US3454456A - Fire resistant plaster product - Google Patents

Description

United States Patent ice 3,454,456 FIRE RESISTANT PLASTER PRODUCT Grant S. Willey, Lake Forest, Ill., assignor to United States Gypsum 'Company, Chicago, Ill., a corporation of Illinois No Drawing. Filed June 1, 1965, Ser. No. 460,498 Int. Cl. D2111 3/00; B32b 13/08 U.S. Cl. 161-162 7 Claims ABSTRACT OF THE DISCLOSURE A set calcined gypsum cast of improved fire rating and suitable as a core for gypsum wallboard comprises unexpanded vermiculite of specified particle sizes distributed throughout, the relative proportions being selected so as to impart to the gypsum cast a linear contraction of about 0.0 to 0.2% after being heated and substantially no thickness expansion after being cooled. At least about 0.1%, by Weight, of mineral fibers and boric acid may also be dispersed throughout.

This invention relates to an improved fire resistant plaster product, and more particularly to one containing unexpanded vermiculite particles.

Products containing set calcined gypsum, such for example as gypsum wallboard, have been used for many years as a fire barrier in buildings, particularly housing. While the usual or /2 thick gypsum boards have a high fire resistance there are certain conditions under which a still higher fire rating is desirable, such for example in multiple dwellings. In order to effect this higher fire resistance, it was necessary in the past to use an extrathick wallboard which not only was more costly but also more difficult to erect due to increased weight. Hence, it became desirable that the latent fire resistance of the gypsum board core be improved so that the thickness and hence the weight can be decreased.

In a study of the action of gypsum wallboard when exposed to fire, such as in a fire test, it was evident that there was a substantial shrinkage of the wallboard core at high temperature with consequent cracking which not only materially contributed to the passage of excessive heat and hot gases through the wall but also hastened the actual disintegration of the board under such adverse conditions. Hence, if this excessive shrinkage could be overcome or decreased the fire resistance would be substantially increased.

Patent No. 2,681,863 teaches that the even distribution of individual short textile glass fibers throughout a gypsum cast markedly reduces cracking when the cast, such as in the form of a core of a gypsum wallboard is subjected to fire. While the addition of the glass fibers to the board core did not markedly decrease the shrinkage, it enabled the calcined core to be held intact, and hence so materially increased the fire rating of the wallboard that a thick board was acceptable to the Underwriters Laboratories for a one hour fire rating when used as the ceiling under wooden joists with oak flooring placed thereover, i.e. their Design No. 1-1 hr. It has been established in a great number of building codes that such a construction must qualify for the above rating.

In Patent Nos. 2,526,066 and 2,744,022 there is set forth a further improvement in the fire resistance of a gypsum cast by the addition of minus 28 Tyler, 30 U.S. Standard, or minus 35 Tyler, 40 U.S. Standard, sieve sizes of unexpanded vermiculite particles to the calcined gypsum-water mix prior to the setting thereof. This addition substantially decreases the adverse contraction of a gypsum cast subjected to high temperature conditions be cause of the expansion of the vermiculite when it also be- 3,454,456 Patented July 8, 1969 comes subjected to the high heat. In fact, the expanding vermiculite could cause an expansion of the core which, if excessive, would be undesirable as the board would buckle and become loosened. The expanding vermiculite not only aids in overcoming shrinkage, but also increases the heat insulation with consequent retardation of the flow of heat to the back of the board.

As mentioned in column 1, lines 39-48 in Patent No. 2,744,022, the vermiculite as it expands causes the core of the board to disintegrate due to spalling which takes place in a progressive manner so as to destroy the wall within a relatively short time This patent, as well as Patent No. 2,681,863 propose that the adverse effect of spalling be overcome by the addition of bodying fibers, such as those made from glass. While ceramic fibers enable the afore-mentioned unexpanded vermiculite to be used, they do not prevent the excessive expansion of the larger, unexpanded vermiculite particles which causes the core to become weak, limp and crumbly. This expansion is manifested by surface spalling when a gypsum cast containing such a vermiculite is subjected to high heat; and hence it spalling could be overcome and a stronger ultimate core formed, a marked advance in the art would ensue.

It is therefore an object of this invention to provide a cast formed from set calcined gypsum and unexpanded vermiculite which will not adversely spall and hence not form a weak, limp and crumbly core when subjected to fire.

It is an object of this invention to provide an unexpanded vermiculite of novel particle size distribution which is especially of value in improving the fire resisttance of set calcined gypsum products.

It is a further object of this invention to provide a /2" gypsum wallboard which when attached as a ceiling to wooden joists supporting an oak floor, such as disclosed in Underwriters Laboratories Design No. 42-1 hr., will enable such a construction to qualify for the one hour fire rating.

It is an added object of this invention to provide a novel means of preventing the formation of large fissures upon the surface of a gypsum wallboard which has been subjected to fire.

It is a still further object of this invention to provide a /2" gypsum wallboard which when used in a particular type of construction will enable such construction t have substantially the same fire rating as that of a /8" gypsum wallboard containing only glass fibers.

Various other objects will become obvious to those skilled in the art from the following detailed description.

The objects of this invention are accomplished by the discovery that very low fire shrinkage with practically zero spalling in a set calcined gypsum cast can be effected by a unique patricle size distribution of unexpanded vermiculite in such cast. It has been found that unexpanded vermiculite of a particle size range which will pass through a No. 50 mesh sieve and be retained upon about a 140 mesh sieve, preferably that passing a 70 mesh and retained from a mesh, U.S. Standard sieve sizes, will not permit an adverse spalling condition but will still control shrinkage or contraction when a gypsum wallboard containing such particles is subjected to fire. It is possible, however, to have a small amount of vermiculite particles slightly coarser than those which will pass a N0. 50 mesh sieve, but such should be less than about 10% of the whole and substantially all should pass about a No. 30 mesh sieve. Particles substantially smaller than those retained upon a 100 mesh sieve can be used as they do not cause spalling; however, they are not as effective in controlling the fire contraction as the coarser sizes nor are they as effective in decreasing the flow of heat. As will be explained in more detail later, these finer size particles contribute best in controlling contraction due to fire when used with plus 100 mesh particles. These smaller particles do prevent the formation of large surface fissures as will be subsequently set forth.

An unexpanded vermiculite ore, of a particle size distribution, similar to that known as B from the Travellers Rest deposit of the Zonolite Division of W. R. Grace & Company has been found to be very satisfactory in carrying out this invention commercially. Use of this ore will be described in detail later.

The importance of the particle size in preventing spalling, improving core integrity, and controlling expansion under fire conditions is clearly illustrated in Table No. 1. This table summarizes the results of a fire test of wallboards having set calcined gypsum cores. Each of these cores contained a different particle size range of unexpanded vermiculite from the Travellers Rest deposit of the above mentioned Zonolite Company.

These boards were all made as nearly the same as possible with the exception of the particle size range of the unexpanded vermiculite added to the calcined gypsum core composition prior to set. The composition used to make each board also contained substantially the same amount of certain of the well-known set accelerators, consistency reducers and bonding aids usually used in the commercial 4 Underwriters Laboratories publication 263, dated May 1959. A temperature of 1550 F. was reached in /2 hour, 1700 F. in one hour and 1850 F. in 2 hours at which time the test was stopped.

Each test however, was made upon the board only and not as used in a construction system. Great care was exercised to assure duplicate conditions for each test. While a number of observations were made during each test, the most important were the linear and thickness contractions of the board core and the core integrity..

The linear contraction is obtained by spring actuated dial gauges attached to probes extending into the furnace and pushing against opposite edges of the wallboard under fire test. The thickness contraction was determined by measuring the core thickness when cold before and after the test, allowances being made for the thickness of the paper. The integrity is ascertained during the test by visual observation of the board and by a study of the core, such as the change in core thickness, i.e. the thickness contraction at the conclusion of the fire test.

In actual construction under fire test, the linear contraction, if excessive, will cause underisable cracks to form in the board and/or openings to form between the boards and permit the hot gases therethrough. The core integrity is also very important as it is indicative of the way the board Will stand up under fire conditions.

TABLE 1.--THE EFFECT OF THE PARTICLE SIZE OF UNEXPANDED VERMIOULITE (NO GLASS FIBER) IN A THICK GYPSUM WALL- BOARD CORE UPON SUBJEOTION TO FIRE Particle size range of Max. linear unexpanded vermiculite contraction during test, Description of core Passing Retained 011 percent after test Board No.:

1 sieve sieve Disiutegrated in 9 mins. 40 sec. and

fell into furnace. 2 40 sieve. sieve Disintegrated in 12 mins. 40 sec. and fell into furnace. 3 50 sieve-.-" 70 sieve 025 Moderate spalling, core intact, good integrity. 4 70 s1eve 100 sieve 1. 8 Very fine shallow map cracks on surface. 5 100 sieve.. 200 sieve 51 Surface smooth, some small map cracks. 6 200 sieve 78 Similar to board No. 5,

onlysurface smoother. 7 50 sieve 100 sieve 17 Very fine shallow map cracks. Very slight spalling. Core intact. 8 B05 see Table 3 36 Very fine shallow map cracks. No spalling. 9 B04 see Table 3 Center of board fell into the furnace in 13 mins. 10 N o vermiculite 73 Very deep fissure, some as wide as .090.

manufacture of a gypsum wallboard in order to duplicate actual plant conditions as closely as possible. The same quantities of foam and water were used in all cases so as to obtain substantially the same core density. This along with the thickness was controlled so that the variations were no more than about -plus or minus 1%. No glass fiber or boric acid was used. Only unexpanded vermiculite was added to the gypsum core composition along with, of course, the usual board plant additives Well known to the art and referred to above. All of the boards contained the same amount of unexpanded vermiculite, i.e. 6% based upon the weight of calcined gypsum used to form the boards. The core composition was enclosed between the same type of paper surfaces as used in regular commercial wallboards.

These boards were all thoroughly dried in the same manner so that none had burnt or under-dried cores. They were fire tested in a horizontal position as a ceiling under very closely controlled, uniform conditions in a small gasfired laboratory fire test furnace, which was designed so as to approximate on a small scale the fire test of the Underwriters Laboratories. For each test the furnace temperature was held as closely as possible to the timetemperature curve relationship shown upon page 22 0f All tests were stopped after 120 minutes. All sieve sizes were US. Standard.

The above Table 1 clearly illustrates the critical importance of the proper particle size range of the unexpanded vermiculite in order to have good integrity of the board core along with practically zero contraction such as shown in the minus 50 plus 70 mesh range (board No. 3); the minus 70 plus mesh range (board No. 4); and the minus 50 plus 100 mesh range (board No. 7). Note the destructive spalling resulting from the use of the coarser size particles such as the minus 30-plus 40 and the minus 40-plus 50 ranges in the core. These are Within the size ranges for the unexpanded vermiculite which the above identified patents predicted would cause spalling and disintegration of the core unless glass fibers are also used.

Board numbers 5 and 6 show that particle sizes smaller than 100 mesh are not as effective in preventing shrink age, though they do not cause spalling. The effect of the sizes of the unexpanded vermiculite particles upon core integrity and shrinkage when boards containing different particle size ranges are subjected to fire is clearly set forth in a comparison of boards 8 and 9 made from two commercially available ores from Travellers Rest, South Carolina, i.e. B05, and B04 respectively. The former board contained unexpanded vermiculite within the contemplated range and had good core integrity, while the vermiculite used in the latter was outside the size range of the invention and hence the core disintegrated within 13 minutes.

In order to determine the efi'ect of both glass fiber and unexpanded vermiculite upon the properties of the core of a gypsum board when subjected to fire, boards were made and tested similar to those referred to in Table 1. These boards contained 6 lbs. of chopped glass fiber or .4% by weight, per 1000 sq. ft. of /2" gypsum board evenly distributed throughout the core along with the 6% unexpanded vermiculite. As followed in making the boards for Table 1, all percentages are based upon the amount of calcined gypsum used to prepare the board.

difference between Design No. 1*1 hr. and Design No. 42-1 hr. other than the former required a board and the latter the /2 board of this invention. It should be noted that the quantity of glass fiber in the /2" board of Design No. 42-1 hr. is the same quantity as is the A2" board of Design No. 1-1 hr.

The use of the thinner board not only saves material but also decreases the board weight with consequent savings in shipping cost and application labor. The gypsum wallboard used in the above Underwriters Laboratories fire tests employed a /2" wallboard made with regular production equipment. The core of the /2 wallboard contained 6% of the B05 ore referred to in Table No. 3 along with 6 lbs. per 1000 sq. ft., or about .4%, of chipped glass fiber roving of between to /2" in length evenly distributed throughout the core, and .75 of boric TABLE 2.THE EFFECT OF THE PARTICLE SIZE OF UNEXPANDED VER- MICULITE AND GLASS FIBER IN 6 THICK GYPSUM WALLBOARD CORE UPON SUBJEOTION TO FIRE Particle size range of Max. linear unexpanded vermiculite contraction during test, Description of core Passing Retained on percent after test Board No.:

11 sieve. 4O sieve 08 Very bad spelling but;

did not fall.

12 4O sieve sieve 058 Severe spelling and poor integrity.

13 50 sieve 70 sieve 033 Slight spelling, core intact. Surface map cracks.

14 70 sieve.-- 100 sieve 025 Shallow surface map cracks.

15 50 sieve 100 sieve 06 Very fine shallow map cracks. Very slight spelling.

16 B05 Ore 21 Very fine shallow map cracks. No spalling.

17 B04 Ore 19 Siggilar to board No.

18 N0 vermiculite 62 Deep fissure up to .050 wide. Core intact.

From the above, it is evident that the glass fibers aid in preventing a complete disintegration of the core due to spalling as mentioned in certain of the above identified patents. The glass fibers also reinforce the core after it has become weakened due to calcination. They also further aid in overcoming linear contraction.

This invention has been found to be most useful in the formation of gypsum wallboard cores whereby a markedly improved fire rating for constructions employing such wallboard is obtained. The practice of this invention has made possible an Underwriters Laboratories approval for a one hour fire rating on a construction referred to as Design No. 42-1 hr. This construction employs a one half inch thick gypsum wallboard containing chopped glass roving and uneXpanded vermiculite having particle sizes within the range comprising this invention. The Design No. 42-1 hr. comprises a gypsum wallboard ceiling under wooden joists with a wooden floor applied over the joists. This same board also qualifies for Design No. 41-1 hr., a construction similar to the above in which the board is furred from the joists with a resilient channel as set forth in Patent No. 3,090,164. This board also qualifies for Design No. 2212 hrs.

In the past a gypsum wallboard could not be used in constructions similar to Design No. 42-1 hr. A A" thick gypsum wallboard containing glass fiber but no vermiculite was required, i.e. Design No. 1-1 hr. referred to above. Except for minor details, such as the width of the finished floor and the size of nails, there is very little acid. A bond aiding material, a foaming agent and a set accelerator, all well known in the art were also used. All percentages are based upon the weight of calcined gypsum used to form the core. The finished /2 wallboard had a minimum weight of about 1925 lbs. per 1000 sq. ft. and afmrsimum average thickness across the field of the board 0 8".

As evident from the foregoing, chopped glass roving is not necessary to offset any distintegration of the core due to spalling when vermiculite particles within the particle size range of this invention are employed. However, glass fibers are always added to special fire resistant commercial gypsum wallboards so as to further improve core integrity thereby minimizing the possibility of failure due to stresses which the board may be subjected to as a result of fire. Good core integrity will enable the core to withstand the Underwriters Laboratories hose stream test. As is evident from a comparison of Tables 1 and 2, the chopped glass roving also functions as a means to provide a further reduction in the fire or heat shrinkage. This is also set forth in Table 6.

From the tabulated data the boric acid appears to decrease the fire contraction of the wallboard. The use of boric acid, however, is not necessary in all cases 'but when used at least .1% by weight of the calcined gypsum should be included.

While 6 lbs/M of chopped glass roving about /2 in length is used in the examples set forth in Table 2, at least about 1 /2 1bs./ M or about .1% by weight of the calcined 7 gypsum should be .used in order to attain noticeable results.

. 8 age of this board was .63% without glass fib'ers and .41% with. This clearly shows that the finer vermiculite TABLE 3.-PARTIOLE SIZE DISTRIBUTIQN OF UNEXPAN DE D VE RMIC ULITE v Underwriters Laboratories Specification l Sieve Silo L USS N 0. Preferred B 05' (Libby) B04 Min. Max.

Retained mesh (Tyler 20) 0 0 0 0 0 v 0 Retained 30 mesh (Tyler 28) 0 0.83 0 7. 9 0 1 Retained 50 mesh (Tyler 48) O 5. 95 3. 2 58. 0 0 9 Retained 70 mesh (Tyler 65) 0 21. 9 Retained 100 mesh (Tyler 100) 100 36. 95 54. 4 8. 4 35 65 Pan 0 55. 10 41. 9 3. 8 65 221-2 hrs.

Except where mentioned, all of the unexpanded vermiculite used was obtained from the Travellers Rest deposit of the Zonolite Company. It is not the intention however, to limit the invention to a particular source of vermiculite. Other sources such as the Libby, Montana deposit (L05) of the above identified company are also suitable for the purpose.

The amount of unexpanded vermiculite needed to carry out this invention depends upon the extent of the fire shrinkage which in general, is determined by the degree of expansion, the particle size range and the amount of impurities contained in the ore used. It is difficult to obtain a commercially available ore which is ideal. A preferred ore would be one which has a particle size range of between about 70 and 110 mesh and is free of impurities or non-expanding particles and has a high expansion when heated. The use of about 1% of such an unexpanded vermiculite in a gypsum wallboard will be noticeable; but for less heat contraction and hence more favorable results when subjected to fire a substantially higher amount should be used. However, such an ideal product is too costly, and in large scale manufacture only a commercially available ore can be used. Thus, instead of the ideal 70 to 100 mesh size, coarser particles such as those which will pass a 50 mesh sieve with some retained thereon (such should be less than about 10%) with substantially all passing about a No. mesh sieve may be included within the ambit of the invention. The latter vermiculite approximates the largest particle size range which is workable in the invention disclosed.

The commercial ore will also contain a substantial amount of particles passing a 100 mesh sieve; these small particles, especially those passing a 140 or a 200 mesh sieve, are much less effective in preventing the shrinkage during a fire test but they are present in commercial ores.

These smaller size particles appear to augment to a certain extent the contraction control of the larger or pins 100 mesh particles. For example, a laboratory made gypsum wallboard containing 6% of a minus 50 mesh sieve vermiculite ore of which 47.5% was plus 100 mesh or 2.78% based upon the same percentage relationship as 6% was found to have a heat shrinkage of 36% without glass fiber and .22% with. Another, similar board was made using the same type of unexpanded vermiculite but with only the plus 100 mesh portion of the 6% used, i.e., 2.78% of the minus 48-plus 100 fraction. The heat shrinkparticles have some value in shrinkage control when used with the coarser particles. For, when 6% of only aminus -plus 200 mesh fraction was used the contraction was .5l% and .78% with 6% of the minus 200 mesh sizes, both without glass fiber. While 6% of a minus 4 8-plus 100 mesh range had .a contraction of .22% without glass fibers and .06% with. V I

An unusual property of the finer sizes is that'while they have a high fire shrinkage close to that of .73% of a board with no unexpanded vermiculite added to the core, there are not the large deep fissures due to sintering as experienced when no unexpanded vermiculite is used. At least about 1%, based upon the weight of unset calcined gypsum used, of the plus 100 or mesh unexpanded vermiculite having the larger particle size limitation mentioned above is needed for noticeable results. Additional amounts of the unexpanded vermiculite such as in excess of about 2.5% are preferred. It should be noted that fine vermiculite particles heretofore generally discarded as waste are employed in the course of providing the various advantages of this invention.

The vermiculite should be one which expands well for otherwise an excessive amount will beneeded for purposes of this invention. The degree of expansion can be determined by suddenly subjecting the vermiculite, in small increments, to a high temperature, preferably over 1000 F., and observing the volume before and after expansion.

Practically all commercially available unexpanded vermiculite ore contains unexpandable particles. These can be determined by water classification of the expanded ore as the expanded vermiculite can be readily floated away. After the remaining particles have been dried, they can be weighed and from the known weight of the ore before expansion, the percent of unexpandable particles in the ore can be readily calculated.

The B05 (Travellers Rest) and L05 (Libby) ores referred to inTable 3,expanded 253% and 230%, with 12.1% and 24.1% of unexpandable material, respectively. The vermiculite was expanded by heating 100 gms. in 10 gms. increments in a 3 /2" x 5" x 2" deep pan at 1000 F. for 3 minutes. The pan, made of stainless steel, was equipped with a handle to enable it to be shaken during the heating operation. The lower volume increase of the Libby ore is probably due, primarily, to the extra amount of unexpandable material present.

9 The following Table 4 illustrates the effect of the amount of unexpandable particles and source of vermiculite upon fire linear contraction.

10 over 50% of the ore. If the ore contains a lesser percentage of the minus 100 mesh portion then a higher percentage of the plus 100 mesh portion will be needed. As is TABLE 4.EFFEGT OF SOURCE OF UNEXPANDED VERMICULITE AND UNEXPANDABLE PARTICLES UPON THE FIRE LINEAR CONTRACTION OF A $6 GYPSUM WALLBOARD Max. fire Pertinent ingredients added to the gypsum board linear core contraction Description of core after test Board No.:

19 6% vermiculite from Travellers Rest Deposit, .117 Slight shallow cracks on snrra B05, 12.1% Unex'pandable Bone Acid 6 lbs/M Glass Fibers. 20 and 21 Same as 19 except vermiculite from Libby used, 159 Similar to above but; a darkened LO 5, 24.1% unexpandable. areak with slightly deeper crac s. 22 Same as 20 and 21, excepting 10% of the Libby 067 D L ore and 1% boric acid were employed. 23 6% of the Libby L05 ore 6 lbs/M Glass F1ber. 150 Same but no darkened ar a.

1 Average. 2 Expandable.

Note that with the Libby ore having the higher percentage of unexpandable material more is required to obtain a low maximum linear contraction under fire conditions. Thus, 10% of the ore was used in board No. 22, though 6% of the Libby ore will provide satisfactory results.

The amount of expandable vermiculite needed can be determined by the extent of the fire shrinkage. The more that is added the less the shrinkage. It is not necessary or desirable to add an amount sufiicient to result in actual expansion. It is preferred that the heat shrinkage be not substantially over about .2%. A higher value under certain conditions may be satisfactory. If there is any question as to the suitability of a particular unexpanded vermiculite for the purpose, it is advisable to have the wallboard made therefrom tested by the Underwriters Laboratories. If more than of the minus 50 plus 100 mesh portion of the expandable part of an ore, based upon the weight of calcined gypsum used, is required to give a linear contraction of about .2% or less then it is evident from the above, the presence of glass fibers will influence the amount of the unexpanded vermiculite needed since they also decrease the linear contraction.

The ability of the properly sized vermiculite to not seriously spall yet effectively reduce linear fire shrinkage when used in the core of gypsum wallboard is an outstanding factor in obtaining the Underwriters rating for the designs mentioned above and is an important part of this invention. It should be understood that the presence of the vermiculite in the board core does not only aid in reducing fire shrinkage but it also, when expanded, retards the flow of heat by means of its heat insulation value.

While severe spalling is readily discernible and sometimes causes a complete disintegration of the core, its adverse effect upon the core integrity is sometimes insidious and a more reliable method of evaluation is needed. It has been discovered that an unexpanded vermiculite will give good core integrity when there is substantially no core thickness increase at the conclusion of the fire test, preferably a contraction. This is clearly set forth in the following Table No. 5.

TABLE 5.EFFECT OF VARIOUS TYPES OF UNEXPANDED VERMICULI'IE UPON CORE LINEAR AND CORE THICKNESS OONTRACTIONS OF )6 GYPSUM WALLBOARD DUE TO FIRE Fire Fire Glass Borle linear thickness fibers, acid, contractlon contraction, Unexpected vermiculite lbs/M percent percent percent Board No.:

8.. 0. 36 5. 9- 5. 58 16 0. 21 5. 58 17 I 0. 19 I 6. 35 7 0. 17 1. 15 6 0. 06 5. 93 24 6 0. 025 6. 60 25 08 0. 08 8. 10 26 6 0. 142 8. 20 1 e 0. 117 5. 79 2 6 1 0. 067 6.40 27 6 0. 175 9.10 6 0. 150 6. 9 2g 6 0. 200 6.

1 Disintegrated. 3 Expanded.

questionable that the source of vermiculite is satisfactory. However, most any commercial source of vermiculite is satisfactory.

As mentioned above the percentage of unexpanded vermiculite required to give a good linear contraction, such as no more than about .2% is best determined by a fire test such as that described above. However, in practice it will usually be found that in order to attain a linear contraction of less than about 0.2% that at least 2%, based upon the weight of the unset calcined gypsum used, of the minus 50 plus mesh portion of a commercial ore such as the B05 ore referred to in Table 3, along All of the boards referred to in Table 5 with a core thickness contraction had good core integrity. The core of these boards contained unexpanded vermiculite of a particle size distribution following this invention. Boards made from the B04 ore, which is coarser than that contemplated by this invention either disintegrated due to spalling, i.e. board No. 9, or had severe spalling along with a high thickness expansion, i.e. board No. 17.

From the above, it is evident that the uncontrolled expansion of the vermiculite not only causes spalling upon the surface but such progresses throughout the body of with, the minus 100 mesh portion which, as shown, is 7 the board with consequent weakening of the calcining board core. Hence, when there is excessive spalling upon the surface there is a good indication that the core integrity will be poor. Such cores are limp while the calcined board cores containing unexpanded vermiculite particles following this invention will be found to be solid. Whether the core is weakened by spalling, i.e. excessive increase in expansion of the thickness of the vermiculite, can best be determined by measuring the change in thickness of the core before and after subjection to fire. There should be substantially no increase in thickness, preferably a shrinkage.

The following Table 6 summarizes the effect upon the fire shrinkage of various materials added to the core of a gypsum wallboard.

While glass fibers have been used in most of the above examples it is also within the ambit of this invention to use other inorganicfibers, such for example as asbestos fibers. In the forthcoming claims the percentage of unexpanded vermiculite is the amount of expandable vermiculite present in the ore used, though in actual practice, as followed in the above examples, the source of ore may have an appreciable amount of unexpandable particles present. Also, all sieve sizes are US. Standard and all percentages are based upon the weight of calcined or settable gypsum used to form the core. All the boards used in obtaining the above data were made with the rotary calcined gypsum manufactured at the East Chicago plant of the 'United States Gypsum Company and is that used 5 GYPSUM WALLBOARD Percent maximum contraction during fire tests with- 6 lbs/M Glass 1% fiber Amosite plus asbestos Uuexpanded No 6 lbs/M 4% and vermiculite other Glass boric boric size addltive fiber acid acid minus 50 plus 70 mesh 025 .033 Exp 6% minus 70 plus 100 mesh.- l8 025 6% minus 100 plus 200 mesh 51 6% B05 Ore .36 21 117 1 217 No vermiculite..-. .73 62 48 l A similar board but made with glass fiber instead of asbestos shrank .142%.

In the above table the glass fibers are shown as also contributing, to a certain extent, to the control of fire contraction. This is not as much the case where such is used with unexpanded vermiculite of the coarser particle size ranges as set forth in certain of the above-indentified patents in which the glass fibers primarily function in the nature of a lattice-like structure as a means of supporting the calcining core containing the expanding larger vermiculite particles. Without this lattice-like structure the entire core was shown to disintegrate and fall into the furnace. As used in the instant invention, the glass fibers serve primarily as an added reinforcement to the calcincd core when it is subjected to the stresses developing during the fire. Although the core without the glass fiber remained stiff and intact it is believed that glass fibers impart additional strength to the core enabling it to further resist the stresses developed during the fire.

While the subject of this invention is best practiced in the preparation of a gypsum wallboard of improved fire resistance, it is not intended to limit the scope thereof to such a particular preferred example. The improved unexpanded vermiculite particles can be used with other set calcined gypsum products which are subjected to high heat and in which it is important that there is a minimum of heat contraction without spalling, such for example, as in gypsum block used for partitions or in certain types of metal casting plasters.

A further use for the unexpanded vermiculite following this invention is in the form of a dry calcined gypsum plaster mix which when added to water can be applied to surfaces such as over steel beams and girders to give improved fire protection. As in the case of the use of the unexpanded vermiculite in a gypsum wallboard following this invention, the larger size particle limitation of the vermiculite should be such that at least about 90% of all the unexpanded vermiculite used must pass a No. 50 mesh sieve, with substantially all retained thereon passing about a No. 30 mesh sieve. At least 1% based upon the Weight of dry calcined gypsum used in the plaster, of such vermiculite should be retained upon a No. 100 mesh sieve. Other additives, such for example as those which control the set and consistency to effect certain improvements, well known to those skilled in the art, may also be used.

in the manufacture of regular commercial gypsum wallboard. By calcined gypsum is meant any calcium sulfate type product which will harden when mixed with water such as is known to those skilled in the art, of which this invention is a part.

The foregoing will so fully explain the character of my invention that others may, by applying the skill of the art to which this invention appertains, readily adapt the same for use under varying conditions of service, while retaining certain features which may properly be said to constitute the essential items of novelty involved, which items are intended tobe defined and secured to me by the following claims.

I claim:

1. A gypsum cast suitable for forming the core of a /2" gypsum wallboard which will qualify for an Underwriters Laboratories one hour fire rating, which cast comprises set calcined gypsum having distributed therethrough unexpanded vermiculite particles in an amount in excess of 2.5% by weight of said calcined gypsum, said vermiculite having a particle size distribution such that substantially all of said particles will pass a No. 30 mesh sieve, at least about by weight thereof will pass a No. 50 mesh sieve, and at least 2% by weight, based on the calcined gypsum, will pass a No. 50 mesh sieve and be retained on a No. mesh sieve, said vermiculite being present in an amount sufiicient to preclude substantial contraction of said cast upon being gradually heated to a temperature of 1850 F. and limited to an amount not in excess of about 15% and not sufiicient to produce linear expansion of said cast upon such heating or to produce thickness expansion after being cooled, whereby to avoid spalling and surface fissures in said cast.

2. A gypsum cast as in claim 1 wherein said unexpanded vermiculite is present in an amount of about 6% by weight of said calcined gypsum.

3. A paper covered gypsum wallboard having a core comprising a cast as in claim 1.

4. The product of claim 1 in which mineral fibers in the amount of at least about 0.1% by weight, based on calcined gypsum, is also dispersed throughout the gypsum cast.

5. The product of claim 1 in which boric acid in the amount of at least about 0.1% by weight, based on calcined gypsum, is also distributed throughout the gypsum cast.

6. A gypsum cast suitable for forming the core of a /1" gypsum wallboard which will qualify for an Underwriters Laboratories one hour fire rating, which cast comprises set calcined gypsum having distributed therethrough unexpanded vermiculite particles in an amount in excess of 2.5% by weight of said calcined gypsum, said vermiculite having a particle size distribution such that substantially all of said particles will pass a No. 30 mesh sieve, at least about 90% by weight thereof will pass a No. 50 mesh sieve, and at least 2% by weight, based on the calcined gypsum, will pass a N0. 50 mesh sieve and be retained on a No. 100 mesh sieve, said vermiculite being present in a quantity sufiicient that said gypsum cast exhibits a linear contraction between about 0.0 and 0.36% upon being gradually heated to a temperature of 1850 F. and no thickness expansion after being cooled.

7. A gypsum cast as in claim 6 wherein said vermiculite is present in a quantity sufficient that in said cast exhibits a linear contraction between about 0.0 and 0.2% upon being gradually heated to a temperature of 1850- F.

References Cited UNITED STATES PATENTS 1,732,737 10/1929 Wiggin et al. 106-110 2,526,066 10/1950 Croce 106110 2,744,022 5/1956 Croce et al. 106110 2,853,394 9/1958 Riddell et a1 161-162 3,190,787 6/1965 Muller et al. 106-109 TOBIAS E. LEVOW, Primary Examiner.

I. E. MOTT, Assistant Examiner.

US. Cl. X.R. 106109, 110

"H050 UNl'llill) S'lA'lES Pfi'lENT OFFICE CERfiiiiCfi'ilS 01* C031; BCTlC-N 3, 5 6 D t d Julv 8 1969 e Grant S. willeyf It is certified that error appears in the abovc-identified patent and that said Letters Patent are hereby corrected as shown below:

a, Column 2, line 24, "patriole" should read --{partic1e Column line 22 "'underisable" should read undesirable line 23, "board should read boards line 2 4, after "gases" insert to pass In Table l, for Board N0. under the heading ."Description of core after test "1.8 should read .18 Column 6, line l t, "chipped" should read chopped Column- 7, line 43 "110" should read 100 Column 10, In Table 5,.- for Board No. '9, unde the heading "Fire thickness contraction, percent", delete 5.58" 'In Table 5, for Board No. 16, under the heading "Glass fibers, lbs./'l l",---"2'--- s-hou1d-rea .d--- '6 Column 1 line 1, after "that" cancel "in" SIGNED KND SEALED APR 28 I970 Attest:

Edward M. Fletcher, In. Attesting Officer