Method of adhering mineral deposit in wood fragment surfaces

Abstract

The disclosed process for making cast vegetable/mineral structural products having flame retardant properties utilize a major volume portion of ligneus plant fragments such as soft and hardwoods, sugarcane, cereal and fiber plant stalks, and a minor volume proportion of a mineral binder deposit comprised of magnesium or calcium oxyphosphates and inert filler particles. Fragments having thickness ranging from 0.3 mm to 8 mm including chips, shavings, strips, strands, fibre bundles, slivers, fibres and peeled and sawn veneer sheets, have applied to their surfaces an aqueous solution of ammonium polyphosphate or soluble acid phosphate salt supplying from 0.15 to 0.40 parts of P₂ O #3# 5 as phosphate ion per part of fragments by weight, and particulate cement solids comprised of MgO or CaO or Mg(OH)₂ or Ca(OH)₂ or MgCO₃ or CaCO₃ ranging from 0.25 to 1.0 part per part of fragment, and from 0.01 to 0.80 parts of inert filler particles and the mixture is molded and held under predetermined compaction pressure until the product has rigidified, in about 10 minutes' times. The molded mass is held under compaction with unit pressures in the range from about 0.3 to about 14 kg/cm². The process is practically immune to cement poisoning sugars and polyphenolics which were found to be detrimental to other cement mixes.

Description

This invention concerns an improved process for bonding or combining lignocellulosic fragments with mineral inorganic binder materials and forming the admixture into a structural product.

The invention particularly relates to such processes wherein the binder materials comprise selected alkaline earth metal compounds and solutions of salts of phosphoric acid reactive therewith capable of developing a bond mass which is distributed initially on fragment surfaces as a gel coating that sets rapidly to a rigid stone-like solid strongly bonding the fragments to each other.

ahdtengentialfissuedmagnesian ofhave having cross-sections about 1.5 to 5 mm wide and thicknesses from just under 1 mm to about 1.8 mm, and substantially free of deleterious waxes, gums, siliceous deposits and pitch adhesions. The strip density as determined by dressing fragments to known volumes and weighing, was about 0.22. The pith side was visibly fissured, and some strips were loosely attached along their long edges by fibrous strands with another strip.

To evaluate the effects of varying the amounts of ammonium polyphosphate solution, the total cement solids, compaction pressure and strip assembly patterns on board density and bond strength of binder deposit, eight test board products were formed each using 200 grams of the oven-dried prepared strips. Because the total weight of solution supplies both the P₂ O₅ required for a given weight of reactive fraction of MgO and MgCO₃ solids, as well as the residual ammonium phosphate salt impregnating the ligneus material, and as this component represents the single largest cost input, it is desirable to be able to reduce the weight of solution while maintaining adequate product strength and resistance to flame.

Moreover, a reduction of total mineral solids decreases the cost of expensive magnesium oxide of the high-density grade which provides the desired reaction rate with ammonium polyphosphate solution. By limiting the compaction pressure and by arranging the fragments so as to obtain the most favorable strength properties for the product in relation to its intended use, optimum low-density and high-strength boards, panels and sheets can be formed.

A first series of three boards was prepared in which the weight ratios of the strips, the cement solids, and the aqueous phosphate solution were, respectively, in parts by weight:

1.00:3.00:1.80.

The cement solids comprised 0.75 parts of magnesium oxide of high density and activity, amounting to 150 grams, of density 3.36, and 2.25 parts of ground dolomite analysing 45.1% MgCO₃, the remainder being chiefly silica and CaCO₃, of grain density 2.62, in amount 450 grams. The solids were of grain sizes passing 100 mesh screen and retained on +250 mesh, with 70% passing 150 mesh.

The strips were sprayed with 360 grams of a commercial fertilizer grade of ammonium polyphosphate solution of specific gravity 1.4 having 50% by weight of dissolved salt, and solution analysis 10:34:00 in respect of nitrogen (as NH₄) and P₂ O₅ (as phosphate ion). This solution was filtered to remove suspended dark matter believed to be carbonaceous solids.

The wetted strips were mixed to distribute the liquid evenly over the surfaces. It was noted that different parts of the wood seemed to have greater permeability to the liquid than other areas, the pith side accepting liquid freely. The solution was absorbed by the greater part of the fragment surfaces within about three minutes from its application, to the extent that the strips were slightly moist to touch.

The estimated enclosing volume of the fragments, not taking into account the fissures, was calculated to be about 10.700 cm² per 100 grams of rind strips. The applied solution represented a layer amounting to 33.6 mg/cm² of surface, providing 9.88 milligrams of P₂ O₅ as reactive component.

The cement solids were then sifted over the cane strips while the fragments were tumble-mixed, and the tumbling was contained for about 20 seconds to distribute the free powder until all the cement grains were adhered. One-third of the coated strips were then immediately shaken into a flat mold with side flanges, of bottom plan area 15 cm by 30 cm. The strip lengths of this layer were aligned generally in parallel and the layer levelled, after which a second one-third portion was transferred into the mold and raked to orient the strips parallel with each other and at a small angle with the first layer; the remainder of the coated strip mix was added with these strips being laid nearly parallel with the first layer strips. A compaction pressure of 0.5 kg/cm² of the top area was applied to the composite by a stainless steel pressure plate having a polished contact face, holding the fragments compacted to form a board of thickness about 3.17 cm. The plate was held at this distance from the bottom of the mold for 5 minutes. At this time the plate was removed and the product was taken out while hot and steaming, giving off an odor of ammonia gas, at a temperature of 44°C It was dried in air for 5 days.

In a similar way, boards #2 and #3 were formed, with compaction pressures respectively 0.7 kg/cm² and 7.6 kg/cm². The corresponding board thicknesses, determined by fixing the plate at the distance from the mold bottom according to the initial position for the attained compaction pressure, were respectively 2.12 cm and 1.19 cm.

Two further boards, #4 and #5 were prepared with the same components, but with the cement solids and solution decreased from the amounts used in the boards 1-3, the ratios by weight being:

#4 1.00:2.50:1.50

#5 1.00:2.50:1.25.

Board #4 was compacted at 12 kg/cm² pressure, producing a thickness of 1.90 cm, while board #5 was held compacted at 4.5 kg/cm² producing a board thickness of 2.74 cm. The "springiness" of the assembled strip materials was clearly apparent from the pressure increase needed to compact the fragments by 0.8 cm thickness.

To explore the feasibility of producing strong board products with markedly decreased amounts of mineral cement solids while using a decreased amount of ammonium polyphosphate solution contributing good fragment-stiffening and fire-retardant properties, boards #6 and #7 were prepared by the procedures described for the first five boards. The weight ratios of the components were as follows:

#6 1.00:2.00:1.20

#7 1.00:1.00:1.30

The boards were held under respective compaction pressures of 11.6 kg/cm². The difference in the amounts of cement solids being the primary cause of respective densities 1.07 and 0.79.

To test the efficacy of bonding by markedly lower amounts of cement solids and aqueous solution, board #8 was prepared by the procedures described earlier, using the weight ratios:

#8 1.00:0.93:1.10

All strips were laid flat and parallel in the mold. The strength of the cured board at seven days of drying air was excellent, and the board costs were appreciably lower than for the other boards. This product would be suitable for wall construction as sheathing, having good nailing and fastener-holding properties in common with all boards of the group.

While the texture of the board surface was rough and it was pervious to air movement, it would have use as a decorative finish panel.

All of the boards of the series, #1 through #8, would be equally castable to form a core board having a continuous face sheet on one or both major surfaces, as conventional in the manufacture of wall panels and boards. Such face sheet may comprise any porous strong layer such as kraft paper whether bleached or unbleached, a porous plastic film such as a perforated vinyl or polyester film or a sheet of woven textile made with natural or artificial strands, yarns or monofilaments, or it may be a wood veneer. The application of such facing sheet or sheets can readily be done either during the casting of the board or after mounting of the structural product on a wall or ceiling. An integral face sheet may be laid on the bottom of the mold box prior to assembling the fragments, and if desired a second sheet of the same or a different material may be placed on the upper surface of the fragments prior to applying compaction pressure. While the migration of binder and solution to the contiguous surface of the sheet from outer fragment portions can adequately bind the face sheet to the core, if desired the inside surface of the one face sheet laid on the mold bottom can be loaded by a sprayed-on wetting film of ammonium polyphosphate solution and a dusting of cement solids just before the treated surface receives the coated fragment mass. Similarly, the under-side of the top sheet if one is to be bound to the core may be treated to ensure strong adhesion.

Post-cure application of a face sheet may be done at any time by similar procedure, provided that an adequate pressure is exerted to hold the face veneer, board, paper, card, or woven or mat sheet closely contiguous to the core so that reactant wetting film and powdered cement solids may form junction bond masses linking the face sheet to the outer fragments.

TABLE VII
__________________________________________________________________________
TABLE OF PROPERTIES OF COMPOSITE
BOARD PRODUCTS CAST USING SUGAR
CANE RIND STRIPS FREE OF PITH
AND OUTER SKIN PORTIONS
Compaction
Weight of  Proportions of
Pressure
Broad
ligneus
Cement
Ammonium
Initial
Bulk
MOR MOE Thickness
No.  strips
Solids
Polyphos.
Kg/cm2
Dens.
Kg/cm2
cm
__________________________________________________________________________
1    1.0   3.00 1.80  0.5    0.34
10.4
11,300
3.17
2    1.0   3.00 1.80  0.7    0.53
37.8
35,100
2.12
3    1.0   3.00 1.80  7.6    0.93
169.5
65,100
1.19
4    1.0   2.50 1.50  12.0   1.05
150.7
56,300
1.90
5    1.0   2.50 1.25  4.5    0.75
30.1
17,700
2.74
6    1.0   2.00 1.20  11.6   1.07
177.1
62,350
1.67
7    1.0   1.00 1.30  11.6   0.79
142.0
41,100
0.75
8    1.0   0.93 1.10  4.5    0.34
32.7
31,300
1.72
__________________________________________________________________________

The exceptionally high MOR values measured for boards Nos. 3, 4, 6 and 7 pertain to thicknesses substantially lower than were obtained at the lower compaction pressures used for boards Nos. 1, 2, 5 and 8. Because cane rind is a highly compressible fibrous material, the strips are intimately brought together along their entire lengths with adjacent strips, and the relatively small void space resulting is largely filled by fluid expressed from the highly porous woody material as the wood volume decreases. This decrease may be 30% of the free bulk volume of the strips. The binder mass which forms adjacent to the major surfaces of the mold and which therefore is at the extreme fiber position of the board contributes high beam strength and stiffness. The interface between fragments provides a maximum area of junction in shear capable of resisting bending moment due to applied load. The MOR values reflect the advantages of using the longest possible fragments such as strips, stalks, slivers, flakes and veneers where a very strong product is desired.

It is also evident that where the quantity of residual salt solution held in the woody matter and the continuous cladding by mineral solid bridging substantially all inter-fragment spaces is favorable, as in the denser cane rind assemblies discussed, strength enhancement of the fragments is at a maximum, as in board No. 6. By way of comparison, conventional wood-wool/Portland cement composites widely used at the present time of even higher density have MOR values below 7 kg/cm², and lack fire-retardant qualities.

EXAMPLE VIII

Fire-retardant low-density structural board products were made from sugar cane residue, i.e. bagasse resulting from mechanical expression of cane juices by passing the stalks which had been stripped of leaves through crushing rollers, and rinsing. The material comprised the rind fiber structure, the pith, and all other plant components namely waxes and siliceous deposits normally present in surface zones of sugar cane.

The bagasse was hammer-milled to produce fragments of random size ranging from dust-like particles to slivers of lengths up to about 8 cm with thicknesses in the range of 1-4 mm. The dust and finest fragments were screened out by passing the fragments over 20 mesh screen and by air-current separation. The dry fragments had average specific gravity about 0.21 and had visibly open and non-smooth surfaces except at rind exterior faces. The calculated total internal space, based on density and solids other than lignocellulosic material, was about 83% of the fragment volume, and tests with aqueous ammonium phosphate solution of density 1.4 indicated that 65% of the fragment volume could be readily impregnated by this liquid.

The estimated enclosing surface area of the selected fragments excluding minute fissures formed by the fragmentation process was 19,000 cm² per 100 grams of dry fragments.

A test quantity of 200 grams of air-dry fragments was measured out. Because of the fineness of the fragments and their very low bulk density, very low density formed products would be feasible to produce even under considerable compaction of a composite of which a major volume proportion--e.g. above 85%--is constituted by ligneus matter. A formulation of reactant cement solids was accordingly selected to provide a vesicular binder mass having 30% or more of its volume as gas-filled minute vesicles, gaining in binder mass volume and enhancing junction edge volumes of binder through lateral flow-out of the more fluent gel material during compaction. The void volume objective of not over 10% cavities was considered to gain the highest structural strength of the composite, for randomly-laid fragments.

A cement solids mixture was prepared comprising 150 grams of a dead-burned magnesia and 450 grams of dolomite, each component being of grain size to pass 100 mesh and retained on 200 mesh, with 70% by weight passing 150 mesh. The quality of the magnesium oxide grain as judged from measured density of 3.36 was inferred to furnish about 11.5%, 17.3 gms reactable to form magnesium oxyphosphate; accordingly 132.7 gms would serve only as a dense inert filler solid. The dolomite reactivity as judged from the measured density of 2.62 and CO₂ combined was inferred to contribute 10.5 gms of MgO to the reaction, the remainder, 428 grams, being inert filler solids.

360 grams of commercial fertilizer grade ammonium polyphosphate aqueous solution, of specific gravity 1,4 1.4 and analysis 10% nitrogen as NH₃ and 34% P₂ O₅ as phosphate, was poured over the bagasse and the fragments were tumble-mixed for two minutes. The surfaces then appeared visibly moist, those fragment portions less than one mm thick appearing to be surface saturated. This amount supplies about 9.5 milligrams of solution per square centimeter of enveloping surface, equivalent to a film of thickness 68 microns, supplying 3.23 milligrams P₂ O₅ per square centimeter.

The powdered cement solids were sifted onto the wetted bagasse while the fragments were tumble-mixed, the mixing being continued for 22 seconds from the time of first contact of the solids with the films. The mixture was immediately dumped into a flanged board mold measuring 15×30 cm. levelled, and compacted by a pressure plate closely fitting in the form at unit pressure of 7.5 kg/cm². A slight relaxation of pressure at constant thickness of 2.53 cm was observed in the first 10 minutes.

The pressure was released at 12 minutes time after casting, and the board was removed and air dried for five days. At the time of release the product temperature was 54°C, and issue of ammonia was noted. The cured board thickness was 2.52 cm.

The resulting board product had a cured volume of 1017 CC and weighed 998 grams, indicating a bulk density of 0.980. The product along its edges and on its larger surfaces exhibited open structure, and could be blown through. By noting the initial displacement on immersion in a column of water, the cavity volume was measured as 67 CC, representing 6.6% of bulk volume. The MOR and MOE values obtained in static bending were respectively 79.4 and 17,900 kg/cm². These strength values are exceptionally high, for a composite having randomly-associated fragments. It is inferred that the high values result from the high volume ratio of binder mass (about 255 CC) relative to fragment bulk volume (1000 CC) and average binder deposit of 75 microns, there being significant deposit thickening adjacent to fragment junctions.

A computation of the residual ammonium phosphate solution remaining as dried salt crystallized within pores and cell spaces of the bagasse indicates that an amount in excess of that meeting Class A requirements for interior finish wood materials by building codes was held in the fragments. Two tests were performed to rate the product.

In a first test, the board product was exposed to direct flame of 800°C temperature over an area of 4.5 by 4.5 cm, for 15 minutes. At the end of that time only minor surface scorching was evident; there was no flaming, minor gas evolution, and no smoke, but within the first two minutes water vapour was observed to be given off. The flame side was red-hot, but the applied flame was not enlarged by any flame contribution from the heated surface. At the end of the test, the face opposite to the test face was only slightly warm, with a localized surface area at 40°C

In a second test an intense heating flame with flame temperature of 1200°C-1400°C was directed against a face area measuring about 15 cm diameter, until the board was perforated. The impinged surface area became glowing hot in about 30 seconds and a gradual ablation of oxidizing ligneus material was noted in a crater area developing. At 31/2 minutes, the opposite surface was glowing. The time to burn through the 2.52 cm-thick board was between 31/2 and 5 minutes from the commencement of the test. This test which is intended to be destructive showed that the fire-retarding properties impeded significantly combustion in the presence of very hot oxidizing flame as compared with conventional composites made of ligneus plant fragments and a mineral binder.

EXAMPLE IX

In order to evaluate the effects on shear strength of the mineral bond mass deposited as a joining material between adjacent surfaces of wood fragments contributed by different reactive cement solids, and the effects of very small particle size and mode of application, a series of wood veneer laminates were made up wherein a minor portion of the cement solids was chosen from magnesium and calcium oxides and hydroxides and this portion was used as an initial application to the wood surface which had previously absorbed ammonium phosphate solution, followed by application of the major portion of cement solids comprising coarser grains of dead-burned magnesia and dolomite.

Twenty-one wood veneers measuring 15 by 30 cm were sawn from clear billets of the wood species:

black cottonwood (Populus trichocarpa)

Douglas-Fir (Coast type) (Pseudotsuga menziesii)

Sitka spruce (Picea sitchensis).

Each wood strip had smooth flat surfaces and thickness 2,8 2.8 mm.

Mineral solids for use as the minor portion of reactive alkaline earth metal compounds were prepared as follows:

A quantity of calcium hydroxide was prepared by hydrating fresh-burned lime and dispersing the hydrate as a slurry with vigorous beating in warm water (55°C). A suspension of the finer particles was reserved, containing about 80 grams solids per liter.

Fresh-burned lime (CaO) was prepared by calcining lump limestone having 97.8% by weight of CaCO₃, the remainder being chiefly silica and clay. The material was ball-milled to a fine dust, which was screened to remove all particles larger than about 15 microns. The dust was held in dried state at very low relative humidity.

A quantity of magnesium hydroxide was also prepared as a precipitate using epsom salts (MgSO₄) and sodium hydroxide, the solids being filtered, washed, and slurried in cold water to form a suspension containing about 100 grams solids per liter.

A quantity of dead-burned magnesia grain of high density (3.36) was ball-milled to very fine particles, and screened to remove particles larger than about 15 microns. The fine dust was slurred slurried in ether to form a suspension carrying about 45 grams per liter immediately prior to its use in the tests to be described next.

In all tests the 3-ply laminate product was air dried for 6 to 8 days before application of test loads. Where necessary the board was slit lengthwise to provide one specimen for evaluation of MOR and MOE in static bending, and a second specimen for measuring bond shear strength. The second specimen was trimmed to 5 cm width and a groove was recessed into each outer face to pierce two wood layers, the grooves being spaced in the length direction of the board to provide a shear zone of area 5 cm by 5 cm; the end portions of the specimen provided gripping areas for applying tension load. This test would reveal failure of the weaker "glue" line of the pair of gluelines placed in shear stress by the load. The test data is presented in Table VIII.

One board was prepared from each wood species for tests (1), (2) and (3) by first tempering the wood to a measured moisture content. A wetting film of ammonium polyphosphate solution of commercial fertilizer grade of analysis 10:34:00 of specific gravity 1.4 having 50% solids content was applied to both faces of the interior ply and to the inner faces of the two outer plies, estimated to provide about 70 milligrams of solution per cm² of surface. The plies were clamped together and dried for 8 days, when the moisture content was again measured. The outer faces were grooved to pierce two wood layers to define a 5-cm long shear area, and the edges of the laminate were trimmed to leave a width of 5 cm.

The tests indicate that the dried salt residue at the interface between plies had limited shear strength, the value for cottonwood being almost 0.6 kg/cm², hence negligible contribution to measured shear strengths in other boards of the series could be attributed to salt residue.

Tests (4), (5) and (6) were carried out by applying aqueous suspension of Mg(OH)₂ to the ply faces in amount to supply about 7 milligrams solids per cm² of surface. The plies were air dried for three days to remove most of the water, when the faces showed a thin white coating of mineral. Ammonium polyphosphate was applied as a wetting film at the rate of 70 milligrams per cm² and the plies were assembled under about 7 kg/cm² unit pressure and air-dried while clamped. The boards were grooved and trimmed as described earlier, and tested to failure in at least one glue line. The results show that the deposit is weak in shear, due to insufficient volume of binder mass and the inability of the reaction products-chiefly magnesium oxyphosphates-to coalesce as a continuous deposit.

Tests (7), (8) and (9) were carried out to establish both the shear strength of a mineral bond mass formed by reaction of -100 mesh cement solids comprised of dead-burned magnesia and dolomite with absorbed ammonium polyphosphate solution as well as bending strength data, where no minor portion of very fine cementing compound is applied. The laminate board products were made by applying 70 milligrams of solution per cm² of surface, then dusting onto the wet films cement solids comprising 20 mg. of dead-burned magnesia of density 3.36 and 60 mg. of raw dolomite of density 2.62, both solids being of grain sizes between 100 mesh and +250 mesh. The plies were clamped at 7 kg/cm² compaction pressure and held for 30 minutes, then released and the boards were air dried for 7 days.

A 3-ply board specimen was made with veneers sawn from each of the three wood species to carry out tests (10), (11) and (12) to investigate the combined effects of an initial minor surface deposit of magnesium hydroxide of sub-micron particle sizes, and a subsequent loading of a major portion of cement solids as in boards (7) to (9). Each contact surface of three plies was initially wetted with an application of ammonium polyphosphate solution to supply about 40 milligrams of solution per cm² of surface area, after which the plies were allowed to absorb the liquid but not to dry out on their surfaces. An aqueous slurry of magnesium hydroxide was distributed on the solution-treated faces to apply about 7 mg of Mg(OH)₂ per cm², and the surface was allowed to dry in air until barely moist. A further application of 30 milligrams per cm² of the phosphate solution was made, followed by application of 80 milligrams of the mixed cement solids used in boards (7) to (9). The boards were pressed and cured as described. The marked increase in shear strength over the control specimens of tests (7) to (9), and the degree to which failure of the wood surfaces had occurred rather than failure through the binder mass, is apparent. The highest wood failure, in cottonwood, and the lowest, in Douglas-fir, correlates with reported shear strengths for these woods parallel to the grain, which are respectively 71.8 kg/cm² and 80.3 kg/cm². Because of the strength enhancement of the surface zone of wood contiguous to the mineral binder deposit, as has been reported in the other examples, it is speculated that failure in a plane through the binder lying between the ply surfaces occurs because the shear strength of the mineral binder is slightly less than that of the strengthened wood zone. It should be noted that the comparable strength of laminates bonded by phenol-formaldehyde resin, using the same woods, did not exceed 7 kg/cm².

The large increase in shear strength for combined applications of Mg(OH)₂ and magnesia/dolomite cement solids of particle sizes as mentioned may be ascribed to the improvement in penetration by magnesium oxyphosphate compounds into the pores, lumens, and microscopic wood fissures opening to the fragment surface, so that a strongly-anchored crystalline oxyphosphate mass occupies microspaces of the surface zone, which zone extends at least one cell layer in depth and possibly extends to a depth of 80 to 100 microns from the outer surface.

When structural products are made from ligneus fragments bonded by cement solids reacted with ammonium polyphosphate solution as recited in these tests, and subjected to static bending, the beam strength will be improved in direct proportion to the capability of the junction bond masses to withstand horizontal shear stress. To evaluate all strength properties of products made with combined forms of cement solids, tests (13), (14) and (15) were made using very fine particles of MgO of high density.

The contact faces of three plies of each wood species were wetted by polyphosphate solution as in control tests (7) to (9), and when the film had become absorbed into the wood leaving a surface moist to the touch, the prepared MgO (in ether suspension) was applied to the moist surfaces to supply about 6 mg of MgO per cm², and the ether allowed to evaporate, after which the cement solids of the control tests were applied. The laminates were quickly assembled and held clamped for 30 minutes, after which the boards were air-dried for 7 days. After slitting lengthwise, one specimen was tested for binder shear strength and the other tested in static bending.

The very high shear strength observed data indicated benefit to be obtained by using an initial minor loading of very small particle sizes before applying the major portion of the cement solids, as compared with the control group of laminates. It is of interest that both the control laminates and the laminates of tests (13) to (15) showed MOR values for cottonwood and spruce exceeding those for whole wood, which for the three species are respectively 600 kg/cm², 824 kg/cm² and 718 kg/cm². The gain may be ascribed to the combined enhancement of strengths by penetration of oxyphosphate compounds into the ply surfaces, and the residual ammonium phosphates crystallizing within the interior portions of each ply. The cell character and sequestered materials in Douglas-fir may account for the inferior MOR and MOE measurements of laminates.

Test laminates (16), (17) and (18) followed the same procedures as recounted above for tests (13) to (15), except that the alkaline earth metal compound applied in minor amount was CaO, as a suspension in ether, to supply about 7 milligrams per cm² of veneer face area. The test data in TABLE VIII indicates that such compound improves the strength properties as compared with the control tests.

Further test laminates (19), (20) and (21) repeated the procedures of tests (13) to (15), except that an aqueous suspension of Ca(OH)₂ in amount to supply about 9 milligrams per cm² of veneer face was applied. The strength measurements indicate that such compound improves the loading properties of the mineral binder mass into the surface wood zones.

TABLE VIII
__________________________________________________________________________
WOOD
FAILURE
STATIC BENDING
SURFACE         IN    MODULUS
MODULUS
COTTON-
DOUGLAS-     MOISTURE
LOADING  SHEAR  SHEAR OF     OF
WOOD  FIR    SPRUCE
CONTENT
AMOUNT   STRENGTH
AREA  RUPTURE
ELASTICITY
Test No.           %      mg/cm2
kg/cm2
%     Kg/cm2
Kg/cm2
__________________________________________________________________________
1                 8.5    *Amm. Phos.
0,6#x2205;6
nil   --     --
2           8.0    Amm. Phos.
0,5#x2205;5
nil   --     --
3    8.0    Amm. Phos.
0,37 #x2205;37
nil   --     --
4                 10.2   Amm. Phos.
0,6#x2205;6
nil   --     --
6.86 mg Mg(OH)2
5           9.1    Amm. Phos.
0,6#x2205;6
nil   --     --
7.80 mg Mg(OH)2
6    10.0   Amm. Phos.
0,55#x2205;55
nil   --     --
7.2 mg Mg(OH)2
7                 7.8    Amm. Phos.
4.36  15    1030   12,600
80 mg cement+
8           6.5    Amm. Phos.
4.60   5    670     9,700
80 mg cement
9    7.2    Amm. Phos.
6.21   5    150    11,900
80 mg cement
10                 10.2   Amm. Phos.
13.08  40    1150   13,200
6.86 mg Mg(OH)2
80 mg cement
11           9.8    Amm. Phos.
12.04  15    990    12,000
7.80 mg Mg(OH)2
80 mg cement
12    10.0   Amm. Phos.
11.28  25    860    11,600
7.2 mg Mg(OH)2
80 mg cement
13                 10.2   Amm. Phos.
12.4   50    1100   12,800
6 mg MgO
80 mg cement
14           9.8    Amm. Phos.
11.8   20    870    10,300
6 mg MgO
80 mg cement
15    10.0   Amm. Phos.
8.6   15    706     9,800
6 mg MgO
80 mg cement
16                 10.2   Amm. Phos.
9.3    7    1050   12,700
7 mg CaO
80 mg cement
17           9.8    Amm. Phos.
8.7    6    715     9,900
7 mg CaO
80 mg cement
18    10.0   Amm. Phos.
8.8   17    890    12,100
7 mg CaO
80 mg cement
19                 10.2   Amm. Phos.
12.7   22    1080   13,100
9 mg Ca(OH)2
80 cement
20           9.8    Amm. Phos.
11.1   12    741     9,900
9 mg Ca(OH)2
80 mg cement
21    10.0   Amm. Phos.
10.6   17    802    12,800
9 mg Ca(OH)2
80 mg cement
__________________________________________________________________________
*Deposn. of Amm. Phosphate solution supplying 23.8 mg P2 O5 per
cm2, all resn.
+ Cement solids 20 mg MgO and 60 mg MgCO3 (new document) - 100
mesh no + 200 mesh.

EXAMPLE X

A number of woody plant materials of low density, such as the straws of cereal seed plants (wheat, oats, rye, barley, millet) and stalks of oil seed plants (rapeseed, flax, cottonseed) which have relatively low strengths and may include pith portions, all of very low bulk densities, may be used to form useful products which, though being generally strong enough to be handled and transported without support, are not sufficiently strong in compression or bending to be properly called "structural" products, due to the very low formed densities when assembled without compaction. The chief utility for panels formed from such materials at densities under about 0.35 is as insulative forms intended to occupy space between inner and outer structure walls of cold storage chambers or as wall, ceiling and roof insulation in houses.

To evaluate the properties of extremely low density panels which may be formed from optimum arrangements of straw, as low as about 0.10 bulk density, having fire-retardant properties, mature stalks of winter rye were harvested from the field just above ground so as to avoid bending or crushing the stems, and the heads were removed, leaving average stalk lengths of 91 cm, relatively free of leaf residues. It was found that a mold box measuring 2.5 m by 0.50 m by 15 cm depth would accept parallel-laid straw weighing 11000 grams in air-dry state, alternate layers being reversed so that butt-ends overlaid head-ends, and vice versa.

The 11000 grams of straw was laid out on a flat surface as a layer about 6 cm deep and 11270 grams of aqueous ammonium polyphosphate solution was sprayed, in stages, to apply a surface wetting film. The solution was commercial grade 10:34:00 material of 50% solids, solution analysis 34% P₂ O₅ as mixed ortho, pyro, tri- and tetra-polyphosphates of ammonia, and included a minor portion of higher polyphosphates. The straw was rolled over three times while maintaining parallelism as spraying continued. The straw was again laid in shallow, parallel layer and 2400 grams of cement solids comprising 11270 grams of dead-burned, fine-grained magnesia of density 3.36 and 9500 grams of silica grain of density 2.63, all being of particle sizes passing 150 mesh screen and retained on +250 mesh, was sifted in stages onto the wetted straw. The straw was rolled up into a bundle after about one-third of the cement solids have had been applied, then laid out flat while a second one-third portion was sifted onto the stalks, again rolled into a bundle, spread, and the final portion of cement sifted. The application was completed within about 90 seconds from the initial contact of cement with the film. All parts of the straw appeared to have a continuous adhered coating of the powder.

The coated straw was lifted in portions which were spread uniformly in parallel arrangement, aligned with the 2.5 m direction of the mold, and further portions were arranged similarly as layers. A sheet of paper was laid upon the straw, and a polished, waxed mold plate was laid upon the paper, exerting only a few grams pressure per square centimeter of plan area, i.e., without compacting the straw.

The formed product was able to be removed within a half-hour and was cured by drying on a rack for 6 days at low relative humidity without application of heat at room temperature. The formed product has had a gross weight when cured of 38.15 kg representing a bulk density of 203.45 kilograms per cubic meter.

The straw could not be ignited by being exposed to direct flame of a gasoline blow-torch for 10 minutes, but charred in the heated zone.

Such panels are remarkably strong in bending despite being of very low density, due to the very long fragment length and lap joint area. The hollow tubular stems represent a favorable cross-section for resisting conduction of heat transversely of the assembly. The product stiffness was such that it could receive plaster or stucco, yet its stiffness was not too great to prevent the panel from being pressed as a friction-held insert between framing members.

Because this panel was far stronger than necessary, a further panel using one-half of the cement solids, and about 60% of the solution used in this example, was prepared as described below.

EXAMPLE XI

Rye straw as in the preceding example was weighed to provide a batch of 11,000 grams in air dry state. Following the wetting by 65,000 grams of ammonium phosphate solution, the wetted stems were clad with powdered cement solids as before, and the board was cast and cured, using a kraft paper liner in the mold bottom and a bleached light calendered card sheet of 0.4 mm thickness as face sheet, both sheets being left untreated otherwise than by capillary absorption of liquid from the straw in contact with the sheet surface.

The dried panel has a weight of just under 25 kilograms, representing a bulk density of 133.34 kg/cm³. Even at this low density the binder formed on the free straw surfaces was of a thickness about 75 microns, and contiguous stems were well bonded by bridging deposits of mineral binder. The panel was amply strong for its purpose as insulation, and had good fire retardant properties.

The foregoing examples illustrate that many plant stalks and stems which are of low bulk density, notably reed grasses and cattails, may be assembled economically to form relatively rigid and strong composites having good insulative properties. Where such composites are integrally cast with a strong outer face sheet such as a wood veneer, a hardboard, or card or paper stock, the unit serves limited structural applications as well as being an insulating body.

Ammonium polyphosphate in solid form is also useful as impregnating salt and in one-step admixing of calcium or magnesium compounds with fragments, salt, and water. The utilisation of a solid such as TVA 11-55-0 or 12-54-0 powdered salt can be equally effective, but different procedures are indicated as compared with aqueous ammonium polyphosphate solutions, as shown in the following example.

EXAMPLE XII

Two batches of composites using Red Alder air-dried wood flake fragments of average thickness 1.3 mm and width and length dimensions 2 cm by 6 cm were made up, one by the single-step procedure of admixing salt, cement solids, and wood fragments then adding water, the other by first making an aqueous solution from the salt, and combining the solution with mixed flakes and cement solids.

In the first batch, 69 grams of 11-55-0 salt were added to 215 grams of air-dry wood fragments of 8.3% moisture content, together with 250 grams of particulate cement solids comprised of 25% dead-burned dense MgO of density 3.36 and 75% of dolomite both solids being of a range of grain sizes between 149 microns and 55 microns and having a packing ratio in tamped bulk form of 0.87. The mixture was stirred while 100 grams of water was sprayed onto the mix, and 30 seconds later the batch was turned out into a mold, raked and levelled, and compacted under 3 kg/cm² pressure between stainless steel mold plates. The product was allowed to set under pressure for 10 minutes, then released and air dried for 7 days before testing.

A second batch, using the same materials, was made by preparing a solution with the salt and the water, combining the cement solids and wood flakes, and adding the solution with mixing for 30 seconds before molding.

Both board products were adequately strong and had fire-retardant properties. The comparative strengths are shown in

TABLE IX
______________________________________
Board
Mixing    Thickness Density  MOR   MOE
Test Mode      cm        kg/m3
kg/cm2
kg/cm2
______________________________________
1    1         1.9       540      18.6  10,200
2    2         1.9       540      20.0  12,100
3*  3         1.9       870      36.9  17,500
______________________________________
*EXAMPLE VI, TABLE VI, Alder Waters.

EXAMPLE XIII

Tests were made to investigate the influence of absorbed ammonium phosphate solution on the adhesion of the mineral binder deposit to lignocellulosic surfaces, by comparing composite products made with preimpregnation by solution and subsequent loading with cement solids, with products made by a process wherein the reacting components are simultaneously applied to the fragments. The tests also investigated the extent to which an effective amount of the solution can be withdrawn into the wood surfaces from a wetting layer in which MgO grains are reacting, so that a sufficient residue of ammonium phosphate salt will remain following the completion of curing of the product to contribute fire-retardant properties.

The relative openness of hardwoods to entry of the solution despite their higher bulk density and smaller lumen diameters results from their large-section pores which form conduits extending deeply into the wood. Softwood fragments offer tracheid lumen apertures which are substantially larger than in hardwoods, but penetration beyond the first cell layer by migrating solution is obstructed by the very narrow ports at cell junctions.

Seven sawn veneer samples were prepared from air-dry Yellow Birch, which is a hardwood species high in hemicellulose content, having high strength. Seven similar veneer specimens were prepared from Douglas Fir (Coast type). All veneers had flat, smooth faces, thickness 3 mm, and length and width dimensions of 28 cm and 5 cm.

One veneer of each species was tested in static bending to establish reference ultimate bending strength, MOR and MOE for the whole wood.

One Birch veneer specimen had both faces coated uniformly with ammonium polyphosphate solution of specific gravity .Badd.1,40.Baddend. 1.40 and analysis 10-34-0, by a wetting layer weighing .Badd.26,4.Baddend. 26.4 milligrams per square centimeter of surface area. Such layer if applied to a non-absorbent material would have a thickness of .Badd.0,188.Baddend. #x2205;188 mm. It was quickly drawn into the veneer surface leaving a slightly moist film. Two other veneers had only one face wetted similarly. A measured weight of cement solids of grain size between -100 mesh and +200 mesh, comprised of 25% dead-burned MgO and 75% of dolomite, was sifted onto each moist surface, to supply a uniform loading of .Badd.56,3.Baddend. 56.3 milligrams per square centimeter. The coated veneers were immediately clamped to form a three-layer laminate, with clamping pressure 2 kg/cm². The clamps were removed after 30 minutes and the board dried for 5 days before testing.

An identical procedure was followed to make a board with three sawn veneers of Douglas Fir.

By a modified procedure, a 3-layer board was made with the three remaining Birch veneer pieces, and a 3-layer board was also made with the remaining Douglas Fir pieces. The amounts of ammonium phosphate solution and of cement solids calculated for each face as explained above were first mixed briefly together in a round-bottom bowl until a viscous slurry was formed, and this material was quickly transferred to and spread evenly on the face while the veneer was seated in a jig having side flanges so that a straight-edge could be used to trowel and strike off an even layer about .Badd.0,3.Baddend. #x2205;3 mm thick. The coated faces were immediately assembled, clamped under 4 kg/cm² compaction pressure, and cured as described hereinabove.

The boards were tested to their ultimate bending strength, obtaining MOR and MOE values. All boards failed in tension wood fracture of the bottom layer.

An end portion of the test board then was grooved by a pair of transverse grooves, one on each outer veneer and the grooves being spaced 5 cm apart along the board, severing two wood layers. The strength in shear was determined by applying a steadily increasing tension load to stress the bond layers in shear. The data is presented in Table X below.

TABLE X
__________________________________________________________________________
COMPARISON OF PRODUCTS MADE BY ONE-STEP MIXING
AND BY PRE-IMPREGNATION FOLLOWED BY CEMENT LOADING
Wood Ultimate
Shear
Failure,
Static
Mod. of
WHOLE WOOD
Strength
% of Bending
Elast.
MOR MOE
SPECIES
Process
kg/cm2
face kg/cm2
kg/cm2
kg/cm2
kg/cm2
__________________________________________________________________________
YELLOW one- 7.88  0   590  11,000
1188
148,200
BIRCH  step
pre- 17.20
15   870  16,000
1188
148,200
impreg.
DOUGLAS
one- 4.20 0-5  330   7,900
844
135,700
FIR    step
(Coast)
pre- 14.64
20   710  10,900
844
135,700
impreg
__________________________________________________________________________

The high shear strength and the high MOR of the Birch laminate product made with pre-impregnation of ammonium phosphate solution indicates the ability of the junction bond mass to resist longitudinal shear, which is directly related to the extent to which the mineral solid has become anchored physically into the hardwood surface.

The lower MOR value obtained with each wood species when the reactant materials are admixed before application to the wood points to the formation of the magnesium oxyphosphate product which is less intimately associated with the surface of the wood. The migration of the colloidal oxyphosphate under osmotic gradient toward the wood is apparently impeded so that crystal formation within the near-surface zones is less well developed by the surface application of mixed cement solids and ammonium polyphosphate solution which have already commenced to react with each other.

EXAMPLE XIV

To test the durability of composite material cast using wood fragments, ammonium polyphosphate solution and cement solids consisting of dead-burned dense magnesia and silica grain, board products were molded from a mix using flakes of dry poplar of average thickness 1.3 mm and widths and lengths 2 cm by 6 cm. The magnesia was of grain sizes ranging between 105 microns and 53 microns and of density 3.36; the silica was screened to the same sizes and had specific gravity 2.62. The flakes were wetted with ammonium polyphosphate solution in the proportion 112.5 grams to 215 grams of wood flakes to 250 grams of cement solids, of which 125 grams comprised magnesia grains. The solids were applied to the wetted fragments with agitation by tumble-mixing until all free powder was adhered to the flakes.

The mix was molded at 4 kg/cm² pressure between stainless steel polished plates. The plates were removed and the board taken out at 10 minutes time, and cured at room temperature for 6 days when the board was fully air-dried.

The board was sawn into two parts and one was tested in static bending, having a MOR of 25.7 kg/cm² and MOE of 12,340 kg/cm², at density 0.551.

The other part was alternately immersed in tap water for 15 minutes, drained and re-dried in air for 3 days, then immersed in sea water for 15 minutes, drained and dried for 3 days, and the cycle repeated for 21 days, ending with prolonged immersion and rinsing in tap water before final air-drying. The weight in air-dried state as measured before the first immersion was compared with its weight at the end of the cycle.

A slight reduction (7%) in MOR was measured, and very low weight loss (3%) recorded for the weathered portion.

The relatively high bonding rate will make the present invention especially useful for continuous casting and molding of high strength light weight hollow circular or rectangular tubular bodies especially such as used for drainage tiles, culverts and longer pipe sections. Usefulness of this invention for such application becomes especially evident as porosity and permeability to fluids is made technically possible by selecting manufacturing conditions under which porous light weight tubes such as required for low pressure underground irrigation can be had. On the other hand, by selecting conditions under which migration of colloids to surface is promoted closed surface drainage tiles and culverts can be fabricated. Similarly, for poles where high bending strength and light weight might be important assets the continuous molding process may include low rpm spinning of the mold during or after casting to cause simultaneous densification of the wet mass and colloid migration to the surfaces on account of the centrifugal forces created by the spinning motion. By variation of this technology a large variety of useful molded products can be formed.

EXAMPLE XV

To 400 g of air-dry spruce wood fragments, passing 5 mesh and retained on 20 mesh screens, 700 g of green wet process ammonium polyphosphate solution of 1.4 specific gravity having 50% salt solids and the solution analysing 10:34:00 nitrogen (as NH₄) and P₂ O₅ was added with mixing. The phosphate-wet fragments were allowed to stand for 5 min whereupon they were dusted with 1200 g of cement solids comprised 2.25 parts of ground dolomite in amount of 900 g and 0.75 parts of dead burned MgO in amount of 300 g both in powder form passing 100 mesh and retained on 250 mesh screens. The resulting mass was briefly mixed to obtain uniform coating of the wet fragments before it was filled into a 45 cm long plexyglass tube mold having a 15 cm inside diameter and fitted with a centre core piece of 11 cm. The mix was allowed to cure for 20 min inside the tube whereupon it was carefully removed and air-dried for 7 days. The hollow tube had a density of 0.53 and weighed 1967 g. It could carry a uniformly distributed load of 2 kg/cm length without breaking and was relatively pervious to water over its full length.

EXAMPLE XVI

The same mix as described in EXAMPLE XV when filled into the same mold and the mold spinned at 3 rps for 5 min was cured into a hollow tube having a wall thickness of 1.2 cm and a density of 0.85. The product was strong and contrary to the previous product it was impervious to water.

EXAMPLE XVII

In a subsequent trial larger tubular sections were cast with ease using air-dry spruce wood slivers of 0.5 to 3 cm length and maximum diameter of 3 mm. The proportions of cement solids and ammonium polyphosphate were similar to those described for the foregoing examples. The plexyglass tube measured 30 cm inside diameter and was 50 cm long with a 15 cm solid core positioned exactly in the center. The assembly was filled with the freshly mixed mass and rotated simultaneously at a speed of 5 rps. Following removal from the mold and airdrying for 7 days the section was found to have a density of 1.25 and a wall thickness of 9.7 cm. The calculated MOR for the section was 580 kg/cm² and should have had a 372 kg allowable bending load capacity on a 10 m section when held at one end. Obviously stronger sections could be fabricated with judicious use of stronger fragments such as oriented flakes that could be "wrapped around" the outside circumference resisting the high stresses and by incorporating high strength fibers and steel reinforcement in the high-tension areas.

Claims

1. #3# The A method of making a molded composite product including by joining a plurality of individual members, each of which is made of a ligneus ligneous plant material with surfaces having internal pore spaces, which is bonded with a mineral binder which comprises the method comprising the steps of:

(a) providing applying an aqueous solution of an ammonium phosphate or ammonium polyphosphate on the plant material so as of the individual members in a manner to allow the aqueous solution to be absorbed within the pore spaces and to wet the surfaces, along with and a particulate alkaline earth metal oxide, hydroxide, or carbonate on the wetted surfaces which reacts with the ammonium phosphate or ammonium polyphosphate in the pore spaces and on the wetted surfaces to form an alkaline earth metal oxyphosphate wet paste as the a binder within the pores and as a wet paste coating on the surfaces surface; and

(b) molding curing the wet paste coated material until the oxyphosphate is solidified within the pores and on the surfaces to bond with the ligneous plant material and to join the individual members to form the composite product.