FIRE-RESISTING BUILDING MATERIALS
A bulletin describing the fire-resistive qualities of building materials will be issued within a short time by the United States Geological Survey through its Technologic branch. Various materials used in building were subjected to the direct application of heat for two hours in the laboratories of the National Board of Fire Underwriters in Chicago, the temperature reaching that of a conflagration. After being exposed to the fire the materials were withdrawn from the furnace and quenched with water, in order to get the conditions that would obtain after the firemen had begun their work on a burning building.
The investigations, which were conducted under the direction of Richard L. Humphrey, engineer in charge of the structural materials laboratory of the survey, are the first of a comprehensive series undertaken with the object of determining the fire-resistive properties and rates of heat-conductivity of various building materials -and the comparative efficiency of the various methods of fire-proofing. The investigations have the further object of a greatly lessening the liability of loss by fire in government buildings which are not insured, and greatly reducing their cost through more efficient methods of construction. Y htie the tests are conducted primarily for the purpose of obtaining information of essential value to the government, it is believed that the results will prove of much importance to the public.
Mr. Humphrey in his bulletin speaks of the desirability of publishing a detailed account of them, although the results are preliminary and inconclusive, for the purpose of making public the information required, not to he used as a basis for making conclusions, but merely as preliminary data regarding the relative fire-resistive qualities of the building materials tested. The conditions under which these tests were made were unusually severe, and nom of the materials passed perfectly. The tempert / ires used would hardly be reached in an ordinary fire. Much of the damage done to the building materials in this series of tests was occasioned by internal stresses, the gas flame of the furnace heating one face of the test-pieces much more rapidly than the other. All the materials tested, including the hydraulic pressed brick cracked more or less. The concrete cracked least; but the tests indicate the necessity for using metal reinforcement in concrete walls to distribute he effect of the expansion.
The fact brought out most clearly by these tests is the low rate of heat transmission of Portland cement concretes and mortars. This is one of the desirable qualities in buildings intended for fireproofing purposes. For markittg the cement blocks, linen tags were fastened by wire nails to the interior walls at the time of moulding. Most of these tags remained in place during the test, and. when the walls were demolished, the tags, in every case, were found entirely undamaged.
An average of all the tests shows that about 90 per cent, of the maximum temperatures attained by the faces of the panels were reached in one hour, while, in the case of the mortar blocks, the increase in temperature of the backs of the panels in one hour was only about 20 per cent, of the total increase in the two hours. Of the solid concrete panels, the cinder concrete showed the smallest increase in the temperatures of the backs of the walls, and the granite concrete the largest. The total increase in the temperatures of the mortar panels was about the same as the total increase in the temperatures of the backs of the concrete; but the increase in one hour was somewdiat greater in the case of concrete.
The damage done to the faces of the mortar and concrete panels would probably be caused at a temperature about half that of the maximum temperatures reached, owing to the water of crystallisation being driven from the mortar. This thin layer of dehydrated mortar may have formed a protective coating and prevented dehydration. The backs of the brick panels also showed a small increase in temperature; but the natural building stones and tiles proved poorer nonconductors of heat.
It is apparent that the strength of the webs of ordinary hollow blocks is insufficient to resist the stresses set up in these tests, as, in many tests, the rapid rise in temperature and the subsequent quenching of one of the faces of the blocks caused the webs to split. It was noticeable that the richer the mortars used in these blocks the better they withstood the tests. The amount ol water used in mixing the mortars had a similar effect on the fire-resistive qualities; the mortars mixed with the greatest percentage of water gave the best results.
When the blocks were cracked or spalled before the application of water, the damage appeared to he greater in the dry mixtures containing the greatest percentage of sand, and it was further observed during the fire test that the richer mixtures warmed up more slowly than the others. It is apparent that one of the causes of weakness in the cement hollow building blocks under these fire tests was the weakness of the concrete, a too dry’ and lean mixture, which, coupled with the thinness of the webs, provided insufficient strength to resist the stress due to the rapid expansion of the face. It is quite possible, as was shown in some of the block tests, to make blocks which w.ll pass the conditions perfectly: the web must be thick enough to give the necessary strength.
The brick panels probably withstood the tests better than the other materials. The common brick tested comprised unused new Chicago brick, and used St. Louis brick. Fifty per cent, of the new bricks were split, while 60 to 70 per cent, of the old bricks were not damaged. Limeknots seemed to be responsible for most of the damage to (he new bricks, as thev were found at the bottom of nearly all the cracks. The bricks at the back of the panels were entirely unaffected. While the strength tests are not conclusive, there is apparently little difference in the strength of the bricks before and after firing.
The hydraulic-pressed brick withstood the test very well. No damage was apparent after the firing and before the water was applied, and, al though a number of the bricks cracked. 70 per cent, of them were found to be sound after quenching. The sand-lime brick did not withstand the test as well as expected; but the sample tested, which was purchased at random, appeared to be somewhat below average quality. The con crete behaved remarkably well, and it was difficult to determine whether the limestone, granite, gravel or cinder concrete sustained the least damage. The faces of all the panels were more or less pitted by the fire and washed away by the Stream of water. The test was unfair to the cinder concrete, as the cinder was very poor, containing a large percentage of unburned coal; the sample selected, however, was the best of six or eight investigated in St. Louis. During the fire test the coal ignited and left the surface of the concrete very rough and badlv pitted. The limestone aggregate in the face calcined, and the granite aggregate split and broke away from the surface-mortar. rhe granite concrete probably behaved better than any of the above mentioned concretes. Damage in no case extended deeply, probably not more than !½ in. The evidence show’s that even at this depth the temperature was comparatively low. The high stresses produced in the panels by the rapid rise of the temperatures of the faces, while the backs remained cool, caused cracks. On taking down the panels, the blocks of concrete were found to be cracked vertically for some distance from the face. The serious damage done to the various natural building stones precludes any comparison among them.
The tile tested was bought in the open market, and in one panel was taken from a lot of material about to be erected in a building. A large percentage of the faces of the tiles were washed a way by the water, and the material composing the faces became soft and could easily be crumbled in the fingers. There was a comparatively rapid rise in the temperature of the backs of the panels. The plaster on the partition-panel tile fell off a few minutes after the test was started.