential effects of unevenness, cone calorimeter tests were carried out. Each group consists of four or eight specimens with 11 to 186kg/m3 of chemical content. The specimens were heated to ignite under 50kW/m2 intensity, and heat release rate and mass loss rate were observed for 20 min.
To evaluate the changes in flame spread characteristics during their service lives, model box tests were carried out for group-S and U. Both groups had two specimens with chemical content of approximately 70 and 120kg/m3, which respectively corresponds to the recommended amount to achieve fire-retardant material and quasi-noncombustible material in Japanese building standard laws. A small room (W0.84×D1.68×H0.84m) with one opening (W0.3×H0.67m) was finished with each specimen, and burnt for 10min by a diffusion burner set at one corner. The burner was 17cm square, and heat release rate of the source was 40kW constant. During the experiment, heat release rate and the air temperatures were measured.
Based on the cone calorimeter tests on group-U, it was clarified that the amount of chemicals for suppressing flame combustion lies between 23 to 42kg/m3, and above this amount, both peak and total heat release rarely decrease. Compared to group-U, the combustion of group-B was generally enhanced, and flame combustion partially occurred with areas less than 34kg/m3. As a result, the integrated heat release up to 20 min also increased compared to that uniformly treated. On the other hand, in group-S, flame combustion as the surface carbonizes is suppressed, so the total heat release up to 20 min were equal or less than group-U. From these experiments, it was expected that efflorescence hardly influences on flame spread as long as the least area exceeds 34kg/m3 for this chemicals. Flame spread in model box tests agreed with this tendency, while the burnt area and eventual heat release rate significantly decreased as the average chemical content increased. It may be that heat resistance that fire-retardant obtains suppressed flame spread. Through these experiments, we concluded that unevenness in fire-retardant-treated product potentially enhance its combustion, but efflorescence hardly enhance flame spread as seen in interior usage.
