Fire barrier coatings

DESCRIPTION

Technical Field

This invention relates to a fire barrier 5 composition, a substrate coated with the composition, a laminate basqd on the coated substrate and a method of making the composition.

Numerous methods have been developed for controlling the ravages of fire. The rationale 10 for each of these methods is found in the physics of combustion, the chemistry of flame, and the engineering of fire control systems. The disclosure herein presented treats with the use of coatings as a fire control method; a barrier system intended 15 to prevent the ignition of and spread of flame along a combustible substrate.

Background Art

The prior art discloses four major types of fire-protective coatings: (1) ablative;‘ (2) 20 intumescent; (3) sublimation and gas- or vaporproducing; (4) ceramic or inorganic. Ablative coatings are generally thick and heavy and costly, thus limiting their application to special uses. Intumescent coatings are generally soft and easily 25 abraded, and have limited duration of protection -- 10 to 15 minutes when exposed to fire. Sublimative or gas-producing coatings have even shorter duration of protection -- after their snuffing gases have been produced and wafted away from the surface, 30 they are left without a protecting mechanism. They have found their major use in the treatment of fibers and clothing. Ceramic coatings, as the name implies, require high-temperature curing in order to form the ceramic bond - which many structural or building components cannot withstand. Inorganic coatings, such as Portland cement, gypsum, calcium aluminate cement, phosphate-bonded cement, metalpigment loaded silicate coatings (sodium, potassium, ethyl, etc.), high-temperature silicone, and magnesium oxychloride and magnesium oxysulphate cements have all been proposed. This disclosure treats specifically with magnesium oxychloride in combination with calcium aluminate cements and other high-temperature resistant binders.

It is remarkable that in the long history of oxysalt cements, so few attempts have been made to convert these materials into a fire barrier in the form of thin paints. No doubt this is, in part, due to the prior major shortcomings of the products and the inability of the developers to overcome them. Amongst these shortcomings are: brittleness and rigidity, along with the proneness to spalling or decrepitating under beat. Unless mixed in stoichiometric ratios, the result is unsightly efflorescence (white bloom) of the gauging salts (magnesium chloride, magnesium sulphate, magnesium carbonate) which continually condenses onto the surface of the coating. In other cases, especially in the oxysulphates, the excess liquids tend to synerese (squeeze out) and leave a greasy wet film at the surface. The coatings exhibit poor aging and weatherability characteristics, as well as solubility in water and rain, thereby making them unusable in outdoor situations.

(The rain and high humidity leach the chloride and sulphate from the formulation, leaving void spaces as the result of the leaching process, which in turn leads to a disintegration of the coating.) They have a strong tendency to be hygroscopic, picking up moisture from the air, and thereby presenting a continuously wetted surface; they show poor resistance to common solvents; and there is a loss of tensile strength and cohesiveness after exposure to flame, causing cracking which exposes the substrate to ignition and subsequent combustion. In addition, some coatings are relatively soft and easily scratched and abraded, and unable to resist scrubbing (oxysulphate coatings are markedly softer than oxychloride coatings). Thin coatings, although they adhere well to substrates, frequently shrink in drying, and show mud-cracking (crazing), thus exposing the underlying substrate to direct flame impingement.

Thus, there exists a need in the art to develop a durable, quick-setting, non-combustible thin coating capable of protecting a variety of susceptible flammable substrates from ignition ( (fire-barrier), one which at the same time has all the attributes of a good paint and, additionally, aids in overcoming the above-listed shortcomings in the prior art. Another need is to develop a fire-barrier coating which does not spall or decrepitate or crack under the heat of flame and thus expose the underlying substrate. A further need is to develop an inexpensive Class A fire-retardant coating as defined by the requirements of the National Fire Protection Association (NFPA), utilizing inorganic non-toxic components, and based on magnesium oxide-magnesium chloride oxysalt chemistry. Other needs are to develop a coating which would bond to wood, plywood, and gypsum (plaster wallboard, to fiberglass board and batting, 5 and to other substrates, imparting to them fire-barrier properties and increased flexural strength when incorporated as an element of the V laminate structures. In addition, a barrier coating must possess a suite of other physical 10 characteristics: hardness, compressive and tensile strengths, cohesiveness, adhesion, and a variety of other properties, including cosmetic attributes, which keeps the coating's integrity under the stresses of fire exposure, so that it does 15 not crack, spall, flake off, or otherwise lose its ability to act as a fire barrier. The coating itself must be noncombustible, so that it does not contribute to the fuel, nor aid the progress of a flame front.

Disclosure of Invention

This invention aids in fulfilling these needs in the art. This disclosure crosses over and unites four major technologies: as a composition of material, it falls into the category of 25 air-setting hydraulic binding cements, belonging to the group of caustic magnesia cements, generally called Sorel Cements, and more specifically as magnesium oxychloride cements; as a product it falls into the category of a coating or paint; 30 as a use or application, it falls into the category of fire-control devices; and its associated characteristics permit it to be used for -- and thus categorized as -- a material of construction.

Specifically, presented herein is a descrip35 tion of a novel magnesium oxychloride cement coating, co-bonded with high alumina calcium aluminate cement and a colloidal silica, which overcomes the numerous deficiencies in the prior art coatings of this type, and the application of this coating, either alone, or in combination with a non-woven spunbond polyester fabric or a non-woven or woven glass fabric in the form of a thin laminate (composite coating) to fire susceptible substrates to form a fire barrier capable of preventing ignition and flame spread of the structure. The coating is non-combustible in itself, and has adequate f tensile and cohesive strength to maintain its integrity for the duration of the fire. It has excellent adhesion to a wide variety of substrates, and can be included as a structural component in the construction of gypsum wallboard, plywood, and other types of laminates.

Prior art magnesia cement compositions consisted of: (a) the binder ingredients; (b) the activator ingredients; (c) fillers; and (d) adjuvants.

(a) The basic binding ingredient, a cement, consists of a lightly calcined (heat-activated) magnesium oxide (MgO) which, when combined with the aqueous MgC12 or MgSO4 salt solution, crystalIlzes into a network of interlocking crystals (and other ionic bonding systems) that gives strength to the system by bonding the fillers contained therein. In the present disclosure, high-alumina calcium aluminate cement and colloidal silica are added as compatible co-binders, imparting to the magnesium oxychloride cement important beneficial characteristics, as detailed below.

(b) An aqueous activator fluid (called a gauging liquid) which is a salt solution of magnesium chloride (MgC12) or magnesium sulphate' (MgSO4), or a combination thereof. Both salts are hydrated salts (6 H20 and 7 H20, respectively) easily soluble in water. Practice shows that to be effective in bringing about the setting of the MgO, the solutions must be concentrated (407□ to 6070), generally expressed as a specific gravity, in either specific gravity or in degrees Baume units (sp. gr. 1.22 to 1.26, or, 26° to 32° Baume). These salts are best added in a stoichiometric ratio for the reaction, which does not leave an excess of salt that can effloresce to the surface or leach out in water or humid conditions; or react with the C02 in the air to form compounds which cause a disruption of the structure by occuping more space than originally occupied in the initial setting of the cement. In the present disclosure, the salt'solutions are used as the common hydrating agent for both the MgO cement and the calcium aluminate cement. As demonstrated below, the use of this solution to hydrate the monocalcium aluminate obviates the necessity of maintaining an optional stoichiometric ratio of salt to MgO.

(c) Fillers -- any of a wide variety of types of inert materials, such as sand, gravels, crushed rocks, silica flour, pumice, vermiculite, volcanic ash, perlite, wood shavings, sugarcane bagasse, asbestos, mineral fibers, etc. are added to impart special properties or lower the overall cost of the concrete.

(d) Adjuvants - a wide variety of accessory chemicals have been added to the slurry mix to bring about specific changes in the product, e.g. to make the concrete harder, to speed up the setting, to slow down the setting, to prevent the magnesium salt efflorescence, to waterproof the composition, to prevent shrinking (and hence shrinkage cracks), to prevent the absorption of atmospheric moisture, ¹to decrease harshness in the mix and impart spreadability (smoothness), to decrease water or fluid demand, and etc.

As in most hydraulic cementitious systems, the binding ingredient or cement is intimately blended with the dry filler materials and any dry adjuvants and stored until needed. The dry ingredients are then mixed with the liquid ingredients, and the slurry poured or placed. The chemical reaction takes place, which sets the cement and forms the concrete. The mass, which is fluid at first, gradually thickens, becomes less workable, forms a gel (initial set) which gradually and continuously becomes harder over a four-hour period. The hardening continues and strength increases for several days so that by the third to fifth day maximum strength has been reached.

The coating, subject of this disclosure, consists of a two-part mixture: Part A, the dry powdered binder and filler components; Part B, the gauging or activator liquid components. The binder contains magnesium oxide, with specific characteristics, and high alumina calcium aluminate cement, along with fillers (silica flour, titanium dioxide, and, optionally, expanded vermiculite and alumina trlhydrate). The liquid component consists of magnesium chloride solution of 1.26 specific gravity, colloidal silica dispersed in dimethyl formamide (DMF), and an anionic surfactant, sodium tetradecyl sulphate. As is common practice in the art, part A is added to part B, with adequate mixing to form a slurry of the proper consistency for painting or coating, and the slurry permitted to set into a hard monolithic refractory 5 layer. When used as a composite or laminated coating, either a non-woven spunbond polyester fabric, or a non-woven or woven fiberglass fabric is interposed between two thin layers of the coating. The coating or composite coating is applied to the desired 10 substrate. When exposed to flames of up to 2000°F, the coatings retain their integrity and prevent the flame front from advancing. They themselves do not burn, and they prevent the direct ignition and flaming of the substrate for as long as 30 15 minutes -- sufficient time, in most cases, for the original fire source to die out or bypass the protected area. Although they are efficient thermal insulators, as long as any water of hydration remains in the coating they are not designed to be long-term 20 heat barriers, hence, after varying periods of time when wood substrate temperatures reach 700° to 800°F, wood substrates will char or carbonize, but do not ignite or burn. ^,Similarly, plastic substrates, foam or sheet or block depending on their 25 particular time - temperature response, will char but not ignite, and thus do not contribute to the advancing flame front.

While the excellent high temperature and fire-resistive properties of coatings made from 30 magnesium oxysalts have long been recognized, relatively little advantage has been taken of these in attempting to adapt thin coatings to commercial use as fire retardants or fire barriers, due, in large measure, to major deficiencies in physical properties. There is widespread use of these materials in thick layers, 1/2 to 4 or even 6 in thickness, as insulative and fire-barrier mastics, grouts and cements. Although these thick insulations are frequently termed coatings, this disclosure treats only with the use of magnesium oxychloride concrete as a thin paint or paint laminate.

Best Mode for Carrying Out the Invention

The coatings and coating laminates described herein have the function of preventing the spread of fire by defensively protecting combustible substrates from igniting, preventing the flame from spreading and extending its area of destruction along the substrate, and thus prevent the accumulation of the combustible (and highly toxic) gases in the structure. Preventing is an absolute term which would signify fire-proofing. However, fireproof, non-combustible and even fire-resistant are all ambiguous terms unless qualified by a description of the fire conditions under a recognized test method. The American Society for Testing Materials (ASTM) does not recognize the term fire-proofing since nothing, under the most severe conditions, cannot be fired. Commercially, the term is used to indicate the slowing down or delaying -־ and occasionally preventing -- of the ignition and spread of flame of the substrate underlying the protective coating or cladding.

Currently available commercial fire-retardant coatings, such as intumescent-type fire-retardant paints which when exposed to high temperatures form an insulating blanket, are intended to function as their designation implies, to delay ignition for about 10 to 15 minutes, and reduce the surface burning rate of interior combustible finish materials. They do not render the underlying materials non-combustible, but serve only to retard 5 ignition for a brief period of time. That delay furnishes an extra measure of escape time, affords time for implementation of other fire-fighting methods, and reduces structural damage. In contrast, the fire-barrier coatings, subject of this disclo10 sure, are designed to prevent ignition and firespread of the underlying substrate by interposing an incombustible layer between the flame and the substrate. Fire-retardant' coatings are rated as to flame spread when tested under standard widely 15 accepted test conditions (e.g. ASTM E-84, Underwriters' Laboratories [UL] #723, National Fire Protection Association [NFPA] #25, or ASTM D-3806).

In development work, for comparative evaluation of candidate coating formulations and assess20 ment of the numerous variables which enter into such formulations and compositions, ASTM Standard Test Method D-3806, Small-Scale Evaluation of Fire Retardant Paints (2-foot Tunnel ' Method is preferred to the much more elaborate and time-consum25 ing 8-foot tunnel test (ASTM E-286) and 25-foot tunnel test (ASTM E-84). D-3806 determines the protection a coating affords its substrate and the comparative burning characteristics of coatings by evaluating, the flame spread over the surface 30 when ignited under controlled conditions in a small tunnel. This establishes a basis for comparing surface burning characteristics of different coatings without specific consideration of the end use parameters that might affect surface burning character35 istics under actual fire conditions. In addition to the experimental flame spread rate, the test determines the weight of the panel consumed, fuel contributed to the fire, time of afterflaming, smoke emission and other characteristics. The Standard measures and describes the properties of the coatings in response to heat and flame under controlled laboratory conditions, but it warns that it should not be used to describe or appraise the fire hazard of assemblies under actual fire conditions. However, the test may be used as an element in fire risk assessment.

The test consists of impinging directly onto the surface coating a 2000°F gas flame supplying heat at 5085 BTU's < per hour for 4 minutes, while the 2-foot long by 4-inch wide test panel is suspended above the burner, coated surface facing downward, at an upward angle in a chamber. The angle of the test panel and the draft of the chamber encourages any flame in the coating and substrate initiated from the burner to proceed upward along the substrate. The extent of this flame front is measured and compared with the known substrate standards (0 and 100).

Flame spread is the ratio of the measured extent of the burning of the substrate protected by the candidate paint, contrasted with a known incombustible rated standard substrate (cement-asbestos board) as 0 and an unprotected highly combustible wood substrate, such as Red Oak, with a rating of 100.

Mo re specifically, Flame spread is a ratio, not inches. The maximum extension of the flame on the sample is measured. The maximum extent of the flame (flame spread) of the asbestos board is considered as 6.5, with an adjustment factor of 4.45. The length of the asbestos board flame is subtracted from the flame length of the test specimen, and the difference is multiplied by the adjustment factor to yield the flame spread of the test sample. For example, in Table II, Test j.2 of the Type II coating alone on isocyanurate foam board, the flame spread on the specimen was 7 long, minus 6.5 length for asbestos board, which yields an 0.5 difference, and when multiplied by the 4.45 factor, equals the reported flame spread of 2.2.

Percent Loss is the total weight of the specimen prior to the test minus the total weight of the specimen after the test, x 100. This includes loss of moisture due to steam vaporization, as well as losses due to combustion. In the case of the coatings, subject of this disclosure, most of the loss in weight when applied to susceptible substrates is due to this water-of-hydration loss from the coating under the area of flame impingement, without significant deleterious effects on the coating.

Twenty readings are taken in determining smoke density, and summed. This is divided by 2,000. The value of Red Oak, considered a standard with a value of 1924, is subtracted, and the resultant is multiplied by 100 to convert the specimen's value into a percentage which relates to the value of the Red Oak as 1007־ and asbestos board as 07־.

Flame spread is considered the main factor associated with testing fire retardant coatings. The fire-retardant coatings currently available are rated as low as 10, or as high as 60 to 70.

The lower the rating, the longer the coating will retard flames. In the higher ranges, e.g. 60 to

70, the coating affords little or no fire protection.

The coatings and laminates described herein, which are rated 0 flame spread (as well as 0 fuel contributed to the fire and 0 smoke developed) are a significant advance in the art of protective coatings in that they are fire barriers, not merely fire-retardants. They are designed to interpose a noncombustible, coherent, durable layer between a flame and a combustible substrate, and thus prevent the ignition of the substrate and and the subsequent flame-spread along the substrate. Under operational fire conditions, following a period of time, the initial fire source will have exhausted its local fuel supply and burned itself out. * w *

TABLE I lists the several formulas for the fire barrier coatings developed in this disclosure. The numbers in parentheses listed below correspond to the numbers in the Materials column of TABLE

I.

The first five items of part A are all finely comminuted solid powders, while the sixth is a granulated solid of -16 mesh. Each material is first dried in a hot air oven at 200°F, mixed together, and, except for item (6), intimately blended by mulling. The surfactant (9) is added drop-by-drop to the MgC12 solution (7) and the colloidal silica (10) is added slowly to (7) + (9). This causes an opalescence in the clear solution, but no precipitate. The reaction is highly exothermic. There is a marked increase in the viscosity of the solution, which decreases again upon standing. The mixed powders, (1) + (2) + (3) + (4) + (5) + 6), are then added to the liquid phase (part B) 7) ־) or (8) + (9) + (10), slowly, with constant mechanical stirring, so that each particle is completely wetted out by the liquid. A smooth suspension of paint-like consistency results. This is applied, while still fluid, by any common method of coating application (brushing, rolling, spraying, etc.). Pot life is short 20 to 30 minutes, with the paint gradually thickening until it becomes a soft gel. The gel gradually hardens with initial set at 1 to 1 1/2 hours, and final set in 2 to 4 hours. Pot life and set time can be controlled by varying the solids to liquid ratio -- most easily by increasing (or decreasing) the amount of the fluid phase. Final set time of the coating is significantly decreased by elevating the cure temperature to 100°-120°F. In these formulations, the final set of the coating, that is, the hardness and other surface characteristics, is relatively insensitive to the solids: liquid ratio. Upon drying, the coating is a brilliant white; porcelain like; refractory hard (Mohs 5.5-6.0); and crackfree, with essentially zero shrinkage. After 48 to 72 hours of cure time, the coating is highly resistant to direct flame impingement, and withstands 1900° to 2000°F without cracking for 30 to 45 minutes and longer, when suitably supported on a fibrous matrix (non-woven spunbond polyester, woven or non-woven fiberglass fabric, fiberglass fibers,

5 etc.) which absorbs the thermal stresses of differential expansion and crystallographic alterations accompanying the firing.

* A .k

TABLE I: TYPE FORMULAS - FIRE BARRIER COATINGS

Materials

I II III IV

Part A (Binder Components -- Powders)

<td> (1) Magnesium Oxide (MgO)</td><td> 400 gms</td><td> 400 gms</td><td> 400 gms</td><td> 400 gms</td>

<td> (2) High Alumina Calcium Aluminate Cement</td><td> 100 gms</td><td> 100 gms</td><td> 100 gms</td><td> 100 gms</td>

<td> (3) Silica</td><td colspan="4"></td>

<td> Flour</td><td> 100 gms</td><td> 100 gms</td><td> 100 gms</td><td> 50 gms</td>

<td> (4) Titanium Dioxide (TiO₂)</td><td> 15 gms</td><td> 15 gms</td><td> 15 gms</td><td> 15 gms</td>

<td> (5) Alumina</td><td colspan="4"></td>

<td> Trihydrate</td><td> —</td><td> - -</td><td> - -</td><td> —</td>

<td> (6) Expanded Vermiculite (-16 mesh)</td><td></td><td></td><td></td><td> 70 gms</td>

Part B (Gauging or Activator Components -- Liquids) (7) Magnesium Chloride (MgCl₂) (1.26 sp.gr.) 380 cc. 440 cc. 380 cc. 490 cc.

(8) Magnesium

Sulphate (MgSO₄) (1.25 sp.gr) (9) Niaproof

Surfactant 4.0 cc. -- 4.0 cc. 4.0 cc.

(10) Colloidal

Silica (357־ in

DMF) 80 cc.

- ר 1

TABLE I: TYPE FORMULAS - FIRE BARRIER COATINGS (Cont'd)

<td> Materials</td><td> V</td><td> VI</td><td> VII</td>

<td colspan="2"> Part A (Binder Components --</td><td> Powders)</td><td></td>

<td> (1) Magnesium Oxide (MgO)</td><td> 400 gms</td><td> 400 gms</td><td> 400 gms</td>

<td> (2) High Alumina Calcium Aluminate Cement</td><td> - -</td><td> ר .— *</td><td> 150 gms</td>

<td> (3) Silica Flour</td><td> --</td><td> 200 gms</td><td> --</td>

<td> (4) Titanium Dioxide (Ti02)</td><td> 15 gms</td><td> 15 gms</td><td> 15 gms</td>

<td> (5) Alumina Trihydrate</td><td> 100 gms</td><td> --</td><td> 50 gms</td>

<td> (6) Expanded Vermiculite (-16 mesh)</td><td> 50 gms</td><td></td><td> 70 gms</td>

<td> Part B (Gauging or</td><td colspan="2"> Activator Components</td><td> - Liquids)</td>

<td> (7) Magnesium Chloride (MgC12) (1.26 sp.gr.)</td><td> 400 cc.</td><td> - -</td><td> - -</td>

<td> (8) Magnesium Sulphate (MgSOz,) (1.25 sp.gr.)</td><td> - -</td><td> 400 cc.</td><td> 560 cc</td>

<td> (9) Niaproof 4 Surfactant</td><td> 4.0 cc.</td><td> 4.0 cc.</td><td></td>

<td> (10) Colloidal Silica (35% in DMF)</td><td></td><td> 80 cc.</td><td></td>

(1) Three types of MgO (magnesia) have been used in the principal binding system which differ in reactivity as a result of originating from different sources and receiving differing heat treatments. In the formulations listed in Table I, OXYMAG grade from Basic Chemicals, Inc. was used -- a calcined magnesia ore from Nevada, with 1.5%“ lime (CaO), 95% MgO, 3% loss on ignition, and'with 99.6% passing 200 mesh. It is treated to eliminate any tendency to expand due to the free lime. Its bulk density is twice that.of seawater-derived magnesias, but has great binding power for fillers and high reactivity with MgC12 and MgSOz,. Alternatively, both MAGOX 98 HR and MAGOX 98 LR magnesias (also from Basic Chemicals, Inc.) have been used with equal success in the fire barrier coating, but for different applications. These MAGOX materials are calcined magnesia from sea water (brine) sources. HR refers to high reactivity; LR refers to low reactivity. Both contain 97% to 98% MgO, 1% lime (CaO) and only 0.15% Fe2O3, so that they are white. Due to the method of calcination, the specific surface area of the HR grade is 68 M²/gm, while the LR grade shows 36 M²/gm. Both have a median particle size of less than 2 microns. Both have been used successfully -- the only modification required is the quantity of MgC12 solution necessary to make a paint of the required consistency. Since their densities are lower than the OXYMAG, they bulk more (30 lbs/ft³ vs. 48 lbs/ft³) and a greater volume of material, with a larger surface area, is required to yield the weights shown in the type formulation. This requires approximately 2070 to 25% greater fluid volume to wet out and produce a fluid paint. Where open pot time is more important than rapid initial set, the LR grade can be used. On a volume-of-paint basis, the MAGOX grades are less expensive than the OXYMAG grade. Other 5 producers, such as Martin-Marietta Corporation, produce a variety of sea water (brine) magnesia grades (which correspond to the HR grades of Basic), MAG-CHEM 40, a very high reactive grade; MAGCHEM 30, a high reactive grade; and MAG-CHEM 10, a low reactive grade.

The order of blending and mixing is of some importance. The solids are added to the fluid mixture with constant stirring, so that each powder particle is wetted out and no lumps remain. After mixing 15 is complete, the paint, using the HR grade of MgO, has a pot life of 15 minutes before thickening to a point which makes brushing or rolling difficult. Using the LR grade, the pot life is extended to 20 to 25 minutes, with the final setting extended 20 to 3 to 4 hours .

(2) High Alumina Calcium Aluminate Cement: Both ALCOA (Aluminum Company of America) CA-25 (High Alumina Refractory Cement) and LONE STAR LaFARGE, INC. SECAR 71 or 80 have been used as 25 a co-binding system with the magnesium oxy cements, with equal success. The materials are equivalent. The principal constituent of these HAC (high alumina cements) is mono-calcium aluminate, (CaO + 2.5 AI2O3). They are hydraulic cements containing a 30 high percentage of alumina (AI2O3), approximately 80%, and calcium oxide (CaO), approximately 17%-18%. (In contrast, Portland cements are calcium silicates; Plaster-of-Paris cements are calcium sulphates). Water is the generally-accepted activating agent,

Ο which causes the essentially anhydrous mixture (only 1.870 loss on ignition) to form a strong interlocking crystalline bond in setting. Setting is rapid, initial set taking place in 1/2 to 1 5 hour, and final set in 4 to 5 hours, and maximum strengths are achieved in one day. Useful service temperature of the set concrete is 3200°F. It is widely used in castable refractories. The HAC lose strength at intermediate temperature ranges of ]_q 900°-1200°F, but achieve much higher strengths through ceramitization at the 1500°F and higher temperature ranges. They resist the corrosive action of calcium and magnesium sulphates and chlorides. The cements are white; they are finely ground, with a specific surface of about 10,000 cm^/gm. (Blaine). Compressive strength of a 1:3 mortar at 7 days is about 9400 psi, while neat cements, depending on the cement:water ratio, can achieve 8,000 to 18,000 psi.

2q The strengths developed with these high alumina cements are significantly greater than when lower grade calcium aluminate cements are used, e.g. U.S. Steel Corp., Atlas Cement Division’s LUMNITE cement, with 47.0% AI2O3, 34.3% 25 CaO and 7.4% Fe203, or REFCON cement, with 58.5% AI2O3, 33.570 CaO and 7.47□ Fe203, both of which are brown in color; or Lone Star LaFarge FONDU cement, with 38%-407_o AI2O3, 397_O CaO and 157־ Fe203.

The calcium aluminate cement is not to be confused with alumina (AI2O3), frequently called out as either a high strength filler in cemented products, or as a binder in refractories and ceramics. In the latter case, it functions as a high-strength and high-temperature binder only after high-temperature fusion in a furnace.

In the listed formulations the water of hydration for the monocalcium aluminate is supplied by the magnesium chloride (or magnesium sulphate) solution. Magnesium hydroxide, the principal ingredient of the mag-oxy cements, according to one current theory, retards the setting of the alumina cement; but the magnesium sulphate high concentration accelerates the set. Similarly, magnesium chloride in low concentration retards the set, but in high concentration accelerates the set. In the formulations, listed in TABLE I, the entire body of cement!tious materials develop an initial set in 1 1/2 to 2 hours, and a final set in 3 to 4 hours.

The set cement is highly resistive to the leaching action of water, and is impervious to the corrosive action of sulphates and chlorides (both of which have severe effects on Portland Cement Concrete).

The addition of the high alumina calcium aluminate cement as a co-binder is completely compatible with the MgO binder system, and adds considerably to the hardness and abrasion resistance of the coatings, as well as improving resistance to leaching of the magnesium salts in the set cement by rain, thus improving the waterproofness of the product. In addition, it effectively prevents the efflorescence of the dried magnesium salts and the syneresis of the soluble salts.

' (3) Silica flour. A standard commercial commodity available from many producers, consists of finely ground (smaller than 200 mesh) silica sand (silicon dioxide, Si02) low in iron, glistening white. It is used as an incombustible low-cost filler. It affords rheological (flow) control to the fluid paint, while supplying hardness and abrasion resistance and innumerable sites for crystal stress relaxation in the cured films, especially at elevated temperatures when the silica goes through its own phase changes.

(4) Titanium Dioxide (Ti02). A standard commercial commodity available from many producers. This ingredient, used at a relatively low concentration, functions as a whitening agent --- not only decoratively, but more importantly as an infrared radiant heat reflector. Ti02 has great hiding power, and imparts a brilliant white to the product, obviously of very low emissivity (a), and high spectral reflectance (ξ ) in visible light, but also at infrared wavelengths. Since burning gases and flames emit a high percentage of their heat as radiant heat, the low emissivity and high reflectance of these coatings contribute to lowering the surface temperature on the coating and its underlying substrate.

(5) Alumina Trihydrate (AI2O3.3H2O) with _רapproximately 357־ bound water, is widely used as a fire retardant filler for plastics. Its mode of action is similar to that of the hydrated magoxy cements. The hydrate retards flame by serving as a heat-sink, absorbing heat in the process of decomposition, and liberating water vapor (steam) which serves as a gaseous blanket protecting the substrate. The grade used in the type formulas was SB 331 from Solem Industries, Inc., Norcross, Georgia, with a mean particle size of 10 microns, and TECHFILL A-112, with a mean particle size of 12 microns, produced by Great Lakes Mineral Co.

of Troy, Michigan. Both products are brilliant white with high specular reflectivity. The alumina trihydrate is used to substitute for the silica flour as a filler. While the coatings it produces are not as hard as the silica flour coatings, this decrease is not significant, and is compensated for by the additional water of hydration it adds.

(6) Expanded Vermiculite. Because of its low K factor and non-combustibility, expanded vermiculite has long been used for pour-in insulation, insulating concrete, and a wide variety of high-temperature insulations. Its fusion point is rated at 2200° to 2400°F, and it has a specific heat of 0.20. When added as a filler to the part A binder components, while its contribution to the heat-barrier properties of the coating is not insignificant (since it occupies, by volume, almost 20% of the fluid paint, and 257□ of the dry paint) its main function is to serve as a stress absorber. The grade of expanded vermiculite used in Type Coating IV is size No. 3, with a density, of 5 to 6 pounds per cubic foot. This commercial grade is screened through a -16 mesh sieve, and only the material passing through the mesh is used. Domestic U.S.A, vermiculite tends to be dark brown in color, while South African grades tend to be a light cream color and require lesser Ti02 to hide the off-color it imparts to the coating. In plant manufacturing, the particles should be slightly wetted with the MgC12 (or MgSOz!.) solution prior to adding to the total mix, but for field application, it is added as a dry ingredient to the rest of the powders in part A, and ennobled by these powders in the mixing. The particles disperse quite readily in the mix, and there is little tendency to segregate. In the dry paint, they impart to the coating .a pebbly surface, which contributes to the reflectivity of the coating.

Thus, in painting the particles are evenly dispersed.

During exposure to flame at 1800° to 1950°F, the crystallite components of the oxychloride cement undergo phase changes which result in compression stresses on the integrity of the film. The vermiculite particles are not subject to these phase changes at those temperatures, and retain their integrity. They are thus capable of serving as stress-relaxation points. Due to their lamellar structure with air voids between the layers, the particles are compressible, and thus take up the stresses of the thermal expansion which accompanies the phase changes of the cement components. This prevents the popping (spalling or decrepitating) and subsequent cracking of the coatings common to many rigid and brittle ceramic films. As noted in the examples presented below, not all formulations require the inclusion of vermiculite particles, and they are not normally added to those coatings included in the laminates made with fabric interlays, which take up the stress points in a similar manner.

(7) MgC12 Solution in Water - 1.26 sp. gr. (30° Baume). Dow Chemical Co., Tech, Flake Grade, supplied in a dry form 'as 99.0% MgC12.6H2O. The material is a thin white opaque flake, easily soluble in water, but high deliquescent. The hexahydrate material contains 46% to 52% MgC12. The hexahydrate is a commodity article of commerce supplied by many chemical manufacturers. Because of the hygroscopicity of the hexahydrate, it cannot be packed dry with the MgO and other solid materials into a one-package magnesium oxychloride cement product.׳ Formerly covered by ASTM Specification #C 276 (now discontinued), the Dow Chemical Co. Product Data and Sales Specification is generally accepted as the standard. The MgCl₂.6H₂0 is soluble to the extent of 155 mg/100 cc distilled water at 20°C. A 28% solution based on MgCl₂ yields 30° Baume at 60°F. In practice, 1500 gms of the hexahydrate salt dissolved in 1000 cc of city tap water yields 1800 cc of a 1.26 sp. gr. (30° Baume) solution, with a pH of 6.0. It is recommended that the solution be allowed to mature for 24 hours prior to use; and it is stable indefinitely, after preparation, in closed containers. Practice shows that a minimum concentration of a sp. gr. of 1.18-1.20 is required to successfully bring about the SOREL cementitious reaction. The more concentrated the solution, the harder the final product. Too great a quantity of MgCl₂ leads to efflorescence under humid conditions. Both the salt and solution are corrosive to ferrous metals, so precautions must be taken in the preparation of the salt solution. Also, because of the hygroscopicity, the flake salt must be kept from exposure to the air.

This solution, in the proportion disclosed in the formulas listed in TABLE I, is the activator or gauging solution, used to wet out and hydrate the MgO and the calcium aluminate cement, and is the principal liquid vehicle for the aqueous-based fire-barrier paint.

(8) Alternatively, the gauging solution may be MgS04 at a specific gravity of 1.25, or 29° Baume. This is the common Epsom salts, (MgSO4.7H2O, a flaked, colorless crystalline material), a commodity product, produced by many chemical manufacturers. Formerly covered by ASTM Specification #C-277, this specification is now discontinued. The hydrated salt is readily soluble in water and highly endothermic - chilling the water in the process of dissolving. 500 gms of the salt, dissolved in 1 liter (or 4.169 lbs in 1 U.S. gallon) of water yields a sp. gr. of 1.25 or 26° Baume. This makes a net 41% solution. In tap water, the resulting pH is 5.3. The reaction between the MgO and MgSO4 is well known in the art of making fire-resistive mastics (and stucco), but the resulting concrete is noticeably softer than when made with MgC12. Also, the product shrinks considerably more, forming surface and deep cracks. This tendency to shrink is counteracted by the calcium aluminate cement and/or by the addition to the sulphate solution of phosphate salts or MgC12 salts.

Because of the lower hygroscopicity of the MgS04.7H20, it is frequently packaged with the MgO into a one-component oxysulphate cement base requiring only the addition of water. However, because the preferred formulas listed in TABLE 1 contain other fluid ingredients besides the salt solution, they require a 2-component packaging system. A third alternative formulation calls for the mixing of the above-formulated magnesium chloride solution with the magnesium sulphate solution in proportions of 1:1, 2:1, or 3:1. This formulation hardens the product, and reduces the tendency to shrink and crack.

(9) Niaproof Anionic Surfactant 4, supplied by Niacet Corp., Niagara Falls, N.Y. 14302, is a 277־ aqueous solution of sodium tetradecyl sulphate, [C4HgCH(C2H5)C2H4CH(SO4Na)CH2CH(CH3)2]. Whereas many surfactants (dispersants) were precipitated by the high magnesium salt solution, or cause a coagulation or gelling of the colloidal silica, or cause objectionable foaming during the mixing of the paint, NAS 4 met all compatibility requirements. Its function is to reduce the surface tension of the liquid phase (B) of the formulation, and thus enable it to wet out the solid particles more efficiently. It acts as a dispersant for the powders in the formula. As such, it reduces the quantity of fluid required to make a satisfactory paint viscosity (the water of convenience), and thus yields a stronger paint film, with less tendency to shrink-crack (craze) while excess fluid evaporates.

(10) Colloidal Silica, or silica sols. NALCOAG 2325, a 357־ colloidal dispersion of silica in the non-aqueous polar solvent dimethyl formamide (DMF), from Nalco Chemical Co., Chicago, Ill.

The particles of silica are discrete, non-crystalline, spherical, submicron (average particle diameter of 20 millimicrons; average surface area 150 M^/gms). The dispersion contains 357־ silica as Si02- When diluted with water, the DMF dispersion yields a pH of 5.

The dried particles of the silica develop strong adhesive and cohesive bonds. Under high temperatures of up to 200°F, they show little crystallographic transformation. In the Type I and Type VI formulas in which used, the colloidal silica performs the function of a densifying agent; a waterproofing agent;

a hardening agent; a surface gloss-producing agent (porcelainizing); a co-bonding agent along with the magoxy and calcium aluminate cements; and an adhesive aiding bonding of the coating to non-porous substrates. Colloidal silicas have been widely used as binders for fillers in the production of high-temperature mineral and inorganic fiber shapes (principally thermal insulation, e.g. APOLLO spacecraft heat shields); as frictionizing (anti-skid) agents’ in coatings; as catalysts; as anti-tacking agents in polyolefinic ־films; in protecting coatings to increase abrasion resistance of decorative laminates; as polishing agents (e.g. silicon wafers) etc.

By far the most commonly used type of dispersion is the aqueous, negatively charged sodium ion-stabilized form, usually supplied as 157«, 307,, 407־ or 507□ Si02 concentration, depending on the particle size of the silica. All are highly alkaline. Other types are stabilized with ammonium ion or chloride ion (positive particle charges); and instead of a suspensoid in water, are dispersed in acidic solutions or oils, or coated with aluminum acetate, etc. Colloidal silicas are produced by DuPont Industrial Chemicals Division of E.I. DuPont de Nemours; by the PQ Corp, (formerly Philadelphia Quartz Co.); by Nalco Chemical Co., and by several others in the U.S.A., as well as in Japan, Germany, England, etc.

While all types of colloical silicas were not tested in the development of the listed formulat 1,0ns, five types were tested, with sufficiently differing results (mainly, comparabilities amongst the various other fluid and powder ingredients) to lead to the selection of the DMF as the preferred type of dispersion. When 30% and 40% alkaline aqueous dispersions of the commonly used sodium ion stabilized form are added to the high salt content MgC12 or MgSO4 solutions, the silica precipitates out and 5 leaves a soft semi-solid gel, containing numerous silica clumps. Alumina acetate coated silica sol (NALCOAG ISJ-613) is aqueous based and completely miscible with the liquid phase of the listed formulations. 10 Containing 3% colloidal alumina, plus 19% colloidal silica, ISJ-613 is acidic, and has a pH of 4. the AI2O3 and Si02 particles are positively charged, the ISJ-613 yielded excellent porcelain-like hard waterproof coatings, but the paint is harsh, i.e. 15 while the fluid paint is smooth, it shows poor flow and brush-mark fill in. It also reduces the pot life of the paint. However, it markedly accelerates the

<td></td><td> initial</td><td> and</td><td> final setting of the coating to</td><td> as</td><td> low</td>

<td> 20</td><td> as 1 to In</td><td colspan="2"> 1 1/2 hours. the type formulation I, calling for</td><td> the</td><td> use</td>

<td></td><td> of the</td><td> DMF</td><td> dispersed colloidal silica, it</td><td> may</td><td> be</td>

<td></td><td colspan="2"> noted that</td><td> the addition of the NALCOAG 2325</td><td> to</td><td> the</td>

magnesium chloride solution generates considerable heat - a highly exothermic reaction ־ but while it 25 causes an opalescence in the salt solution, there is no precipitation. There is an increase in solution viscosity, which is compensated for by the addition of a surfactant, NAS 4.

Colloidal silicas should not be confused with 30' other types of silicas or silicates. They are significantly different from soluble silicates (e.g. sodium, potassium, lithium or other alkali metal silicates); organic silicates (e.g. ethyl silicate); fumed silica (Cab-o-Sil', Aerosil); microfine silicas; amorphous silicas; silica gels; silica flour or silica powder.

* * *

To accomplish the objectives listed above, the basic SOREL cement reaction was modified at several critical points. In order to increase the fluidity of the paint so that it could be brushed, sprayed, or rolled onto the substrate, with good flow and fill-in to eliminate streaks, the ratio of activator or gauging fluid was increased beyond that customary in the art. At such ratios of fluid MgO to solids (approximately 1 cc of a 1.26 sp. gr. MgCl₂ solution fluid per gram of MgO), the customary result would be excessive syneresis of the gauging fluid resulting in an oily surface, followed by crystallization of the salt, depositing a heavy efflorescence of white powder, followed in turn by hygroscopic adsorption of atmospheric moisture, leaving a continuously wet surface. The novel formulation removes the criticality of maintaining the correct stoichiometric balance between the magnesium salts and the magnesium oxide in order to prevent the efflorescence and syneresis, since the high alumina calcium aluminate cement, utilized as a co-binder, in its own hydration/crystallization reaction binds all the excess salt solution. The formulation facilitates the adding of sufficient fluid to permit ease of working, and smooth coating flow without the deleterious effects of excess activator solution. Excess fluid normally results in high film shrinkage and unacceptable crazing of the coating. The formulations herein presented, by utilizing the salt (

solutions as the water of hydration for the calcium aluminate cement have completely obviated this problem, and the coatings are crack-free. They have remained crack-free during almost two years of aging under exposure to the fumes and fluorescent lighting of an industrial chemical laboratory.

The surfactant (9), Niaproof 4, by wetting out the powdered binders and fillers, and performing this function in the high salt content milieu, decreases the fluid requirements for a workable formulation, thus contributing to the strengthening of the coating.

In the setting process, the coatings show essentially zero shrinkage, and in addition to the absence of shrinkage cracking, they do not lift off from the substrate. They exhibit excellent adhesion to metal, wood, fiberboard, concrete, wall paper, fiberglass fabric and board, clay and ceramic tiles, asbestos cement, and many plastics -- including plastic foams. The coatings do not adhere to fluoroplastics nor to polyester film (however, they adhere tenaciously to polyester fabrics). The separation from polyester film is complete, imparting a smooth, glossy surface to the coating. This release paper characteristic is made use of in lining the surfaces of molds into which the formulations are cast; or by covering the coatings with the polyester film so that when the film is stripped off, it leaves a smooth glossy surface. They can be made to bond to the films by using an intermediary tack coat of an acrylic latex or other adhesive.

The coatings set equally well whether covered by the polyester film or exposed to air.

The addition of the calcium aluminate cement and/or colloidal silica, both acting as co־binders, imparts a surface hardness (Mohs 5.5-6.0) and abrasion resistance exceeding that of prior art oxychloride coatings, and brings to the normally much softer oxysulphate coatings a hardness (Mohs 4.0-5.0) almost equal to that of the oxychloride coatings. (It should be noted that one of the major drawbacks to the use of oxysulphate floors and walls, stucco, fireproof mastic coatings, despite the many advantages of this material, resides in its softness. The use of calcium aluminate cement overcomes this failing).

A high alumina calcium aluminate cement (70% to 80% AI2O3) is preferred to the cements which contain lower concentrations of alumina 47% to 58%). These latter are brown in color and detract from the high reflectivity white which is desired in the product, but, more importantly, contain much higher concentrations of calcium oxide (CaO), 34% to 39%, versus 17%, which has a disruptive effect on the crystallites of the cement at high temperatures, and leads to catastrophic cracking of the coating when exposed to flame.

Novelty also rests in the־use of colloidal silica, dispersed in dimethyl formamide (10) as a compatible co-bonding system with the other two co-bonding systems (1 + 7) and (2 + 7). Thus the system contains three bonding systems, all of which have differing characteristics, and retain their cohesiveness at different temperatures. The magoxychloride loses its water of hydration at 570°F, and starts to calcine at 1000°F. The colloidal silica is stable to 2000°F, and the calcium aluminate cement to 3200°F, with transitory weakness at 800° to 1200°F while it undergoes phase transformation.

Coatings I - V are self-levelling and with final set, usually in 2 to 2 1/2 hours, they are brilliant white, smooth, hard (Mohs 5.5 -6.0), dense, difficult to scratch except with pointed steel objects, and capable of faithfully reproducing the finest detail of a mold. They develop a high gloss if cast against a smooth plastic film (e.g. polyester or polypropylene) or a plastic mold without the need of a release agent. If cured with the surface exposed, type I's surface stays glabrous (shiny), but the others develop a matte finish. The coatings may be tinted or colored by the addition of water- or oil-based mineral pigments, such as the Hoechst Corp. Colanyl pastes (yellow, brown, green, etc.); and the coatings may be overprinted by silk-screening or other printing methods to give a decorative pattern. They may be covered with wallpaper, with the coating serving as a high temperature resistant adhesive, if the paper is applied while the paint is still in the gel stage!

The coatings are resistant to water leaching and to the action of many polar and non-polar and aromatic solvents. They are abrasion resistant. They are crack-free, age without noticeable changes, and exhibit little shrinkage in curing.

As discussed above, the T1O2 imparts a brilliant whiteness to the coatings, which reflects most of the incident infrared radiation of a flame, and thus contributes to the lowering of the surface temperature of the substrate. Most of the heat transferred by a flame is radiative (radiation). While flat black paint will absorb 92-977־ of this incident radiation, a white glossy paint will absorb only 217־ and reflect the other 79%. This difference between a white reflective surface and a dull absorptive surface can mean a lowering of several hundred degrees surface temperature in the former. NASA has determined that increasing the reflection of radiant heat from 307־ to 80%F resulted in a decrease of surface temperature from 2300°F down to 1900°F. Thus, the high radiation reflective surface imparted by the glossy white coatings described in this disclosure should operate at considerably cooler tempertures and require a longer period of time before their protected surfaces reach ignition temperature.

Another novel feature of these types of formulations, compared to the prior art, is the use of colloidal silica dispersion in DMF (10), which further increases the hardness, density, abrasion resistance, and hydrophobicity, and waterproofs the coating, further decreasing any tendency for rain to leach the magnesium salts. It also prevents the coating from picking up moisture from the air. The colloidal silica imparts ,a high sheen and porcelain-like surface to the. coating.

The coatings can be applied by brushing, spraying, dipping, roller coating, silk-screening, or by any other convenient method. Several layers of coating can be applied, on top of one another. The freshly applied layers show excellent bonding to the older base layer.

The coatings can be texturized by adding sand, gravels, vermiculite, expanded perlite, etc.

The formulations, currently, because of the short shelf life after mixing, must be packaged as a ready-to-use two-component system: Part A, the blended powdered solids; and Part B, the liquid components, to be mixed at volumetric ratios which correspond to the metric weights shown in TABLE I.

When coated onto both surfaces of a free-standing woven or non-woven natural fiber or synthetic resin fabric or fiberglass fabric, the cured coatings impart complete fire-retardancy to the fabric. Under the

1950°F temperature and high velocity of a propane torch for 30 to 60 minutes, the fabric does not combust, and there is zero flame-spread from the area of the cone of impingement of the flame. The organic fabric pyrolyzes into a carbonized skeleton within the ennobling coating, and the fiberglass melts, but the structural strength of the coating holds the fabric together; nor does the flame penetrate to the opposite side of the barrier. The fabric serves both as a stroma or supporting framework, and as a network of stress-relief foci. The numerous minute fibrils interlock with the crystallites as they form, leaving compressible points at which the volumetric changes associated with the heat-engendered phase changes in the cements have an opportunity to release their thermodynamic energies. They also serve to increase the tensile strength of the coatings.

The thermal dissipation of polyester or glass fabric is not as high as a metallic screen would be, it is better than that of the magnesium oxychloride cement itself, and thus reduces localized temperature build-up. No optimum or preferred fabric porosity has been established, but should be in the range of 50% to 95%. The H0ESCH Trevira spunbond grade 1110 polyester fabric used in the examples given below weighs 3.3 oz. per square yard, has a porosity of 85%, and has a melting point of 485° to 540°F.

A variety of laminating fabrics has been used; non-woven spunbond polyester (geotextile); woven polyester; nylon; jute (burlap); cotton, wool; steel wool; aluminum and carbon steel, -16 mesh window screening, woven cloth (graphite), cloth's woven aramid (Kevlar) fabric; etc. Materials of choice are the woven fiberglass, where it is desired to keep to a minimum smoke and gases generated by the pyrolysis of organic materials, and the spunbond polyester, which imparts desired physical properties (high tensile strengths) until called upon to serve as a barrier in an actual fire.

This laminate system of a coating on both surfaces of a fabric can be laid on a susceptible substrate (wood, fiberboard, composition board, plastic laminates, foamed plastics, etc.) with similar results. The coating on the undersurface acts as a high-temperature bonding adhesive for the fabric-to-substrate interface. The paint layer on the top surface forms the refractory-hard coating. If the coating laminate is laid between two mating substrates, the top layer also acts as another high-temperature bonding adhesive. Whether on the surface of one fire-susceptible substrate, or bonded between two substrates, the coating laminate serves as a fire barrier. The coating laminate adds considerable tensile strength to the structure. Since the coating material fills the voids and spaces within and between the fibers of the mesh of the fabric, the laminate carries more weight.

The efficiency of the supporting textile in the laminate is due to a combination of: 1) fabric porosity (i.e. the volume of voids divided by the total fabric volume, measured as a percentage); 2) the heat resistance or melting point of the fibers; and 3) the thermal conductivity of the individual fibers. This latter allows for a thermal gradient to be established between the hot spot immediately beneath the flame, and cooler areas further away from the flame. Since the duration of effectiveness of a fire barrier is in part a function of the quantity of hydrate material available for release of water vapor, a laminate is a better insulator than the coating alone. Although two or three coats of paint alone can be applied to a surface to build up a dry thickness of 15 to 18 mils, a single laminate layer can build up to 25 mils thickness, while two or three layers of laminate can build up to 1/8 or more -all possessing the characteristics of the paint: refractory hard and abrasion resistant, and brilliant white; capable of serving as a fire barrier for 20 to 30 minutes when exposed to a flame of 1950° to 2000°F.

Both the coatings alone, and the coating laminates -- single or multiple -- can be sawn, nailed, stapled, or fastened with screws, without fracturing.

Instead of using woven or non-woven fabrics for increasing the tensile strength and serving as the stress-relief face for the coatings, a variety of types of chopped fibers could be used to serve the same functions, e.g. glass, steel, polyester, aramid, nylon and other synthetic and natural materials .

EXAMPLE I

Type I formula (TABLE I) intended for outdoor use, was made by the method indicated in the teachings (vide supra). In this formula, the colloidal silica (10) not only becomes a densifying agent and a waterproofing agent, but also becomes a co-bonding agent along with (1) and (2). Since the viscosity increases fairly rapidly, and the pot-life of a flowable coating is 15 to 20 minutes, batch size is limited to the quantity which can be applied in that period of time. It may be noted that when the colloidal silica (10) is added to the magnesium chloride, the reaction is highly exothermic. An opalescence develops, along with an increase in viscosity, which is reduced by the addition of the surfactant (9). The coatings were brushed onto a variety of substrates: 6 x 6 fired ceramic tiles; 3 x 6 concrete cylinders; 6 x 15 of woven burlap (jute) fabric, and HOECHST Trevira non-woven needle-punched spunbond polyester fabric, grade 1110, 3.3 oz. per square yard; 6 x 15 x 14 white pine board; 6 x 15 x_3/8 plywood. The fabrics were coated on both surfaces. However, the coatings do not adhere to polyester film (Mylar) and peeled off cleanly, leaving a smooth glossy surface.

The coatings started to gel in 1 hour, and initial set took place in 90 minutes. Final set varied from 2 1/2 to 3 hours. The dried, fully-cured paint has a density of 1.56, equal to a weight of

97.4 lbs/ft³.

All coatings dried crack-free and refractory hard, resisting scratching by a steel knife blade (Mohs hardness 6.0). The coatings were rigid, and, in thin sections, brittle, brilliant white, smooth, highly glossy, and have remained unchanged in appearance for 20 months up to the time of the preparation of this disclosure. Unlike the other coatings -- II, ΙΠ, IV and VII, the types I and VI, which contain colloidal silica, do not pick up moisture from the air, and the weight of the dry coating remains constant. Since the coatings were brush-applied, they varied in thickness from 4 to 6 mils thick, when dry. The samples were submitted to a battery of tests. The coatings were permitted to cure for 1 to 2 days prior to being exposed to a standard laboratory flame test.

Flame Tests: Using a needle-tip propane torch of 1950°F flame temperature, so that the tip of the blue cone of the flame played directly onto the painted surface from a distance of 3/4 to 1, the coated burlap fabric and polyester fabric, mounted vertically, were separately exposed to the flame for 15 and 30 minute periods. Tests were replicated four times with similar results. When tested on the uncoated base fabric, the burlap Ignited and burned, creating a flaming torch within 1 second. The polyester fabric ignited within 1 second and melted, and became self-extinguishing when the flame was removed, but in the continued presence of the flame, burning continued until the fabric was reduced to a molten mess. The coated surfaces were rendered fire-proofed under conditions of this test. There was zero flame spread. The 1 diameter area exposed to the flame first became red hot, then white hot, and stayed in this condition for the full 15 and 30 minutes of the test. The fabric beneath the coating pyrolyzed and outgassed only in the area directly under the flame and in a penumbral circle of an additional 1/2 to 1 diameter. The coatings themselves did not crack, and while they did calcine, they retained their integrity and the source flame

־ did not penetrate through the opposite side. Paper and other flammable materials held against the back of the fabric, directly under the flame impingement, carbonized but did not ignite. The flame front 5 did not extend beyond the area of direct flame exposure.

Two layers of paint were applied on consecutive days to the wood panels (white pine board and plywood) with the result that the dried coatings were 10 approximately 8 mils thick. The adhesion between the two coats was excellent, and no line of demarcation could be noted. When exposed to the propane torch test, if the full energy of the 1950°F flame was applied immediately to the coating surface, ! 15 the superficial layer would pop or decrepitate, due to the rapid volatilization and expansion of the contained residual moisture (not water of hydration). Bringing the surface temperature up gradually over a period of 1 minute, by first playing 20 the flame over a larger surface area, obviated this decrepitation. For burn periods of 15 minutes, the results on the two boards were similar. The coatings held their integrity and there was zero flame spread of the substrate from the area immediately beneath 25 the flame. The coating calcined only in the 1 diameter area directly under the flame and in an additional 1 penumbra (total diameter of approximately 3). The substrates, after an initial

5-minute period, showed signs of carbonizing (pyrolysis), giving off thin wisps of smoke, but did not ignite or combust. The depth of the carbonization gradually continued for the full 15 minutes before the first indication of slight browning reached the back of the 1/4 thick white pine board.

Smoke density increased up to the 10 minute mark, after which it reached almost zero. At no time did the substrate flame, nor was there a flame-front. The area of pyrolysis of the substrate was limited to that of the area of calcination of the coating (total 3 diameter). The degree of pyrolysis varied with the depth. At the end of 15 minutes while carbonization was complete immediately beneath the coating, there was a gradual diminution until at the rear of the panel there was only the initial indications of carbonization. The plywood panel smoked more densely due to the use of organic resin and adhesives between the layers, but also showed a lesser degree of pyrolysis. In contrast, the uncoated boards both flamed within 1 minute, and the flame front extended up the board unchecked while being fired, and the flaming continued after the removal of the activating flame.

The coatings continued to adhere well to the substrate, without indication of debonding. The calcined area of the coating could be scraped away, but did not peel away or flake off of its own accord.

This series of tests indicated that the coatings themselves were non-combustible, and were capable of acting as a barrier to the ignition of susceptible wood and fabric substrates, and thus preventing the extension of a flame front.

In preparation for a certified ASTM D-3804 two-foot tunnel test for flame spread, smoke density, and fuel contribution to be run by an independent test laboratory (Applied Research Laboratories of Miami, Florida), a 23 3/4 x 3 7/8 x 1/4 white pine panel was coated with Type I paint, and a layer of HOECHST Trevira non-woven spunbond geotextile fabric #1110 to form a single laminate. As indicated previously, such a construction consists of a base layer of paint applied to the bare wood, a strip of the polyester fabric of the same size as the board, rubbed or rolled in with a rubber roller or a squeegee, in order to saturate the fabric and a top layer of paint applied immediately. The fabric weighed 6.8 gms; the pine board, 150.9 gms; while the laminate picked up a total of 205 gms of paint (dry weight) equal to 2.23 gms of paint per square inch. The total coating laminate had a thickness of approximately 25 mils. The dried paint was brilliantly white, refractory hard, and crack-free.

Results of the ASTM D-3804 test, with a full 5 minute burn (instead of the customary 4 minute burn) at 2000°F, showed: Flame spread 17.8; Fuel contributed = 0; Smoke density 56.6; Percentage loss in weight 11.8%. The control white pine panel showed: Flame spread = 73.0; Fuel contributed = 0; Smoke density 578.9%; and Weight loss = 74.64%. The coated panel was barely scorched on the surface; there was no significant penetration of the coating or laminate, while the bottom half of the uncoated control was completely consumed, and the upper half reduced to a badly cracked, completely charred stick of charcoal. The smoke density and percentage loss in weight in the coated panel is principally due to the moisture vapor (steam) driven off during the heating of the coating -- as it is supposed to be -- in protecting the underlying susceptible substrates .

The coating adhered tenaciously to the 3 x 6 concrete cylinder, leaving a smooth, glossy refractory-hard highly abrasion resistant impervious coating. Untinted, the coatings were a highly reflective bright white. The addition of minute quantities of HOECHST Colanyl pigments to the basic Type I formula permitted the coloring of the concrete coatings to be bright green, yellow, brown and black.

The coated ceramic tiles were weighed and immersed in tap water for periods of 7 and 14 days, after which they were dried and reweighed. There was no noticeable loss of weight, nor changes in the physical properties of hardness, nor gloss, which would indicate increase in porosity due to leaching of soluble salts.

A series of aggressive solvents was flooded over the surface of the coatings on the ceramic tiles, and each permitted to stand for 15 minutes, the excess drained, and the remainder dried off with paper toweling. The coatings did not soften, and showed only a minor loss of ׳ surface gloss to some of the treatment. Resistance to the following solvents was thus demonstrated: SOLVESSO 150 (a highly aromatic hydrocarbon); ACCUSOL S/T (207. ethylene glycol in monoethyl ether); Copier׳ machine mix (a high-flash mineral spirit); MOBAY SOLVENT BLEND (4 parts cellulose acetate + 1 part xylene + 1 part methyl ethyl ketone (MEK) + 1 part butyl acetate); kerosene, acetone and xylene.

The fluid paint was poured into a patterned plastic cup with raised printing, and rotated. The excess was poured off, leaving a thin shell mold. Upon drying and removal of the cup, the 6-mil thick layer of paint was not only smooth, brilliant white and crack free, but showed a remarkably faithful reproduction of the finest detail of the mold.

Thus the Type I coating not only performs a role as a fire barrier, but in addition has the necessary attributees of a good paint.

EXAMPLE-II

Where outdoor use and exposure to rain is not contemplated, the colloidal silica (10) may well be omitted from the formulation, and Types II and III of TABLE I utilized for interior use. These differ only in the use of the surfactant (9) dispersing and wetting the binder powders of part A so well that this permits a reduction of almost 15% in the amount of gauging fluid required for an equivalent paint viscosity. Type II paint has a pH of 8.5 and weighs 14.26 lbs. per gallon (1.71 gms/cc). When the respective coatings are cured, there is little difference between them. Neither shows any cracking nor appreciable shrinkage. Both exhibit refractory hardness (Mohs 5.5), brilliant whiteness, high gloss and excellent adhesion to various substrates. Hence the results of the testing the fipe barrier properties under different configurations -- will be treated together. Omitting the colloidal silica exposes the coating to moisture pick-up from humid atmospheres, and the cured coatings show an increase in weight until psychrometric equilibrium is established.

During a two-year period, over 100 batches of Types II and III formulations were made up to coat several hundred types, sizes, and configurations of test specimens. The only difference in the formulations was the grade of MgO (1) utilized. The . prior discussion covers the characteristics (vide supra). There was no noticeable difference

ר- 45 in the final results of the cured coatings. However, due to its longer open-pot time, Basic Chemical

Co.'s Magox 98 LR is the preferred grade.

The various configurations of test specimens 5 are categorized below; specimens made up for in-house fire-exposure screening tests and formal specimens made to the standards of ASTM D-3806 and submitted to an independent testing laboratory (Applied Research ¹⁰ Laboratories of Miami, Florida) for certified flame testing in their Two-foot Tunnel.:

a] Coating applied to bare fabric.

1. Non-woven spunbond polyester > 2. Glass fiber cloth (woven fiberglass)

3. Woven jute (burlap)

x. Fabrics 1 and 2 were also used as the fabrics forming part of the coating laminate.

xx. Generally, only a single coating 20 laminate was utilized; occasionally, a double coating laminate was applied. Multiple laminates, consisting of 3 to 5 or more layers may also be constructed to form a rigid panel construction of 1/4 to 1/2 thickness, or

<td> 25</td><td> even up to a 1 thickness.</td>

<td></td><td> b] Coating and/or coating laminate applied to white pine board. c] Coating and/or coating laminate applied to laminated plywood.</td>

<td> 30</td><td> d] Coating and/or coating laminate applied to 5/32 wood panelling. e] Coating and/or coating laminate applied to thin one-ply doorskin.</td>

<td> 35</td><td> f] Coating and/or coating laminate applied to cedar roof shingles.</td>

<td></td><td> g]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

<td></td><td> to</td><td> compressed</td><td> fibergl</td><td colspan="2"> ass board.</td><td></td>

<td></td><td> h]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

<td></td><td> to</td><td colspan="2"> gypsum wallboard.</td><td></td><td></td><td></td>

<td> 5</td><td> i]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

<td></td><td> X to</td><td colspan="2"> polyurethane foam</td><td> board.</td><td></td><td></td>

<td></td><td> j]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

to isocyanurate foam board.

<td> -</td><td> k]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

<td> 10</td><td> to</td><td colspan="3"> polystyrene foam board.</td><td></td><td></td>

<td></td><td> 1]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

<td></td><td> to</td><td> corrugated</td><td colspan="2"> paperboard.</td><td></td><td></td>

<td></td><td> m]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

<td></td><td> to</td><td> wood fiber</td><td colspan="3"> (cellulose) board ceiling</td><td> panel.</td>

<td> 15</td><td> n]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

<td></td><td> to</td><td colspan="3"> asbestos-cement board.</td><td></td><td></td>

<td></td><td> O]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

<td></td><td> to</td><td colspan="3"> plywood covered with vinyl</td><td> wallpaper,</td><td></td>

<td></td><td> p]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

<td> 20</td><td colspan="3"> between two pieces of</td><td colspan="3"> 3-ply laminated plywood.</td>

<td></td><td> q]</td><td> Coating</td><td> and/or</td><td> coating</td><td> laminate</td><td> applied</td>

between two piece’s of 1-ply doorskin, and between 4 pieces of doorskin (4-ply).

a] Type II and Type III Coatings on Bare Fabric.

a] 1. Pieces of HOECHST Trevira Spunbond

Polyester geotextile fabric, grade 1110, 3.3 oz. per sq. yd., were cut into 5 1/2 wide x 14 long pieces. Coated with the paint on both sides. One group permitted to air dry, uncovered; another group was pressed between two sheets of polyester (Mylar) film and permitted to set covered. Coating covered only 10 of the panel, permitting /the uncoated 4 length to be used as a control. All coatings cured into a refractory-hard brilliant white surface, crack-free. The covered panels were smooth and glossy; and due to the compression, the dried coating on each side was 2 1/2 to 3 mils thick and weighed a total of 104.7 gms. The coatings in the uncovered fabric were 4 mils thick on each surface. Each coated fabric was exposed several times to the standard 1950°F propane torch test for 15 and 30 minute periods, all with similar results. The control sections of the panel, flamed and then melted within 1 second. The spot under the direct flame in the coated areas became white hot on both surfaces. The area on both sides calcined, but the flame did not break through to the opposite side, no hole was created in the fabric, and the coating did not crack and decrepitate. Occasionally, upon cooldown, the coating in the heated area did crack. The fabric in the heated areas volatilized and left a space betwen the top and bottom coats of paint. In the 1/2 to 3/4 diameter penumbra surrounding the heated areas, the stroma of the fabric was still intact, but the fabric had carbonized. Beyond the penumbra (3 diameter), neither the fabric nor the coating showed any signs of change.

a] 2. An open-mesh woven fiberglass cloth, called SCANDATEX Wall Covering, 3 oz/sq. yd., with a 90 Ibs/inch tensile strength, cut into strips 5 1/2 x 14, was used as the supporting matrix for a similar series of tests - with comparable results. Even after 15 and 30 minutes of exposure, there is zero flame spread from the point of the flame application, with only minor lateral penumbral involvement. While the glass fibers do not volatilize, they do melt, but the strength of the coating maintains structural integrity of the sheet. Since the organic material content of the glass fabric is less than that of the polyester fabric or jute, it might be expected that the production of volatile gasses or smoke would be less, but this was not measured.

a] 3. An open-mesh woven burlap or jute, 8-mesh,

6.5 oz/sq. yd., cut into strips 5 1/2 x 14, was similarly coated on both surfaces with the paint, and tested by the propane torch method. Similar results were obtained. After 15 and 30 minute test exposure, there was zero flame spread, and the coatings held their integrity, although calcined.

In the above series of tests, which demonstrate the non-combustibility of the coatings and the ability of the coatings to prevent the penetration of open flame to the opposite surface, the fabrics acted as both a substrate and as a stress-relief filler preventing cracking and decrepitation, and indicated the, value of this combination in forming a coating laminate for application to fire-susceptible substrates .

TABLE II lists the various combinations of coatings and/or coating laminates applied to a variety of substrates and the results obtained when these were subjected to flame testing to determine their fire-barrier characteristics.

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES

<td> Refer to Configuration List______________________</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> b] White Pine Board b.l</td><td> 6xl2xl/4</td><td> Propane Torch</td><td> Wood ignited in 15 seconds flaming brand in 30 seconds; complete penetration to opposite side in 1 1/2 minutes.</td><td> Uncoated control. 1</td>

<td> b.2 Coating alone</td><td> 6xl2xl/4</td><td> Propane Torch</td><td> First 6 minutes; slight scorch on back of panel; no combustion; at 8 minutes, coating cracked; at 9 minutes, wood beneath coating ignited. Test terminated.</td><td> * Two Coats ¹⁰ of paint; 1 Coating thickness: 8 mils.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Coni'd)

<td> Refer to Configuration Size Test Method Results List</td><td> Observations</td>

<td> b.3 White Pine Board 23 7/8x3 7/8 Two-foot Flame Spread: x 3/78 Tunnel Test 73.0 ASTM-D-3806 Percent Loss:</td><td> Uncoated control. Proximal</td>

<td> (5-min.burn) 74.65% Smoke Density: 578.9%</td><td> half-length of board completely</td>

<td> **</td><td> consumed, distal half completely charred and cracked. (Compare with b.7 and b.8).</td>

** Based upon results for Red Oak as 100, and asbestos board as 0.

TABLE II; FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations List

<td> b.4</td><td> White Pine Board 23x3״xl/4</td><td colspan="2"> ASTM-D-3806 Flame Spread: Uncoated</td>

<td></td><td></td><td> (5-nin. burn) 69.0 control Percent Loss: Plain White 83.60% Board; almost Smoke Density: completely 321.170 consumed; Fuel Contri- charred bution Factor: remains from 90.6 the distal end.</td><td> 1 U1 H 1</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations

List

<td> b.5</td><td> Competitive 23x3xl/4</td><td> ASffl-D-3806</td><td> Flame Spread: B.Moore</td>

<td></td><td> coating on</td><td> (5-min. burn)</td><td> 15.6 RETARDO in-</td>

<td></td><td> White Pine</td><td></td><td> Percent Loss: tumescent</td>

<td></td><td> Board 1</td><td></td><td> 7.6070 paint, control Smoke Density: (2 coats) 12 46.17□ mils thick. . (Note: Intumescent paints are short-term protection; under propane torch tests, breakthroughs occur when the char is eroded by the flame’s velocity.</td>

TABLE II; FIRE BARRIER PROPERTIES OF TYPES II AKD III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to List</td><td> Configuration</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> b.6</td><td> Coating alone</td><td> 23x3xl/4</td><td> ASTM-D-38-6 (5־rain. burn)</td><td> Plane Spread: 11.1 Percent Loss; 28/5.6% Smoke Density: 79.0% Fuel Contri- buted: 0</td><td> Two Coats of Type II, totalling 100.7 gms. (1.46 gms/ sq.inc.).Percent weight loss and ' smoke density principally due to water driven off in the heating Minimal carbonization under the coating at the point of flame. Coating maintained adhesion to the substrate.</td>

I

U1 w

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations List

<td> b.7</td><td> Coating Laminate</td><td> 23 7/8x</td><td> ASTM-D-3806 Flame Spread: 0</td><td> Single coat of</td>

<td></td><td> on White Pine</td><td> 3 7/8x</td><td> (5- min. burn) Percent Loss:</td><td> paint and single</td>

<td></td><td> Board</td><td> 1/4</td><td> 5. on Smoke Density: 0 Fuel Contri- buted: 0</td><td> layer of HOECHST spunbond non-woven polyester fabric. Superficial ¹ layer of coating m decrepitated; ^Λ undercoat was ¹ not affected; only slight evidence of scorch. Coating maintained excellent bond to substrate. (Compare with b.3).</td>

I¹׳ .1

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size ' Test Method Results List</td><td> Observations</td>

<td> b.8 Coating 23 7/8x3 7/8 ASTM-D-3806 Flame Spread; Laminate xl/4 (5-min. burn) 0 on White Pine Percent Loss: Board 3.897.</td><td> Double layer of polyester fabric, and double coat of</td>

Smoke Density; paint, total-

<td> 9.21 Fuel Contributed: 0</td><td> ling 1/8 thick. Coating under flame emerged in ! perfect contion. No ef- יט fects on sub- ! strate. Since Photoelectric cell does not distinguish between the constituents</td>

<td></td><td> causing the decrease in light intensity, this' smoke is due to steam -</td>

vaporized

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Coat'd)

<td> Refer to Configuration List</td><td> δίζβ</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> b.8 (cont'd.)</td><td></td><td></td><td></td><td> water moisture. (Compare with b.3)</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES

’נ;!

<td> Refer tt List</td><td> ) Configuration</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> b.9</td><td> Coating Laminate White Pine Board</td><td> 23x3xl/4</td><td> ASTM-D-3806 (5-min. bum)</td><td> Flame Spread: 4.5 Percent Loss: 24.12% Smoke Density: 79.01 Fuel Contri- buted: 0</td><td> Single coat of paint and single layer of Swedish SCANDATEX wall-covering open-mesh woven fiberglass, Coating main- 1 tained its integrity. No m significant 1 carbonization of the substrate.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND HI COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size List</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> c] Interior/ 8x8xi/2 Exterior Laminated Plywood c.l</td><td> Propane torch</td><td> Initial flaming started in 20 seconds; zone of combustion extended 5 from epicenter; complete penetration of board, leaving a 1 1/2 diameter hole in 3 minutes.</td><td> Uncoated control.</td>

<td> c.2 Interior/ 23 7/8x3 7/8 Exterior x 1/4 Laminated Plywood</td><td> ASTM-D-3806 (5-min. burn)</td><td> Flame Spread: 73.0 Percent Loss: 58.61% Smoke Density: 543.4</td><td> Uncoated control. After burn test, remnants were a few broken pieces of charcoal, carbonized throughout.</td>

8

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND HI COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations

List

c.3 Coating alone 12xl2xl/2 Propane Torch Gently warmed Two layers of on Interior/ area for 30 Type II paint,

Exterior seconds to 10 mils thick.

<td> Laminated Plywood</td><td> bring temperature up gradually to a full 1950°F. First 3 minutes, no | ״ noticeable change; at 5 minutes ⁰¹ steam escape was noticeable; at 7 ¹ minutes first sign of carbonization of underlying wood (but not combustion); at 9 minutes cracks appeared in the coating; area under flame started to combust at 10 minutes, but no flame spread; temperature on back of the panel was 200°F. ,</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Coat'd)

<td> Refer to Configuration Size List</td><td> Test Method Results Observations</td>

<td> c.4 Single Coating 12״xl2xl/2 Laminate on Interior/ Exterior Laminated Plywood</td><td> » Propane Torch Steam started HOECHST spunbond to emerge at polyester geo- 3 1/2 minutes; textile fabric no decrepita- was used in the ting; at 5 laminate, minutes, coating was white hot; at 1 minutes coating calcined but no cracking; area under flame carbonizing, slight smoke emission; at 10 minutes, heavy smoke emission; 1/4 deep carbonization, but no flame, only directly under impinging flame.</td>

TABLE II; FIRE BARRIER PROPERTIES OF TYPES II AND IH COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont^,d)

Refer to Configuration Size Test Method Results Observations List

<td> Single</td><td> 23'7/8x</td>

<td> Coating</td><td> 3 7/8x</td>

<td> Laminate on</td><td> 1/2</td>

<td> Interior/</td><td></td>

<td> Exterior</td><td></td>

<td> Laminated</td><td></td>

<td> Plywood</td><td></td>

d] Wood Wall 23 7/8x

Panelling 3 l/2x d.l 5/32

ASTM-D-3806 Flame Spread: Compare with (5-min.burn) 0 the uncoated

Percent Loss: control, Test 3.00% c.2 above,

Smoke Density: HOECHST spun9.2% bond polyFuel contri- ester fiber buted: 0 fabric used י in the laminate.

ASTM-D-3806 Flame Spread: Uncoated (5־min. burn) 69.0 control.

Percent.Loss: Georgia77.7970 Pacific, 3Smoke Density: ply untreated 584.270 plywood.

Fuel Contri- CEDAR buted: 58.6 HOLLOW, simulated wood grain printed finish.

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to List</td><td> Configuration</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> d.2</td><td> Coating alone on Wood Wall Panelling</td><td> 23 7/8x 3 l/2x 5/32</td><td> ASIN-D-3806 (5-min. burn)</td><td> Flame Spread: 28.9 Percent Loss: 13,227. Smoke Density: 71. Π Fuel Contri- buted: 0</td><td> Two Coats of coating,alone on front surface. Flametested on rear (uncoated) surface. Indicates that applying the coating laminate to the opposite side of a panel aids in slowing down fire spread.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont^,d)

<td> Refer to Configuration Size Test Method List</td><td> Results Observations</td>

<td> d.3 Coating 23 7/8x3 1/2 ASTM-D-3806 Laminate x5/32 (5־min. burn) Applied to Wood Wall Panelling</td><td> Flame Spread: Coating Lami33.4 nate, using Percent Loss: woven fiber15.05% glass fabric Smoke Density: applied to the 19.770 front surface. Fuel Contri- Flame-testing buted: 0 on rear (un- coateJT sur-</td>

face. ΙηΗΓcates that applying the coating laminate to the opposite side of a panel aids in slowing down fire.spread.

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method List</td><td> Results Observations</td>

<td> f] Red Cedar 18x6 1/4 Propane Torch Roof Shingle xl/8 to 1/2 f.l ׳</td><td> Uncoated sec- Uncoated tion ignited control. in 6 sec. and became a fiery brand, re-, ducing itself to charcoal and ash in 4 min. Flame ¹ spread was a unchecked. 1</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration List.</td><td> Size Test Method Results Observations</td>

<td> f.2 Coating Alone on Red Cedar Roof Shingle</td><td> 18x6 l/4x Propane Torch Coated Two coats of 1/8 to 1/2 section tested coating alone, on back side: applied to oneBoard ignited half of bottom under tip of surface of the flame, and wood shingle. Tested flamed for 5 on top (unmin., but did coated surface), not spread be- Results indi- ¹ yond a 1 1/2 cated that * diameter flame the applying of front, until coating to the ¹ it reached opposite surthe rear of face of the the coating, panel prevents then self- flame spread extinguished. ״ potential 45-min. torch value in woodflame exposure,shingled roof playing direct-construction, ly onto the rear of the unsupported bare</td>

<td></td><td> coating, did not</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations

List

f.2 (Cont;d) extend the flame front and did not affect the coating.

TABLE II: FIRE BARRIER PROPERTIES OF TYPES IT AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size List .</td><td> Test Method Results Observations</td>

<td> g] Fiberglass 6xl2xl board</td><td> Propane Torch Flame started Uncoated to melt the control . glass in 10 Johns-Manville sec., eroding fiberglass away the Air-duct Board</td>

<td> gl 1</td><td> fibers in a without alu- circular hole; minum-foil 1 1/2 diameter facing; 6 hole was comple־ pounds per ! tely drilled cubic foot through the density. 5</td>

<td></td><td> panel in 45 to ¹ 60 sec., leaving a rim of solidified glass beads in the wall.</td>

TABLE II; FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont’d)

Refer to Configuration Size Test Method Results Observations

List

g.2 Coating 12xl2xl Propane Torch 1st Test, 5 Single Coat .Alone on mins., follow- of Type II

Fiberglass ed by dissec- paint, at 1.6

Board tion of the gms/sq.in.

top coating. 20-24 mils 1/4 to 3/8 thick. Exceldepth״ lent adhesion resin had va- to the resinporized, leav- bonded fibering gray un- glass; no bond bonded fiber- de־lamination glass. No sig- after heating, nificant melt- Conclude that ing of glass, the coating had 2nd Test, 15 protected the min., follow- non-combustible ed by dis- fiberglass, section of top permitting it coating. Under-to withstand neath the yet- the erosive whole coating, action of the the fiberglass flames up to had melted, 15 times normal 3/4 deep exposure, leaving a crater 2 in

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AMD III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd) (3130

<td> Refer to Configuration List</td><td> Size</td><td> Test Method Results Observations</td>

<td> g.2 (cont'd)</td><td></td><td> diameter, in whose wall there was the melted glass beads, with a 1/4 thick fiberglass bottom. 3rd Test, 20־min. burn. Similar to 2nd test, but whole was completely drilled</td>

<td></td><td></td><td> through to the bottom of the panel</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer ־to List</td><td> Configuration</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> g-3</td><td> Fiberglass Board</td><td> 24x3 7/8 xl/2 ־</td><td> ASTM-D-3806 (5min. burn)</td><td> 1 Flame Spread: 4.5 Percent Loss: 3.371 Smoke Density: 85.51 Fuel Contributed: 0 ׳&gt;</td><td> Uncoated control. JohnsMansville Airduct Board, without aluminum foil; 6 Ib/cubic foot density, sliced to half thickness of 1/2״ to fit test equipment fixture. Flame had drilled completely thorugh the panel, 2 in diameter, and for 4 diameter, had burned out the resin binder, with a 610-״ng scorch length.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method Results List</td><td> Observations</td>

<td> g.4 Coating 24x3 7/8x ASIM-D-3806 (5- Flame Spread: Alone on 1/2 minute burn) 2.2 Fiberglass Percent Loss: Board 13. TL Smoke Density: 21. IX fuel Contribut ion Factor: 0 י</td><td> Two Coats of Type III paint; 2.6 gms/sq.inc. Direct 5־min. flame had cuased only a 1/4 depth resin burnout under the flame; with only 1/32 depth of glass melting; no other significant changes.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method Results List</td><td> Observations</td>

<td> g.5 Coating 24׳'x3 7/8x ASTM-D-3806 Flame Spread: Laminate on 1/2 (5-minute burn) 2.2 Fiberglass Percent Loss: Board 9,427.</td><td> Single Coat laminate, using HOECHST geotextile poly-</td>

Smoke Density: ester spunbond

<td> 21.1% Fuel Contri-</td><td> fabric. 3.2 gms of paint per</td>

bution Factor: so, inch: coat-

<td> 0</td><td> ing in excellent condition; refractory hard; with excellent adhesion to the fiberglass substrate. Although the top layer of the coating cracked upon cooling, there was almost</td>

<td> -</td><td> no change</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations List

<td> g.5 (Cont'd)</td><td> in the fiberglass substrate; no resin burn-</td>

out; no melting of fiberglass — the laminated coating had afforded complete fire protection.

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method List</td><td> Results Observations</td>

<td> M Gypsum Plaster Wallboard</td><td> The surface paper covering of gypsum wallboard burns off in a fire. When hung horizontally, the plaster will crack, dehisce, and fall ! exposing more _N susceptible sub- .&gt; strates. This ! series of tests is to determine the effect of the coatings and coat-־ ing laminates on preventing the decomposition of the plaster wallboard.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont’d)

<td> Refer to Configuration List</td><td> Size Test Method Results Observations</td>

<td> h.l . Gypsum Plaster Wallboard</td><td> 12xl2x5/8 Propane Torch Paper burned Uncoated off in 10 secs. Control, and substrate calcined pro-</td>

gressively deeper with time. No flame spread, but at the end of 15 min. test, area of calcination, 2” in diameter, had reached opposite side of panel and covering paper was scorched. Powdered calcined plaster easily punctured and scraped off, for full depth of panel.

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations List

<td> h.2</td><td> Coating</td><td> 12xl2x5/8 Propane Torch In a corres- Two layers of</td>

<td></td><td> Alone on</td><td> ponding 15- Type II paint.</td>

<td></td><td> Gypsum</td><td> min. exposure</td>

<td></td><td> Plaster</td><td> test, the</td>

<td></td><td> Wallboard</td><td> coating remained intact over the flamed area without cracking. The depth of cal-</td>

<td></td><td></td><td> cination and powdering of the gypsum was 1/4 to 3/8, below which it retained its full strength, and there was no scorching on the rear purface. The board could not be punctured to its full depth.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to List</td><td> Configuration</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> b.3</td><td> Gypsum Plaster Wallboard</td><td> 22x3 l/2x 1/2</td><td> AS1N-D-3806 (5-fflin. burn)</td><td> Flame Spread: 0 Percent Loss: 3.55% Smoke Density: . 1,122.a Fuel Contributed: 0</td><td> Uncoated Control. The paper coating of the proximal 5 length nearest to the flame source was' ¹ burned off and m 1/4 to 3/8 depth of panel ¹ was calcined and deeply fissured, and cracked easily during handling.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES I! AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Coat'd)

<td> Refer to List</td><td> Configuration</td><td> Size</td><td> Test Method</td><td> Results *</td><td> Observations</td>

<td> h.4</td><td> Gypsum Plaster Wallboard</td><td> 22x3x 1/2</td><td> . ASTM-D-3806 (5־min. burn)</td><td> Flame Spread: 4.5 Percent Loss: 6.07% Smoke Density: 161.8% Fuel Contri- buted: 0</td><td> Uncoated Control Results very similar to h.3</td>

<td> h.5</td><td> Coating</td><td> 22x3xl/2 ASIM-D-3806</td><td> Flame Spread:</td><td> Two Coats of</td>

<td></td><td> Alone on</td><td> (5-min. burn)</td><td> 11.1</td><td> Type II paint,</td>

<td></td><td> Gypsum Plaster Wallboard</td><td></td><td> Percent Loss: 6.811 Smoke Density; 10.5% Fuel Contri- buted: 0</td><td> totalling 82.3 gms or 1.2 gms/sq.inc. Wallboard covering paper charred beneath the paint, and there was a ־ 1/16” depth of plaster calcination.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration List</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> h.6 Coating Alone on Gypsum Plaster Wallboard</td><td> 22x3xl/2 «</td><td> ASTM-D-3806 (5־min. burn) .י.</td><td> Flame Spread: 2.2 Percent Loss: 6.27% Smoke Density: 336.87. Fuel Contri- buted: 0</td><td> Two Coats of type II paint, totalling 136.5 gms or 2.0 gms/sq. inc. Wallboard paper under coating charred, but there was zero calcination of the plaster. Coating under flame maintained excellent adhesion to the charred paper. (Compare to h.3 test).</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to List</td><td> Configuration</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> b.7</td><td> Coating Laminate on Gypsum. Plaster Wallboard</td><td> 22x3xl/2</td><td> ASTM-D-3806 (5-min bum)</td><td> Flame Spread: 0 Percent Loss: 5.68% Smoke Density: 30.370 Fue Contribution Factor: 0</td><td> Single laminate, using HOECHST geotextile spunbond polyester fabric. Slight degree of paper charring but no plaster calcination under the flame. (Compare to h.3 test)</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES Π AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont' d)

Refer to Configuration Size Test Method Results Observations

List

<td> Rigid poly-</td><td> 12xl2x2</td><td> Propane Torch Material</td><td> Uncoated</td>

<td> urethane</td><td></td><td> ignited</td><td> Control.</td>

<td> foam board</td><td></td><td> immediately,</td><td> 6-pound .</td>

<td></td><td></td><td> issuing heavy,</td><td> density</td>

<td></td><td></td><td> acrid, sooty</td><td> rigid ure-</td>

<td></td><td></td><td> smoke.Conti-</td><td> thane foam</td>

<td></td><td></td><td> nued to burn for 15 secs, after flame was removed, and smoldered for several additional minutes, leaving a black char,</td><td> board.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont^,d)

Refer to Configuration Size Test Method Results Observations List

i.2 Coating 12xl2xl Propane Torch 1st 3 minutes Used HOECHST Laminate on a slight amount spunbond

Rigid Poly- of steam issued polyester

Urethane from area under fabric in

Foam Board flame; at 5 min. the laminate, smoke, associated with a total with initial thickness of charring of 40 mils, the foam, commenced. Depth of char continued progressively as heating continued, but foam did not ignite. Coating remained intact, for the 15-min. of the test. Area of foam decomposition ־ 3 diameter. No. flaming, and no flame spread.

TABLE Π: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations

List

Isocyanurate Foam Board

23x3 l/2״x 1

j.1

ASTM-D-3806 Flame Spread: Celotex (5-imn. burn) 26.7 THERMAX isoPercent Loss: cyanurate insu12.3970 lation foam

Smoke Density: board with alu157.970 minum foil

Fuel Contri- facing removed, bution Uncoated

Factor: 0 control.

Although rated as self extinguishing, the proximal 8 of the board was either completely consumed, or reduced to less than 1/2 thickness; flame burn showed along the complete length of the board.

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td colspan="2"> Refer to Configuration List</td><td> Size</td><td> Test Method</td><td> Results</td><td colspan="2"> Observations</td>

<td> j.2</td><td> Coating Alone on Isocyanurate Foam Board</td><td> 23x3 l/2xl</td><td> ASTM-D-3806 (5-min. burn)</td><td> Flame Spread: 2.2 Percent Loss: 7.13% Smoke Density: 25.0% Fuel Contribution Factor: 0</td><td> Foam charred to a depth of 1/4 immediately below point of flame impingement, but no spread beyond the 1 1/2 diameter flame area, plus an additional 1/2 penumbral zone. Board essentially unaffected by the flame exposure.</td><td> 1 00 1</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method Results List</td><td> Observations</td>

<td> j.3 Coating 23x3 l/2xl Propane Torch Flame Spread: Laminate on 0 Isocyanurate Percent Loss: Foam Board 10.48%</td><td> Laminate used was Swedish SCANDATEX woven glass</td>

Smoke Density: fiber cloth.

<td> 104.01 Fuel Contribut ion Factor: 0 A</td><td> Flame exposure caused a 3/8 depth char in the foam directly under the foam and for a 1 diameter penumbral zone of influence. The coating remained intact, without cracks.</td>

5 i ׳ TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING ־''. LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration List</td><td> Size Test Method Results Observations</td>

<td> k] Polystyrene Foam Board</td><td> 12xl2xl Propane Torch Immediately 1 1/2-pound/ upon exposure cubic foot dento flame, the sity, insulapolystyrene ig-tion board.</td>

<td> k.l</td><td> nited and burned, and continued burning quietly, emitting dense, sooty, acrid smoke. After several minutes, the flame was self-extinguished.</td>

' *i

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING ׳ LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size List</td><td> Test Method Results Observations</td>

<td> 12 Coating 12xl2xl Laminate on Polystyrene Foam Board t</td><td> Propane Torch Within the Thin fiber- 1st 15 secs. glass veil the polysty- used as the rene melted be- stroma for the neath the point 14-mil-thick of the flame Type II coat- application - ing. without flaming and without ! smoking, _a leaving an empty space ! behind the laminated coating, which maintains its rigidity and strength. Thus there is zero flame spread.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method Results List</td><td> Observations</td>

<td> 1] Corrugated 23x3 l/2x ASTM-D-3806 Flame Spread: Paper Board 1/2 (5-min. burn) 69.0 Percent Loss: 1.1 87.95X Smoke Density: 296.17. ז 1</td><td> 3-flute, four layer, untreated corrugated paper board. 475 lb. test; used for heavy wall containers, partitions, temporary shelters, etc. Control was almost completely consumed. Only 5 of distal end was recognizable as a corrugated board, and this was badly charred and reduced to 2 flutes; all else was ash</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont’d)

<td> Refer to Configuration List</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> 1.1 (Cont’d)</td><td> &lt;</td><td></td><td></td><td> and carbonized paper.</td>

<td> 1.2 Coating Alone on Corrugated Paper Board</td><td> 23x3 l/2xl/2</td><td> ASTM-D-3806 (5-min, burn)</td><td> Flame Spread: 0 Percent Loss: 13.811 Smoke Density: 01 Fuel Contribu- tion Factor:0</td><td> Only the 4 length closest to the flame source was carbonized under the coating, and • this only partially. The</td>

8 board, although blackened, still showed structural integrity. Rest of panel was unaffected.

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method Results List</td><td> Observations</td>

<td> 1.3 Coating Laminate 23x3 l/2x ASTM-D-3806 Flame Spread: on Corrugated 1/2 (5-min. burn) 0 Paper Board Percent Loss: 12.70Z</td><td> Results.similar to 1.2 The flame showed a 4 long zone of influ-</td>

Smoke Density: ence which 15.870 caused carbonizaFuel Contribu- tion of two

<td> tion Factor:0 f</td><td> flutes of the board under the coating; leaving ¹ a free-standing g coating in this area; but ex- ¹ cellent adhesion of the coating to the paper in the area outside of the ' zone of influ.ence.</td>

TABLE II; FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method List</td><td> Results Observations</td>

<td> m] Wood Fiber 23x3xl/2 ASTM-D-3806 (Cellulose) (5-min. burn) Board Ceiling m.l _Tile</td><td> Flame Spread: Uncoated 37,9 Control Percent Loss: Armostrong 27.27% ceiling panel Smoke Density: (acoustical) 19.770 Type 195 Classic; Fuel Contri- Fed. Spec. SS-S</td>

<td></td><td> buted: 0 118-B, Class C ¹ Flame spread £ Index 76-200. ₍ The panel was scorched along the entire length, and under the point of flame application was completely eroded so that the panel broke into</td>

<td> 4</td><td> pieces. In propane torch tests, this panel, once ignited, after flame is removed, continues to</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method List</td><td> Results Observations</td>

<td> m.l (Cont'd)</td><td> smolder for 6 to 8 hours until entirely consumed by the combustion.</td>

<td> m.2 Coating Alone 23x3xl/2 ASTM-D-3806 on Wood Fiber (5-min. burn) (Cellulose) Board Ceiling Tile</td><td> Flame Spread: Two Coats of Type , 13.4 II paint, total- Percent Loss: ling 80 g or 1.16 ₍ n 12.587» gms/sq.in. The ! Smoke Density: thin coating 21.Π apparently Fuel Contri- cracked under buted: 0 the flame, thus accounting for the short length (4״) of charred substrate; however, beyond this area, the panel was unaffected and intact.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont’d)

<td colspan="2"> Refer to Configuration List</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> nl n ו</td><td> Coating Alone on Cement Asbestos Board</td><td> 23x3 l/2x 3/16</td><td> ASTM-D-3806 (5-min.burn)</td><td> Flame Spread: 0. Percent Loss:</td><td> Purpose: To establish a base-line for</td>

5.427□ behavior and

Smoke Density: performance of

<td> ״47:47</td><td> the Type II coat-</td>

<td> Fuel contri-</td><td> ing, when on</td>

<td> button</td><td> a standard</td>

<td> Factor: 0</td><td> incombustible</td>

<td></td><td> surface. Cement</td>

<td></td><td> asbestos board</td>

resists only 6OO°7־OO°F without disintegrating through loss of cement bond. The rigid board fractured under the coating, leaving the two coats of type II paint as a free-standing sheet.

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td colspan="2"> Refer to Configuration List</td><td> Size</td><td> Test Method</td><td> Results</td><td> Observations</td>

<td> n.2</td><td> Coating Laminate on Cement Asbestos Board</td><td> 23x3 l/2x 3/16</td><td> ASTM-D-3806 (5-min. burn)</td><td> Flame Spread: 2.2 Percent Loss: 12,637.</td><td> Purpose: To establish a baseline for behavior and per-</td>

Smoke Density: formance of the

<td> 38.21</td><td> Type II coating</td>

<td> Fuel Contri-</td><td> laminate, with</td>

<td> but ion</td><td> HOECHST spunbond</td>

<td> Ractor:0</td><td> polyester fabric when on a standard incombustible surface.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

י£

Η Μ Η Ν

W Η ω ο

Refer to Configuration Size

List

Test Method Results

Observations

0] Vinyl Wallpaper 12xl2xl/2

Covered Coating

Laminate on

0,1 Laminated

Plywood

Propane Torch Wallpaper, 3־ply indoor/ under the outdoor plywood, direct flame, covered with burned immedi- type II paint ately in a 1 and HOECHST 1/2 circle, spunbond polybut the flame ester fabric did not spread forming a lamiand stopped at nate; the lamithe edge of nate being used ¹the penumbra, as the adhesive ό

1st 3 min., for applying / only traces a 4-mils-thick ¹ of steam vapor decorative vinyl was noted; at wallpaper to 5 min. smoke the substrate, from the flame;The paint makes at 7 min. top an excellent ply of the wallpaper adhepanel had car- sive, so tightbonized ly bonded that through, tempe-it could not rature on back be separated of panel of from the subabout 160°F strate

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method Results List</td><td> Observations</td>

<td> 0.1 (Cont’d) At 10 min. middle plywood layer started to carbonize; at 12 mins, crack in coating permitted flame penetration to the substrate. At 15 mins, rear surface of board showed scorch marks and initiation of carbonization.</td><td> without tearing. Results are similar to test c.4, TABLE II. The coating laminate forms an efficient bartier in preventing flame spread, which in 5 replicate, tests did not extend beyond the 1 1/2 diameter direct flame impingement area, plus an additional 3/4 diameter penumbral zone of influence.</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations

List

<td> pl</td><td> Coating Laminate 12xl2c3/4</td><td> Propane Torch Surface of wood</td><td> Two pieces of</td><td></td>

<td></td><td> Applied</td><td> ignited immedi-</td><td> 1/4 thick,</td><td></td>

<td> p.l</td><td> Between</td><td> ately and</td><td> 3-ply plywood</td><td></td>

<td></td><td> Two Pieces of</td><td> burned; burning</td><td> together</td><td></td>

<td></td><td> Laminated</td><td> continued, car-</td><td> with Type II</td><td></td>

<td></td><td> Plywood</td><td> bonizing the</td><td> paint, a com-</td><td></td>

<td></td><td></td><td> three top lamina</td><td> bination fire</td><td></td>

<td></td><td></td><td> of plywood,</td><td> barrier and</td><td></td>

<td></td><td></td><td> reaching the</td><td> adhesive.</td><td></td>

<td></td><td></td><td> coating layer in</td><td> Paint was</td><td> 1</td>

<td></td><td></td><td> 4 minutes; this</td><td> tinted with</td><td> \O</td>

<td> k</td><td></td><td> layer was not</td><td> American</td><td> s|</td>

<td></td><td></td><td> penetrated by</td><td> HOECHST</td><td> 1</td>

<td></td><td></td><td> the'flame, and</td><td> 17-3010</td><td></td>

<td></td><td></td><td> for the pext 10]</td><td> Colanyl</td><td></td>

<td></td><td>₎ ו ׳. 1 4&gt;1 , f.</td><td> min. there was</td><td> CaVamel FK</td><td></td>

<td></td><td></td><td> ho combustion of</td><td> pigment, to</td><td></td>

<td></td><td></td><td> the panel behind</td><td> yield a pine-</td><td></td>

<td></td><td></td><td> the barrier</td><td> yellow color</td><td></td>

<td></td><td></td><td> coating. There</td><td> ־־ so that</td><td></td>

<td></td><td></td><td> was only a gra-</td><td> the coating</td><td></td>

<td></td><td></td><td> dual carbonizing</td><td> layer was in-</td><td></td>

<td></td><td></td><td> (but no flaming,</td><td> distinguish-</td><td></td>

<td></td><td></td><td> and no flame</td><td> able from</td><td></td>

<td></td><td></td><td> spread)</td><td> the other</td><td></td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cant'd)

<td> Refer to Configuration Size . Test Method Results List</td><td> Observations</td>

<td> , of the second P’l '^Cont θ' plywood sheet. Smoke emission, after the first 4 minutes, was minimal</td><td> wood ply layers. Coating made a superb/ adhesive, forming a perfect bond between the two plywood ,panels. The bond did not delaminate under the influence of the heat from the propane torch, nor from the ASTM-D3806 test, (see also p.3)</td>

<td> r</td><td></td><td></td><td></td><td></td><td> •</td>

<td colspan="5"> Λ TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING LAMINATES ON VARIOUS SUBSTRATES (Cont'd)</td><td></td>

<td> Refer List</td><td> to Configuration Size</td><td> Test Method</td><td> Results</td><td> Observations</td><td></td>

<td> p.2 .</td><td> Coating Laminate 12xl2x3/4 Applied Between (total thickTwo Pieces of ness) Laminated Plywood</td><td colspan="3"> Propane Torch Results Two pieces of similar to 1/4 thick, 3p.l. Combus- ply plywood, tion restric- bonded together ted to the top with Type II layer only, paint, plus and the flame HOECHST spunspread restric-bond polyester ted to a 34־ forming a single diameter laminate. The circle. laminate serves as an excellent adhesive, forming a bond which showed no line of separation even when the joined panels were subjected to heavy hammer blows. It is probable that the laminate also</td><td> 1 ύ \O 1</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations List

<td> p.2 (Cont'd)</td><td> adds to the tensile properties of the plywood. Both this test and p_?l show that the fire barrier layer interposed between two susceptible layers can prevent the flame spread from the first layer to the second layer.</td>

DOT

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations List

p.3

Coating Applied 23x3x7/8

Between Two

Pieces of

Laminate

Plywood

ASTM-D-3806 a) Flame Objective: To (5-min. burn) Applied to study the’influ1/4 panel ence of a subsurface: merged barrier coating on flame Flame Spread: spread, and the M;4 effect of nearPercent Loss: ness to the surface of this

Smoke Density: 132.9% Fuel Contributed: 0

b) Flame Applied to 1/2 panel surface: Flame Spread: 69.0

Percent Loss: 6.9970 Smoke Density: 69.770 coating. Tests d2 and d3 had shown that barrier coatings, on the rear of wood paneling limited flame spread on the uncoated surface.

Two samples, consisting of 1/4, 3-ply plywood bonded to 1/2 of 41O1

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont’d)

<td> Refer to Configuration List</td><td> Size</td><td> Test Method Results</td><td> Observations</td>

<td rowspan="2"> p.3 (Cont'd)</td><td></td><td> Fuel Contri-</td><td> ply plywood, with</td>

<td></td><td> buted: 6/9%</td><td> one coat of Type II paint were tested; one being burned on the 1/4 surface (2), and the other on the 1/2 thick surface (b). In both cases, the single layer of paint made an ex-</td>

<td></td><td></td><td> f t</td><td> cellent bonding adhesive which did not debond under the heat , of the test even when the carbonized ply had reached the surface of the coating.</td>

H 0

נא

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRA.TES (Cont'd)

<td> Refer to Configuration Size Test Method List</td><td> Results Observations</td>

<td> p.3 (Cont'd) f</td><td> Panel (a) showed flame scorch on only the proximal half of the length; Panel (b) showed flame scorch over the entire length of the panel. In neither case did the flame penetrate the coating. The closer the coating to the wood surface, the greater the flame spread׳protection, (see also p.l) »</td>

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method List</td><td> Results Observations</td>

<td> q] Coatings Applied 8x8x5/16 Propane Torch Between Two and Pieces of 1-ply q.l Doorskin and 6x8xl/2 Between 4 Pieces of Doorskin (4־ply)</td><td> a) Flame quick- _a) t_w0 pieces ly punched a 1 _of 1/8 thick diameter hole wood veneer through the doorskin were top layer, bonded to- until it reached getber using the barrier Type !j _coat. coating. No flame ing. spread beyond the 1 diameter addi-b) Four pieces tional penumbra. _of 1/8 thick Flame was effec- wood veneer tively stopped doorskin by the coating, bonded toOn a 25-min.test gether using the rear of the three layers panel gradually of Type II coatcarbonized, ing. but did not flame; the coat- Unprotected, ing while the wood is white-hot highly remained intact flammable, and did not crack;</td>

VOX

TABLE II: FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

Refer to Configuration Size Test Method Results Observations

List

p.3 (Cont'd) paper place These _tests on opposite (and others) side of the indicate coating the possibility scorched, but of constructing did not flame, a multi-layer

...... plywood panel b) Flame with.firepunched a 1- ^barr!^er P^r0־ diameter hole P^ertles through the. top layer, which with time enlarged י to 3 diameter;

flame was stopped by 1st barrier coating layer; 2nd layer of veneer carbonized under 1st barrier coating, but did not flame___________________

TABLE II; FIRE BARRIER PROPERTIES OF TYPES II AND III COATINGS AND COATING

LAMINATES ON VARIOUS SUBSTRATES (Cont'd)

<td> Refer to Configuration Size Test Method List</td><td> Results Observations</td>

<td> J q.l (Cont'd)</td><td> nor combust. A total of 46 minutes was re״ quired for the rear surface to be carbonized.. but at no time during the test did the 2nd, 3rd or 4th layers combust. The three layers of barrier coating, separated by wood char, had formed not only a flame barrier, but a carbonaceous insulation which prevented flame spread. The pyrolyzed area on the rear of the panel was only 1 in diameter.</td>

EXAMPLE III (Type IV Formulation)

Because of the occasional tendency of Types II and III coatings to decrepitate when heated rapidly, and to crack upon cooldown after exposure to the high temperature test flames, expanded vermiculite was added to the formulation to serve as stress-relief foci (vide supra, in the teachings). The formulations which include this component follow closely Types II and III, with suitable adjustments in the quantity of the gauging fluid to allow for the higher fluid absorption of this material, in order to maintain the same paint viscosity. The substitution of the expanded vermiculite for half of the silica flour did not affect the characteristics of the coating - which dried refractory hard and crack-free.

(a) Coated a 12 x 12 x 1/2 plywood board with a single coat of paint (4-5 mils thick).

(b) 12 x 4 x 1/2 plywood board with a single coat laminate, using Swedish SCANDATEX woven fiberglass wall covering.

(c) 23 7/8 x 3 1/2 x 5/32 wood wall panelling; double coat of paint alone.

(d) 23 7/8 x 3 1/2 x 5/32 wood wall panelling; single coat of paint laminate, using

SCANDATEX woven fiberglass.

(a) and (b) were submitted to propane torch 'tests; (c) and (d) were tested by ASTM-D-3806.

(a) In a series of 6-minute tests, there was noticeable steam emitted at 3 minutes, wisps of

י׳'.!/ smoke at 5 minutes, and, in one case, cracking of the coating. Dissection of the burn area showed no sign of discoloration on the rear of the panel, and carbonized wood to a depth of only 1/8. There was zero flame spread beyond the flame impact area. There was no sign of decrepitation in any test.

(b) In a 15-minute test,’ steaming started at 5 1/2 minutes; at 7 minutes, smoke started; at 12 1/2 minutes, rear of panel showed discoloration. At the end of the test, wood was carbonized to a depth of 3/8 and showed cracks -- only under the flame impingement area. There was no flame spread. The coating remained intact and crack-free; and there was no decrepitation, even though the full flame had been applied immediately to the surface without a preliminary warm-up.

<td> (C)</td><td> Flame Spread:</td><td> 6.68</td>

<td></td><td> Percent Loss:</td><td> ־3.67</td>

<td></td><td> Smoke Density:</td><td> ־43.427</td>

<td></td><td> Fuel Contributed:</td><td> 0</td>

<td> (d)</td><td> Flame Spread :</td><td> 2.23</td>

<td></td><td> Percent Loss:</td><td> ־4.87</td>

<td></td><td> Smoke Density:</td><td> ־17.117</td>

<td></td><td> Fuel Contributed:</td><td> 0.</td>

Both (c) and (d) coatings came through the ASTM-D-3806 test in excellent condition. No decrepitation; no cracks; hardly a sign of flame exposure, except for a slight calcination. The highly flammable wood panelling substrate showed minimum carbonization, and the coatings maintained excellent adhesion to the substrate, under the area of direct flame impingement.

EXAMPLE IV (Type V Formulation)

Type V formula was developed to determine the effects of adding a well-known fire-retardant material, alumina trihydrate, to the mag-oxychloride formula. The resulting coatings were particularly smooth and brilliant white, as well as hard (mohs 4.5), and crack-free; exhibited little shrinkage, and had excellent adhesion to a variety of substrates.

(a) Single coating applied to a 12 x 12 piece of HOECHST non-woven spunbond polyester fabric, grade 1110.

(b) Single coating laminate, using same fabric, applied to a 12 x 12 x 1/2 wood waferboard. Utilized 1.17 gms. of paint per square inch of surface.

(c) Single coat, 4 mils thickness, dry, applied to a piece of 12 x 4 x 1/4 white pine board.

(a) and (b) were tested by the propane torch, method; (c) was tested by the chimney method.

(a) At 1 minute, whiffs of steam were noted; at 2 minutes, slight smoke arose from the carbonization of the fabric stroma in the area under the impinging flame; at 3 minutes, coating started to calcine and the area under the flame became white-hot, both on the front and back surface, and developed cracks which radiated out from the center into the penumbral area. In a 15-minute test, although there was no flame spread, the cracks widened and small chunks of coating fell off, leaving a hole in the central flamed area. The central 1 1/2 diameter area under the direct flame was completely calcined, and there was 1/2 diameter penumbral area surrounding the central area, of lesser degree of calcination. This Type V formulation does not possess the physical resistance of Types I, II or III, in its inability to hold the unsupported coating intact without cracking during the period of calcination.

(b) In 4 replicated tests, at 1 minute steam emission was noted; at 2 minutes paint developed cracks; at 3 minutes paint cracks propagated in all directions and opened wider; at 3 1/2 minutes, although there was no decrepitation, wood underneath the laminate started to combust (burn) - not merely carbonize without flaming, as would have taken place if the coating had remained intact. By 7 minutes, wood under the coating was in self-sustained

ו combustion.

(c) In the chimney test, two pieces of wood (the coated piece and an uncoated control) are set up in parallel, facing each other, and separated by a 1 spacer at the top. The bottoms of the boards are protected by sheet metal, and the unit is set on top of an open can of STERNO (solidified wood alcohol), which is ignited, so that the 1400°F flame rises in the chimney formed by the two opposing test specimens. At 1 minute, the uncoated control was a blazing inferno, the flames licking up the complete 12 length of the board, with severe charring at the base. At 5 minutes, the basal 6 length of the control board was reduced to char, at which time the tests were terminated. In the coated panel, the coating developed longitudinal and horizontal cracks, and some pieces had exfoliated. The pine board showed a 1/16 thick layer of char under the coating, but there was no indication of combustion or flaming.

J ’ EXAMPLE V (Type VI and VII Formulations)

Types VI and VII formulations were developed to take advantage of the reputedly higher thermal decomposition temperature and higher water-of-hydration content of magnesium oxysulphate cements over magnesium oxychloride cements. C.W. REDEKER (1967) found that the oxysulphate cements withstand a 400°-500°F higher temperature. Normally, the oxysulphate cements are markedly softer and more abrasive-prone.

Types VI and VII overcame the surface softness problem, and brought the oxysulphate coatings up to an acceptable 5.0 and 5.5 Mohs hardness .. the former by the use of a colloidal silica (as is used in Type 1. formulation) and the latter by using high alumina calcium aluminate cement, as in Types II, III and IV formulations. Both coatings are brilliant white.

The major drawback of both Type VI and VII formulations is the appearance of surface and deep crazing (mud-cracking) in the dried coatings when applied as coatings alone on any of the numerous types of substrates listed in Example II. This crazing not only detracts from the cosmetic acceptability of the coating, but also, under flame-testing, the cracks open up and expose the substrate to the direct flame, which results in substrate ignition.

The problem of crazing in the Type VI and VII oxysulpbate cement coatings has been eliminated by applying the paint onto a non-woven . spunbond polyester geotextile fabric. Here the interlaced fibrils of the fabric form the necessary stress-relief foci, so that the coatings set crack-free.

When applied to fabric as a free-standing coating, without a substrate, and subjected to the propane torch test, the coating behaved the same as discussed under Type II and III coatings, tests

a.l, - namely that the coatings withstand direct flame exposure for 15 and 30 minutes, without a flame spread beyond the direct impingement area; and without burning holes through the fabric.

Type VI and VII coating laminates, applied to wood substrates. The propane torch test is not sufficiently sensitive to distinguish differences in results between oxychloride and oxysulphate coatings. Both VI and VII type coatings, when applied to the surface of white pine and laminated plywood, were capable of protecting the substrates from ignition and flame-spread for periods of 15 minutes, and with thicker coatings, for 20 to 25 minutes.

While Type VI and VII coatings can be categorized as satisfactory, Types I, II and III are the preferred formulations.

From the teachings and test results presented in this disclosure, it is obvious that the paint, subject of this invention, has numerous applications of great utilitarian value.

1. Non-combustible coatings. Since the coating does not burn, it may be used in those confined areas or enclosed spaces which cannot risk the dangers inherent in a paint that burns.

2. High-temperature coatings. Since the coatings withstand 1900°F for extended periods of time, they may be applied in those situations which require high temperature protective paints (for example, coatings for furnaces, wood stoves, mufflers, or as high-temperature radiation reflectors).

3. Abrasion resistant coatings. Since the coatings are hard and refractory, with great compressive strength, they may be applied over soft surfaces or plastic foams to lend rigidity, puncture-proofness, and abrasion resistance to susceptible substrates.

4. Adhesive bonding agent for wood, wallpaper, etc. The coatings exhibit exceptional bonding strengths for wood and paper, hence may be used as an adhesive for the bonding of wood lamina in the construction of plywood; in the assembly of corrugated paper board or honeycomb panels; or the adhesion of wallpaper or decorative foils to walls or substrates, and at the same time, impart fire-barrier or flame-stop properties.

5. Structural laminates. When multilayers (4 to 12) are built up from paint impregnated non-woven spunbond polyester geotextile fabric, and the layers pressed together while curing, non-magnetic, fire-retardant, structural laminates of tremendous compressive and tensile strengths are formed. Resistant to biodegrading agents, these structural laminates can be made in the form of sheets, tubes, posts, curved panels, etc.

6. Fire-barrier coatings. Potential applications are for roofing; walls; fire doors; lining of school and hospital fire-egress exits; aircraft, boat or ship interiors; coating of horse stables or animal barns; coatings for foam plastics; coatings for plastic pipe, wire and cable conduits; acoustic or decorative ceiling tiles; corrugated paper board containers; fireplace hearths and backing; waste containers; etc.; as well as improving the surface properties of gypsum plaster wallboard.

Disclosed herein are new and useful coatings and coating laminates, in several alternative formulations, which have proven capable of forming fire-barriers which prevent flame penetration, ignition, and flame spread to a wide variety of fire-susceptible substrates.

These formulations have overcome the numerous shortcomings of the prior art; viz. salt efflorescence and syneresis of excess fluids; water solubility; high shrinkage; crazing or mud-cracking; spalling (decrepitation) under fire; proneness to abrasion due to softness; need to maintain delicate stoichiometric balance of the principal components; deleterious effects of the excess fluid required for smooth paint application; etc.

There has been developed a durable, rapid-setting, non-combustible thin coating capable of protecting susceptible flammable substrates from ignition and flame-spread, which, at the same time, has the necessary characteristics of a good paint, and which bonds firmly to a wide variety of substrates .