In: Moslemi, Al, ed. Inorganic bonded wood and fiber composite materials: Proceedings of the 2d international inorganic bonded wood and fiber composite materials conference; 1990 October 14-17; Moscow, ID. Madison, WI: Forest Products Research Society; 1991: 18-27. thermosetting resins. Hydration tem­ Technologies for Abstract perature curves illustrating the time- Existing technologies that reduce the related process of hydration are shown rapid production of production time of gypsum, magnesia, for b-gypsum hemihydrate, magnesi­ mineral-bondedwood and portland cement-bonded wood com­ um chloride cement, magnesium sul­ posite boards are reviewed. Experi­ fate cement, and portland cement in composite boards mental data that verify these technolo­ Figure 1. The course of hydration is gies are given, and new concepts are distinguished by two stages. The first presented. Adding an ammonium lig­ stage or "open time" begins with the nosulfonate retarder, an ammonium addition of water and ends when the Marull H. Simatupang sulfate accelerator, and gypsum dihy­ binder starts to set. During this period, Norbert Seddig drate to provide a crystallization nuclei the cement paste is still plastic and can be formed. Mixing and forming of the Christoph Habighorst resulted in an open time of 4 minutes and a setting time of 2 minutes for furnish must be performed within the Robert L. Geimer boards made with b-gypsum hemihy­ open time. Prior to the start of the sec­ drate. Steam injection during pressing ond stage or ''setting'' period, pressing of magnesia-bonded particleboards re­ of the green board should commence. duced press time to 15 seconds per mil­ Pressing should continue until the max­ limeter of board thickness. Adding car- imum hydration temperature is at­ bon dioxide as a gas injection or a tained. The time from mixing the water carbonate allowed the production of with the binder until the maximum hy­ rapid-setting cement-bonded particle- dration temperature is designated as boards. A fast-hardening cement mixture total hydration time. that is not retarded by soluble carbohy­ In the manufacture of rapid-setting drates and tannins can be obtained by boards, the course of hydration is al­ nixing portland cement, high alumina tered. Generally, the setting time is Short­ cement, and b-gypsum hemihydrate. ened. However, the open time must be Wood composites manufactured with sufficiently long enough to enable pro- iorganic binders may contain 30 to 90 per mixing and forming of the furnish. percent by weight of binder compo­ Retarders and accelerators am often nents. The three most important inor­ used to prolong the open time or short- ganic binders used in such composites en the setting time; retarders and accel­ (gypsum, magnesia cement, and port- erators may be added individually or in land cement) are generally crystalline combination. In general, additives in­ in nature. fluence the hydration reaction by mod­ The setting, hardening, or hydration ifying the crystallization process. A re­ of inorganic binders is a slow process tarder hampers the solution of the when compared to boards bonded with binder, thereby delaying hydration. An

The authors are from the Federal Research Center for Forestry and Forest Products, Institute of Wood Chemistry and Chemical Technology of Wood Hamburg, Germany (Simatupang); Faculty of Natural Sciences, University of Hamburg, Hamburg, Germany (Sedding Habighorst); and USDA Forest Service, Forest Products Laboratory, Madison, Wisconsin (Geimer). Figure 1. - Hydration temperature curve of gypsum, magnesium oxysulfate cement, magnesium oxychloride cement, and portland cement.

18 accelerator may enhance hydration by: fonate is added as plasticizer to reduce proximately 15 minutes, similar to pure 1) forming a supersaturated solution the water requirement. Potassium sul­ gypsum wallboard, but production line leading to early crystallization; 2) pro- fate, water-glass solution, or gypsum speed is limited to approximately 10 ducing crystallization nuclei, such as dihydrate are commonly used acceler­ meters per minute. Any attempt to in- occurs when gypsum dihydrate is added ators. Total cure time is approximately crease the setting rate of gypsum must to b-gypsum hemihydrate; 3) increas- 15 minutes at a production line speed of consider the time needed to mix and ing temperature, thus promoting the about 30 meters per minute. form the composite ingredients. Sev­ solubility of the binder and 4) direct The strength-enhancing advantages eral methods including the encapsula­ reaction with a binder component, such of mixing wood with an inorganic tion of gypsum dihydrate with starch as occurs when carbon dioxide (CO2) is binder were foreseen as early as 1880 (3) have been used to mask the effect of added to portland cement. when a German patent was issued that an accelerator to allow ample open time. This paper reviews existing technol- described a light-weight, gypsum­ Bucking (2) reported a method spe­ ogies that reduce the production time of bonded wood wool board (17). Wood cific to the production of three-layer gypsum, magnesia, and portland ce- composite boards must be pressed or gypsum-bonded particleboards. The re­ ment-bonded wood composite boards. clamped until the gypsum has hard- quired make water for the entire board It also reports results of our research ened sufficiently to withstand the was deposited in the middle layer. In carried out in Hamburg, Germany, and springback forces exerted by the wood addition to air-dried fine wood parti­ Madison, Wisconsin, that verify these particles or fiber. Gypsum particleboa­ cles and gypsum hemihydrate, the fur­ technologies and introduce new concepts. rds are made commercially in a discon­ nish for the outer layers contained gyp- tinuous process requiring 2 hours of sum dihydrate as a nuclei builder. During Gypsum-bonded composite boards clamping time (26). The long press time pressing, moisture diffused from the Gypsum wallboard is produced in is needed because of the retarders used middle layer to both outer layers, acti­ large quantities throughout the world to obtain the required open time. Con- vating the setting of the gypsum hemi­ for use as sheathing material in interior tinuous production of gypsum-bonded hydrate. A retarder can be added to applications. These boards are made particleboard on a pilot-plant scale has regulate the setting time of the gypsum from foamed b-gypsum hemihydrate been reported by Bucking (2), but the hemihydrate in the middle layer. Ac­ (plaster of paris) and derive their bend- 6-minute press time is still too long for cording to Bucking, the setting time of ing strength from a thin layer of paper profitable commercial operation. gypsum is generally 2.5 times the open overlay Addition of additives regulate Gypsum fiberboard has less Spring- time. Since the open time was regulated the setting behavior and permit pro- back than gypsum-bonded particlebo­ to 24 minutes, the setting or pressing duction on a continuous basis using a ard and is produced commercially on a time was 6 min. In another variation of belt conveyor (36). Generally, lignosul- continuous basis. Cure times are ap- this method, water was added as gran­ ulated ice (1). In our laboratory, we examined com­ binations of ammonium lignosulfonate as a retarder and ammonium sulfate as a latent accelerator for the production of gypsum particleboards. Three meth­ ods to mask the acceleration effect were applied: 1) encapsulation of ammoni­ um sulfate with polyvinyl-propylene (PVP);2) absorption of ammonium sul­ fate in fine wood particles, and 3) ab­ sorption of ammonium sulfate in a super absorber. In all cases, an ammo­ nium lignosulfonate retarder solution was incorporated into the furnish by mixing with the wood particles. In the first method, encapsulation of ammo­ nium sulfate with the water soluble PVP proved to be very effective in mask­ ing the acceleration effect. The masking effect lasted approximately 5 to 6 min­ utes, which is too long to be used in a continuous process but may have ap plication in production with a single- opening discontinuous press. Figure 2. -Schematicof total hydration time of regulated gypsum using various In the second method, we treated a additives. Plain area represents open time; shaded area represents setting small portion of fine wood particles time. with an ammonium sulfate solution and

19 subsequently dried at 4°C. The ammo­ by reacting magnesium hydroxide with tion of inogonic-bonded wood compos­ nium sulfate was deposited as crystals carbon dioxide and heating the formed ites. The setting time is about the same as in the wood particles. The treated wood reaction product. magnesium oxychloride cement (7). particles were then mixed with the re­ Magnesium oxide is often referred to Cement prepared from caustic-cal­ mainder of the furnish and used to as magnesia. Magnesium oxide may be ined magnesia and carbonic acid is make gypsum-bonded particleboards. either caustic-calcined magnesia or dead- designated as magnesium oxycarbon­ The results of bending strength tests burned magnesia. Dead-burned mag­ ate cement. It is prepared by obtaining showed the feasibility of using a combi­ nesia is also called heavy magnesia. The magnesium carbonate trihydrate from nation of a retarder and a latent acceler­ choice of the magnesia form to make the treatment of a magnesium hydroxide ator. The ammonium lignosulfonate magnesia cement is predicated by the slurry with exhaust gas from burners or prolonged the open time of the binder chemical composition of the minor com­ engines that contain carbon dioxide. from the usual 3 to 6.5 minutes. The ponent. Caustic-calcined magnesia is The precipitated magnesium carbonate ammonium sulfate proved to be a good obtained by calcining or light burning trihydrate is filtered off, mixed with the accelerator. The setting time of the of magnesium carbonate, magnesium fiber material, and formed into a mat. boards was 3 minutes. hydroxide, or dolomite. Dolomite is a When the mat is heated above 80°C, In the third method, the ammonium naturally occurring mineral containing magnesium oxycarbonate is formed sulfate accelerator was absorbed in a qual amounts of magnesium carbon- along with by-products of carbondioxide super absorber (Liquasorb HC 9797). ate and calcium carbonate. If dolomite and water (8). Magnesium oxycarbon­ One gram of the super absorber can is calcined below 825°C, only magne­ ate cellulose fiberboards, sometimes con­ absorb up to 34 grams of a salt solution. sium carbonate is decarbonized, leav­ taining glass-fiber reinforcement, are Super absorbers are commercially used ing the calcium carbonate intact. Such currently produced in Japan. Another in the manufacture of baby diapers and material is designated as half-calcined method to make magnesium oxycar­ as a moisture retainer in sandy soils. dololomite and is also suitable to prepare bonate cement includes heating a mix­ The swollen super absorber is de­ magnesia cement. Heavy magnesia is ture of gel-like basic magnesium car­ stroyed by light pressure, releasing the obtained by high-temperature treatment bonate, magnesium hydroxide, pulp, absorbed solution. The combination of about 1650°C) of either magnesium car­ and water. This method produces a ammonium lignosulfonate in the fur­ bonate or magnesium hydroxide. lightweight board with good strength nish and ammonium sulfate in the su­ Caustic-calcined magnesia is suitable properties (34). per absorber enabled an open time of 4 to prepare magnesia cement with the Paszner (21,22) utilized heavy mag­ minutes and a setting time of slightly addition of magnesium sulfate, magne­ nesia with a water-soluble ammonium less than 3 minutes. If additional gyp- sium chloride, or magnesium nitrate. polyphosphate to produce magnesium sum dihydrate is added to the furnish, Heavy magnesia can be used in combi­ oxyphosphate-bonded wood compos­ a setting time of 2 minutes can be nation with a water-soluble ammo­ ites. The setting rate was in the range of achieved. Our results are summarized nium polyphosphate. 2 to 5 minutes. The open time was too in Figure 2. Magnesia cement prepared from caus­ short to perform proper mixing and mat tic-calcined magnesia and magnesium forming. Therefore, dolomite powder Magnesia cement-bonded sulfate is designated as magnesium was added to regulate the hydration composite boards oxysulfate cement. The chemical com­ process. Magnesium oxyphosphate­ Magnesia cement-bonded excelsior position of the set cement is determined bonded wood composites have better boards were the first inorganic-bonded by the temperature during setting (6). properties than other magnesia cement- composite manufactured on a continu­ Setting time decreases with increasing bonded composites. However, the pro­ ous production line. The binder sets temperature. Magnesium oxysulfate cess has not been industrially applied, within minutes under the influence of cement is used in the production of probably as a result of the high cost of higher temperatures (31). Magnesia ce­ Heraklith boards and Tectum boards heavy magnesia. We have made explor­ ment-bonded particleboards were first (magnesia cement-bonded wood wool atory experiments to improve the man­ described by Simatupang and Schwarz boards) (14). ufacturing process. Instead of using expen­ (32). Industrial production commencing Magnesia cement prepared from caus­ sive, heavy magnesia, we succeeded in in 1980 has been described by Loiri (13). tic-calcined magnesia and magnesium utilizing either caustic-calcined magne­ Magnesia cement is classified by Kur­ chloride is designated as magnesium sia or half-calcined dolomite. A sparly dowsky and others (10) as chemical­ oxychloride cement. Magnesium oxy­ soluble ammonium polyphosphate was boundcement. The major compound of chloride cement is faster setting than used as the minor component of the magnesia cement is magnesium oxide magnesium oxysulfate cement (Fig. 1). magnesia cement. The 12-mm-thick The minor component, about 10 to 15 Magnesium oxychloride-bonded parti­ magnesia-bonded boards were pressed percent based on magnesium oxide, is cleboards have been industrially man­ at 120°C for 15 minutes (30). A compar­ an acid radical, generally in the form of ufactured in Finland (13). Magnesium ison of these particleboard properties a magnesium salt or an ammonium oxynitrate cement, prepared by the ad­ bonded with three kinds of magnesia salt. To prepare magnesia cement, a so­ dition of caustic-calcined magnesia and cement is given in Table 1. The modulus lution of the minor component is mixed a solution of magnesium nitrate, is a of rupture (MOR) of the three kinds of with the magnesium oxide. However, relatively new cement that, as yet, has particleboards show little differences. magnesia cement can also be prepared no industrial application in the produc- However, magnesia oxyphosphate-

20 bonded particleboards swell less and (Table2). This was true for both conven­ reduce the setting time of cement-bonded are more water resistant. tional and steam-injected boards. Con­ boards to a few minutes. As previously stated, elevated tem­ ventional pressing produced boards peratures shorten the press time of par­ with higher strength properly values Addition of carbon dioxide ticleboards bonded with either magne­ than did the steam-injected boards as a When carbon dioxide is added to sium sulfate or magnesium chloride. A result of an undesirable precure in the cement according to the following equa­ specific press time of 50 seconds per steamed boards leading to lower face ions, calcium carbonate is formed: millimeter of board thickness was re- strength properties. ported (13). This time is still too long compared to that of thermosetting res­ Portland cement-bonded Calcium carbonate provides the ini­ composite boards in-bonded wood particleboards. Sima­ tial strength necessary for early release tupang (29) showed that the net press In the conventional manufacture of of the board from the press. The carbon time may be short-ended, if the preheat­ portland cement-bonded particleboa­ dioxide can be obtained from the de- ing and the decompression times are rds, the setting time is reduced by the composition of a carbonate or in­ reduced. At a press temperature of 110° addition of additives or admixtures duced as a gas. We examined both vari­ to 120°C, the specific press time was and by increasing the temperature of ations. only 30 seconds per millimeter. How- the clamped boards. The admixtures, Addition of carbonates. - The addition ever, boards with optimum strength formerly designated as mineralizing of carbonates is widely practiced in con­ properties were produced at a press agents, axe usually one of the following: crete technology. The amount of car­ temperature of 160°C. The specific press aluminum sulfate (incombination with bonate, up to about 3 percent by weight, time was longer, 60 seconds per milli­ either calcium hydroxide (lime water) based on cement, is generally added as meter of board thickness, because addi­ or water-glass solution), calcium chlor­ a powder. The make water dissolves the tional time was necessary to de-gas the ide, magnesium chloride, aluminum carbonate, and the heat of solution in- board. We succeeded in reducing the chloride, calcium formiate, or calcium creases the temperature and catalyzes specific press time for magnesium oxy­ acetate. Nearly every manufacturing the reaction. In the manufacture of ex­ sulfate-bonded particleboards to 15 plant has its own recipe or secret formu­ perimental cement-bonded partidebo­ seconds per millimeter by introducing lation. The optimum temperature dur­ ards, as much as 7.5 to 15 percent of saturated steam into the center of the ing setting varies from 40° to 80°C and either potassium carbonate, sodium board during pressing, similar to that is influenced by the chemical composi­ carbonate, or ammonium carbonate is used in thermosetting resin-bonded tion of the wood, the type of cement, added to compensate for the spring- wood composites (5,35). Using three and the use of admixtures (11). Mini­ back force of the wood particles. In the kinds of caustic-calcined magnesia as mumclamping time is 6 to 8 hours. presence of powdery carbonate, the the major binder components, we The addition of carbon dioxide and open time of the cement is between 2 found that less reactive grades resulted themodificationof portland cement are and 5 minutes. in boards with higher MOR values two methods currently being used to Portland cement-bonded partidebo­ ard made with ammonium carbonate Table 1. – Properties of magnesia-bonded particleboards manufactured with varisous acid radicals. has the advantage of being void of ad­ Water ditional cations. These cations, pro­ Thickness swell absorption duced when potassium or sodium car­ Density Internal After 2 After 24 after 24 bonate is used, may decrease the Acid radical ovendry MOR MOE bond hours hours hours long-term durability of cellulose. HoW­ ------(MPa)------(%) ------(kg/m3) ever, boards made with ammonium car­ Magnesium 1,050 to 12 to 14 3,000 0.3 to 0.6 2 to 5 6 47 chloride 1,150 bonate emit ammonia that may be a health hazard. Our limited experience Magnesium 1,050 to 10 to13 3,000 to 0.7 to 1.1 5 10 50 sulfate 1,150 3,900 with sodium carbonates showed that Ammonium 1.060 11.7 -- 0.6 2.5 44 28 boards with acceptable properties could polysphosphate be made if the sodium carbonate is free of sodium hydrogen carbonate. Sodium

Table 2. – lnfluence of magnesia activity and pressing mode on the properties of magnesia-bonded particleboards. Total Maximum Density Internal Binder type hydrationtime temperature Press mode ovendry MOE MOE bond ------(Mpa) ------(°C) (kg/m3) Martin Marietta 35 51.0 99.5 Conventional 1,090 18.5 6,330 0.9 51.0 99.5 Steam inject 1,090 16.5 5,600 0.7 Martin Marietta 50 13.4 124.8 Conventional 1,080 13.6 3,140 0.7 13.4 124.8 Steam inject 1,040 7.2 2,140 0.3 TH 12.2 118.2 Conventional 1,000 10.7 2,630 0.5 122 118.2 Steam inject 1,010 6.4 2,400 0.4

21 hydrogen carbonate causes an undesir­ ates tested. One method we used was a tassium hydrogen carbonate is tabu­ able presetting of the cement and is mixture of potassium carbonate and lated in Table 3. Press temperatures of unfortunately present inmost commer­ potassium hydrogen carbonate. By 70°C produce the highest strength cially produced sodium carbonates. varying the amount of the potassium boards. The setting time was only 5.8 Our results show that the best addi­ hydrogen carbonate, the setting time minutes. However, the actual press time tive to produce rapid-setting portland and the pH of the paste can be regulated must be approximately twice as long to cement-bonded boards is potassium (Fig. 3). The effect of press tempera- de-gas the board and prevent blows. carbonate. However, this compound is on the properties of boards manufac­ Preheating cannot be used with the the most expensive of the three carbon­ tured with potassium carbonate and po- combination of potassium carbonate and potassium hydrogen carbonate ac­ celerators because precure takes place during the open time. Another method using potassium carbonate and water-glass solution will prolong the open time at ambient con­ ditions. In this case, the press time can be further reduced if the furnish is pre- heated. We used a kitchen microwave to preheat the formed furnish. The mat was immediately pressed after heating. The combination of preheat tempera­ ture and actual press time affected the board density and resulting physical properties (Table 4). Acceptable proper- ties can be achieved in a 5 minute press time when the mat is preheated to 75°C. The board was 12 mm thick; therefore, the specific press time was 25 seconds per millimeter; setting time was 2 min­ utes. The properties of boards made with the addition of potassium carbonate Figure 3. – Setting time and pH of cement paste containing various ratios of and water-glass solution are compared potassium carbonate and potassium hydrogen carbonate. to conventionally made boards (Fig. 4). Three kinds of cements, portland (PZ Table 3. – Influence of press temperature on the setting time of cement-boned particleboard with 35F), tuff (Tuff), and high furnace (HFS), carbonate addition. were used as the binder. Samples were Press setting Density Modulus of Modulus of tested after 28 days and 64 weeks of temperature time ovendrya rupture elasticity indoor conditioning and following 32 3 ------(Mpa) ------(°C) (min.) (kg/m ) and 64 weeks of exterior exposure. The 60 6.5 1,130 8.1 2,430 results show that conventionally made 70 5.8 1,130 10.3 3,210 boards (8-hr. press time) have better 80 5.8 1,150 8.6 2,450 bending strength properties than those 95 5.6 1,180 8.5 2,540 made of potassium carbonate and wa­ ter-glass solution. The properties of conventionally made boards also pre­ Table 4. – Influence of microwave preheating temperature on the properties ofcement-bonded wood vail after exterior exposure. particleboards with potassium carbonate and water. Carbon dioxide injection. – Carbon Density Modulus of Modulus of dioxide was reported to be used in some Reheating Press time ovendrya rupture elasticity manufacturing plants to accelerate the (min.) 3 (°C) (kg/m ) (MPa) (MPa) setting of cement in the industrial man­ 50 10 1,200 9.5 2,710 ufacture of wood wool boards and Dur­ 75 10 1,210 9.7 3,510 isol elements (24). Exhaust gases from 75 5 1,170 9.1 2,700 boilers and engines were the source of 85 10 1,230 9.9 4,200 the carbon dioxide. These materials are 85 7 1,200 7.9 2,680 coarse porous, allowing the gas easy 85 5 1,010 3.7 710 95 10 unstable board access to the cement paste. Since the thickness of the cement paste in wood a Cement PZ 35F + 13 percent potassium carbonate + 1.5 percent water-glass solution. Press temperature 85°C; spruce wood particles. wool boards is about 0.3 mm (28), the reaction is nearly complete.

22 Figure 4. – Modulus of rupture of conventionally pressed and rapid-setting cement-bonded particleboards fabricated with three kinds of cement, after conditioning indoors for 28 days and 64 weeks and after outdoor exposure for 32 and 64 weeks. Conventionally pressed boards: press time 8 hours; accelerator 3 percent calcium chloride. Rapid setting boards: press time 18 minutes, accelerator 13 percent potassium carbonate plus 1.5 percent water-glass solution.

According to a British patent, carbon conductivity. The apparatus consists of dioxide is used to make calcium car­ a pressure vessel connected to a carbon bonate-bonded fiber boards with good dioxide some and a vacuum pump. A mechanical properties (16). The green slightly modified spark plug is screwed boards, made on a Hatschek machine, into the pressure vessel. Cement paste were heated to 60°C and subsequently (0.8 mm thick) is applied between the treated with carbon dioxide. two electrodes of the spark plug. When A Japanese patent describes the ap­ a vacuum is applied, the electrical con­ plication of carbondioxide toaccelerate ductivity decreases as a result of mois­ the hardening of cement-bonded fiber ture evaporation. When the specific boards. In this process, the green boards electrical conductivity falls to 0.3 ms Figure 5. – Schematic of apparatus are also heated to lower the moisture cm–1, carbon dioxide is introduced. Typi­ to monitor setting of cement paste by content and improve the porosity (15). cal conductivity at a carbon dioxide measuring electrical conductivity at According to a new process developed pressure of 7 bar is shown in Figure 6. various carbon dioxide pressures. in Hungary (9,25), carbon dioxide is As soon as the carbon dioxide is in­ injected during the pressing of portland duced, the cement paste at the surface curve is considered to be the setting cement-bonded particleboards. The rapidly carbonizes and the electrical con­ time of the cement paste. With increas­ process resembles the injection of satu­ ductivity sharply decreases. Carbon di­ ing carbon dioxide pressure, the car­ rated steam during the pressing of phe­ oxide pressure does not influence this bonization or setting time is shorter. nol formaldehyde-bonded particleboa­ initial reaction. The electrical conduc­ We made experimental boards using rds (5). tivity comes to a stop and rises slightly a press platen that enabled theinjection We have studied the basic reaction again, probably resulting from the for­ of carbon dioxide during pressing (Fig. 7). between cement and carbon dioxide mation of calcium hydrogen carbonate. A heating element was used in the car- using the apparatus shown in Figure 5. Carbon dioxide then diffuses into the bon dioxide pressure supply line to pre- This device monitors the setting of ce­ cement paste and the electrical conduc­ vent the gas from freezing. Gas was ment paste under the influence of car- tivity decreases steadily to a plateau. introduced only from the upper platen. bon dioxide by measuring the electrical The inflexion point in the last part of the A metal screen was placed on the bot-

23 tom platen. A seal was used to ensue a is pressed to the ultimate board thick­ boards, good distribution of the injected gas-tight compartment in the hydraulic ness. However, sufficient density must carbon dioxide was obtained. However, press and to prevent loss Of carbon di­ be attained prior to gas introduction to bending strength values decreased, as a oxide. Experience has shown that with prevent the mat from being blown away result of presetting the cement on the steam injection of particleboards, gas by the gas. When we used this proce­ surface of the wood particles. If carbon should be introduced before the furnish dure to press cement-bonded particle- dioxide was introduced after the fur­ nish was pressed to ultimate thickness, distribution of injected carbon dioxide was poor, as evidenced by the air pock­ ets. However, where air pockets were absent, the bending strength of the boards was good. To avoid presetting and still maintain good distribution, we introduced nitrogen during the early stages of dosing. Subsequently after pressing to the stops, carbon dioxide was introduced. In another variation, we applied vac­ uum after pressing, followed by the introduction of carbon dioxide. Before opening the press, vacuum was again applied to reduce the partial pressure and remove water vapor. Temperatures of 115°C were observed in the center of pressed boards. A comparison of board properties pressed according to various methods is shown in Table 5. Boards with good bending strength can be ob­ tained by pressing to ultimate thick­ ness, followed by vacuum treatment and injection of carbon dioxide. The influence of the carbon dioxide Figure 6. – Electrical conductivity of cement paste at a CO2 pressure of 7 bar. pressure the properties of the boards lnfluence of carbon dioxide pressure on the setting rate of cement paste. is shown in Table 6. At pressures under 5 bar, no boards with acceptable prop­ Table 5. – Influence of mode of carbon deoxide injection on cement-bonded particleboards. erties can be produced. At pressures of Maximum Carbon 5 to 9 bar, boards show comparable Mode of hydration Density Modulus of dioxide properties and hydration times. How- injectiona setting time temperature ovendry rupture distribution ever, the hydration temperature almost (min) (°C) (kg/m3) (MPa) doubles when pressures increase from Preinjection 5 to 9 bars. The reason for these large Carbon dioxide 1.2 95 1,050 8.0 Satisfactory differences is unknown. Nitrogen 1.5 79 1,090 12.5 Satisfactory Glucose and hyrolyzable tannin com­ After closing pounds, when present at concentra­ Direct 1.1 82 1,090 12.7 Unsatisfactory tions of 0.25 percent, severely retard the After vacuum 1.2 88 1,125 12.5 Good setting of portland cement paste and a Carbon dioxide pressure 7 bar; injection from upper platen. Mat endosed by a seal, provided with a valve to push out trapped air. show less detrimental effect in the pres­ ence of carbon dioxide (Table 7). Up to

Table 6. – Influence ofcarbon dioxide pressure on cement-bonded particleboards. Maximum Carbon dioxide hydration Modulus of pressure Setting time temperature Density ovendry rupture (bar) (min.) (°C) (kg/m3) (MPa) 1.5 26 61.0 - -a -- 3 1.8 66.6 - - -- 5 1.2 67.1 1,130 13.8 7 1.2 88.4 1,130 12.5 9 1.l 102.4 1,140 12.6 Figure 7. – Schematic of press platen a Unsatisfactory board. with seal for carbon dioxide injection.

24 about 0.5 percent of these compounds, based on cement weight, can be toler­ ated by the gas-injected cement paste. To determine if conventional equip ment can be retrofitted for the introduc­ tion of carbon dioxide gas, we used the sealed frame shown in Figure 7 in com­ bination with conventional press plat- ens. Screens were placed on both sides of the mat. After pressing to ultimate thickness, vacuum was applied fol­ lowed by the introduction of carbon dioxide. Using this arrangement, the setting times were about 1 minute long­ er when compared with boards pressed with the perforated platens. properties of the boards made according to both methods are comparable with each other.

Modification of portland cement The use of rapid-setting cement to manufacture wood wool board similar Figure 8. – Total hydration time of a rapid-setting cement mixture. to Heraklith magnesia-bonded boards was described by Pletzer others (23). The cement was a portland cement con­ Table 7. – Influence ofglucose or tannin on the setting of cement paste in the pressure of carbon taining 1 to 30 percent by weight of cal­ dioxide at 7 bar. cium-halogen aluminate. Another process Cement Setting time described in the patent literature enables Compound concentration Minutes Relative the production of rapid-hardening wood (%) (%) wool boards using a jet cement and ferric Glucose 0.00 4.3 100.0 chloride accelerator (33). The press tem­ 0.11 4.3 100.0 perature was between 90°C and 120°C 0.23 4.4 102.0 Another type of rapid-setting cement was 0.45 4.5 104.7 described by Shigekura (28). This binder, 0.90 5.0 116.3 a mixture of b-gypsum hemihydrate 2.25 6.0 139.5 and portland cement, sets within 20 Hydrolyzable tannin 0.00 4.3 100.0 minutes. 0.13 3.9 90.7 Okamura and others (19) used et­ 0.25 3.7 86.0 tringite and a hydraulic binder (port- 0.50 3.8 88.4 land cement) to produce wall or ceiling 1.00 4.0 93.0 boards. The ettringite was obtained by 2.50 5.5 127.9 reacting aluminum hydroxide sludge with gypsum and calcium oxide. The Sekisui Chemical Co. (27) proposed the Table 8. – Influence of glucose or tannin on the setting time and maximum hydration use of a mixture of portland cement, temperature of a rapid-setting cement mixture.a high alumina cement, gypsum calcium Cement Setting time hydroxide, and eventually sodium ci­ Compound concentration Minutes Relative Temperature trate to produce particleboards. The United Gypsum Corp. has a patent to Glucose 0.00 7.0 100.0 83.5 manufacture boards with organic ag­ 0.23 4.4 62.9 85 gregates with a binder composed of 0.45 4.2 60.0 85 portland cement, high alumina cement, 0.55 4.4 62.8 84 gypsum, and calcium hydroxide (4). 1.10 5.5 78.6 83 Odler (18) reported the continuous Hydrolyzable tannin 0.00 7.0 100.0 83.5 manufacture of a cement-bonded parti­ 0.30 7.0 100.0 87 cleboard utilizing a rapid-setting ce­ 0.50 6.4 91.4 87 ment, and Paulisan (20) submitted a 0.60 7.5 107.1 86 patent for the production of rapid-set­ 1.20 8.0 114.3 85 ting cement-bonded particleboards. a Cement mixture: 7.2 g PZ 35F + 1.6 g high alumina cement + 1.2 plaster and 4 ml water. The rapid setting of the binders report-

25 ed by Sekisui, united Gypsum Corp., times as short as 2 minutes may be Odler, and Paulisan may be a result of obtained using ammonium sulfate as the formation of ettringite during the an accelerator and gypsum dihydrate setting reaction. to enhance crystallization. However, Little data have been published in adequate open time can only be ob­ scientific journals that describes ettrigite tained using a technique such as encap­ enhanced, rapid-setting cement-bonded sulation (temporary immobilizing the particleboards. According to Kurdowski accelerator in wood particles), or a su­ and others (10), most expansive ce­ per absorber. ments (cement that expands after set­ Magnesia cement-bonded wood ting, in contrast to portland cements composites prepared by mixing mag­ that shrink) are based on ettringite. nesium oxide with either magnesium Such cements are generally mixture of sulfate or magnesium chloride experi­ portland cement and high alumina ce­ ence hydration times of approximately ment with the addition of a third com­ 150 minutes. Much faster rates are ob­ ponent (e.g., gypsum). The setting may tained by mixing the magnesium oxide be controlled by adding calcium com­ with ammonium polyphosphate. All ponents. We have made some examina­ the magnesia boards set faster when tion with such ternary mixtures. The heated. Injection of steam into a magne­ binder sets within 8 minutes at temper­ sium oxysulfate-bonded board reduced atures above 60°C (Fig. 8). The hydra­ specific preps times to 15 seconds per tion is not retarded by glucose or hydro­ millimeter. The technique tends to re­ lyzable tannins. Both compounds, at duce board strength resulting from a concentrations at or over 0.5 percent, precure effect. We also found that the showed slightly accelerating effects slower reacting magnesia types were (Tables 8). A small number of experimen­ best suited for steam injection pressing. tal boards were also made, confirming Injection of carbon dioxide gas into a reported data (18,20) that such ternary portland cement-bonded board produces cements are suitable to produce rapid- calcium carbonate that provides suffi­ setting cement particleboads. Board cient initial strength to allow the pres­ properties such as weather resistance, sure to be removed in a few minutes thickness swell, water absorption, du­ rather than hours. ThiS process is being rability against fungi and insects, and commercially applied in Hungary. The absence of efflorescence can be modified reaction is very fast and requires special by proper adjustment of the proportion techniques, such as preinjection with an of portland cement, high alumina ce­ inert gas or application of a vacuum to ment, gypsum hemihydrate, and cal­ prevent precure. The carbon dioxide cium hydroxide. can also be obtained by adding a car­ bonate to the furnish This compound Conclusions emits the carbon dioxide gas upon heat­ Known methods to reduce the pro­ ing. Cement-bonded boards made with duction time of gypsum, magnesia, and carbon dioxide tolerate much higher portland cement-bonded wood com­ concentrations of glucose and tannins posite boards are reviewed. Expen­ than do conventional boards. mental data that verify these techniques Some success has been experienced are given, and new concepts are pre­ in bonding particleboards using a mix­ sented. In many cases, reduction of set­ ture of portland cement, high alumina ting times by the addition of accelera­ cement, and b-gypsum hemihydrate. tors must be accompanied by the The binder sets within 8 minutes at addition of retarders to provide ade­ temperatures over 60°C and is not af­ quate open time to mix and form the fected by soluble carbohydrates or tan­ furnish. nins. Inorganic-bonded fiberboard expe­ riences much less springback after Literature cited pressing than particleboards. The 15 minute setting time of gypsum has per­ mitted the development of continuous production methods for manufactur­ ing gypsum-bonded fiberboard. Labo­ ratory experiments indicate that setting

26 Printed on recycled paper 27