Carbon Dioxide Application for Rapid Production of Cement Particleboard. Robert L. Geimer, Forest Products Research Technologist

Transcription

1 Carbon Dioxide Application for Rapid Production of Cement Particleboard Robert L. Geimer, Forest Products Research Technologist USDA, Forest service Forest Products Laboratory One Gifford Pinchot Drive Madison. Wisconsin Mario R. Souza Laboratorio de Produtos Florstais Brasilia, Brazil Ali A. Moslemi, Professor College of Forestry University of Idaho. Moscow, Idaho Dr. Maruli H. Simatupang lnstitut fur Holzchemie Hamburg, Germany November 1992 Prepared for publication in Proceedings, lnorganic Bonded Wood and Fiber Composite Materials Conference, Spokane, WA, Sept 27-30, 1992 Keywords: Carbon dioxide, gas injection, cement particleboard, inorganic bonded, species compatibility, pressing The Forest Products Laboratory is maintained in cooperation with the University of Wisconsin. This article was written and prepared by U.S. Government employees on official time, and it is therefore in the public domain and not subject to copyright

2 ABSTRACT Conventional pressing techniques in the manufacturing process of cement-bonded particleboard require an 8- to 24-h damp time. By injecting CO 2 into a cement-bonded particle board, board compaction pressure can be released in 4-1/2 min. The carbon dioxide treatment also reduces the inhibiting effect that many species have on cement hydration. The objective in this study was to determine the effect of changes in fabrication and pressing variables for cement-bonded wood composites on composite properties immediately after pressing and following a 28-day curing period in boards pressed using a carbon dioxide injection pressing technique. Two gas injection pressing methods, along with variations in board specific gravity, wood/cement ratio, and water/cement ratio were investigated, Weight gain, an indicator of gas utilization, was dependent on board permeability and could be related to the extent of wood compaction. In this study, boards made with Southern Pine had a bending modulus of elasticity (MOE) value 1.9 times greater and a bending modulus of rupture (MOR) value 25 times greater than that of boards pressed in the conventional manner. Bending MOR increased with increases in amount and compaction of wood. whereas bending MOE was more dependent on amount of cement. Increasing the water/cement ratio reduced gas penetration and lowered both MOR and MOE values. INTRODUCTION The setting (hydration) of cement involves a succession of overlapping crystallization stages, unlike the curing of thermosetting resins, which is dependent on heat-activated molecular polymerization and cross-linking (9). In the manufacture of a cement-bonded wood composite, the hydration process normally requires between 8 and 24 h to develop sufficient board strength and cohesiveness to permit the release of consolidation pressure. Early work by Berger and others (3) showed that exposure of portland cement to carbon dioxide gas (CO 2 ) reduces the duration of the initial hardening stage. The phenomenon results from the chemical reaction of CO 2 with calcium hydroxide (Ca(OH) 2 ) to form calcium carbonate 2

3 (CaC0 3 ) and water (H 2 C) as shown in Equation (1). (1) The formation of CaCO 3 provides a bond in the initial stages of hydration and precedes other crystallization phases that occur during the next 14 to 28 days. Replacement of up to 5 percent of the cement with Ca(OH) 2 promotes the reaction, permitting removal of mat consolidation pressure in less than 5 min. The CO 2 cure process was applied commercially to the production of a cement-bonded flakeboard in Szombathely, Hungary (1). But, obtaining uniform gas distribution throughout the board was an explicit technical problem, which we believe may have been a major concern. Gas permeability is quite sensitive to three fabrication variables: board specific gravity (SG), water/cement ratio (water/c-r), and wood/cement ratio (wood/c-r). If we assume the cell walls are completely saturated, the SG W of the compacted wood in a pressed cement-bonded particleboard is only changed by the decrease in size of the cell lumens and can be calculated as (2) where is wood/c-r, target ovendry SG of hydrated board, and water/c-r of the unhydrated mat (exclusive of water of saturation). This and other related equations are displayed in Appendix A. Figure 1 depicts the compacted wood SG W for two board SGs, two water/c-rs, and two wood/c- Rs from 0.2 to 0.4. The compacted wood SG encountered in the principal portion of this experiment varied from a low of 0.40 in a 1.2 SG board with a water/c-r of 0.25 and a wood/c-r of 0.28 to a high of 0.76 in a 1.4 SG board with a water/c-r of 0.35 and a wood/c-r of Recognizing that gas permeability is not a 3

4 straight-line function of either board or compacted wood SG, we realized that small changes in these variables would cause large differences in the results of gas injection press schedules, thus requiring special equipment or techniques for different boards. Press schedule timing is critical to assure good gas distribution without precure of the cement. Two methods of introducing steam into a resin-bonded board applied to the study reported here (4). In the first method, a sealed system utilized a spacing frame fitted with a gasket around the edge of the platens to create a sealed chamber between the platens. Gas was introduced into the chamber either through the spacing bars or through perforations in the platen. In the second method, which was an unsealed system, the board itself acted as the seal and gas was introduced through perforated platens into the face of the board. The objective of this study was to determine the effect of changes in cement-bonded wood composite fabrication and pressing variables on the development of composite properties immediately after pressing and following a 28-day curing period in boards pressed using a CO 2 injection pressing technique. Three fabrication variables (board density, water/c-r and wood/c-r) and two pressing systems (sealed and unsealed) were investigated. CONSIDERATIONS An intensive preliminary investigation was necessary to delineate the general responce of cement-bonded wood composites when exposed to CO 2. During the course of this preliminary work (72 boards), we investigated many combinations of press schedules, in conjunction with sealed and unsealed systems, to determine the proper time to introduce the gas, the duration and pressure of gas injection, the proper use of vacuum by itself and in combination with gas injection, and the effect of press closure speeds. We also examined the injection of nitrogen into the mat, prior to reaching final board thickness, as a means to establish better distribution patterns without incurring cement precure. In addition, fabrication details such as water\c-r, wood\c-r, board density, mat formation time and duration of closed 4

5 assembly stage, and the use of screens and edge densification devices were scrutinized. Finally, we made numerous modifications and changes in blending, forming, and pressing techniques. From this preliminary investigation, we defined the range of the variables to be investigated and determined fabrication and pressing techniques that produced meaningful results. Similar but not identical press schedules were developed for each pressing system (sealed and unsealed). We also found it necessary to change gas pressure with changes in board SG. The formal experimental design investigated the difference between boards of two SGs (1.2, 1.4), two water/c-rs (0.25,0.35), and two wood/c-rs (0.28, 0.35). The preliminary investigations and the formal experiments are discussed in separate sections. PRELIMINARY INVESTIGATIONS We decided to use Southern Pine as a raw material because it was readily available and provided the opportunity to observe the beneficial effect of CO 2 injection on a species known to be only partially compatible with cement (6, 10). Southern Pine was found to vary in its inhibition of the hydration of cement by Weatherwax and Tarkow (13). Hydration tests we made on the pure cement and on a Southern Pine and cement mixture confirmed observations by Sanderman and Kohler (8), which classified Southern Pine as moderately inhibitory to cement cure (Fig. 2). Physical properties of cement-bonded wood composites normally take 28 days to fully develop. Time constraints prevented us from waiting for test properties to develop on fully hydrated boards to aid in determining appropriate board fabrication and pressing techniques. Our judgment of the relative effect of changes in fabrication variables was based on maximum board temperature, hydration rate, weight gain, visual appearance, panel thickness, surface roughness, and a bending test, conducted shortly after pressing, on the untrimmed board. No effort was made to optimize gas usage. We discovered gas leakage in the 914- by 914-mm perforated platens, which were fitted with a sealing ring. We felt that this condition would not affect board 5

6 properties when pressing in the sealed mode but it was restrictive to use in an unsealed system. Consequently, unsealed boards were pressed using a smaller, 356- by 356-mm platen. To provide the heat of vaporization in converting the CO 2 to a gas. we used a 1,500-W heater. This heater was marginal at the maximum flow rate of 10 g/s experienced while charging the manifold and platens but was adequate for the 2- to 6-g/s rate used during the main gassing period. To obtain adequate gas penetration in the high-density (1.4 SG) boards, we elected to double the gas pressure from 600 to 1200 kpa. These gas pressures were used when pressing 1.4 SG boards in both the sealed and unsealed systems. For most preliminary investigations and all principal studies, we used a fine-meshed brass screen for the bottom caul as a convenient method of transporting the loose mat into the press. We saw no advantage to using the screens as top cauls in either sealed or unsealed systems or to using screens of larger mesh for the bottom caul. We found that a flat silicone gasket, in the shape of a 3.2-mm-thick by 20-mm-wide square frame, placed on the top surface of an unsealed board densified the edges of the mat and reduced the consumption of CO 2. No gasket was used in making boards for the formal portion of this study. However, a frame supporting the bottom caul screen afforded some measure of edge densification. Our investigations showed that boards of poor quality were produced when no Ca(OH) 2 was added and board quality was marginal with a 2 percent addition. The remainder of our studies were conducted with 5 percent of the cement replaced by Ca(OH) 2. Early in our work, we found that the reagent-grade Ca(OH) 2 could be replaced with lower cost agriculture-grade slaked lime with no detrimental results. All our work indicated that an increase in board SG would improve properties as long as good gas penetration was obtained. Decreasing wood/c-r from 0.35 to 0.28 permitted easier gas penetration into the board and resulted in good initial properties. Decreases in the water/c-r were favorable to board properties and quality. Hydration rate. as indicated by maximum temperature and weight gain, was improved when water/c-r was decreased from 0.35 to 0.25 and was severely restricted at a water/c-r of

7 In several trials, the CO 2 was injected during the press closure stage. The results were marginal and confirmed earlier Investigations (10), which indicated that very precise press schedules would be needed to prevent precure. Trials were also conducted using nitrogen pregassing during the dosing stages. We believed this would permit easier entry of the CO 2 into the board after final board thickness had been reached. However, no advantage could be detected and since the procedure also involved complicated equipment and press schedule changes, we elected to continue introducing the CO 2 after press closure. We found, however, that if the gas entry was delayed until after press closure, an early vacuum stage was essential to obtain sufficient gas penetration. An initial period of 30 s with a vacuum pulled on the bottom of the board and gas introduced into the top platen was adequate. Following this gas-vacuum period, our programs incorporated a stage wherein gas was introduced to both sides of the board. Longer gas exposure times were beneficial in all cases. Decreasing the second gassing stage from 120 to 60 s proved detrimental in most cases. PRINCIPAL STUDIES Procedures The Southern Pine particles used in this study were donated by Georgia Pacific Corporation 1 (Russelville, South Carolina). Particles were screened to obtain those passing Tyler mesh 6 (3.36 mm) and retained on mesh 20 (0.84 mm). Average moisture content (MC) of the wood was 7 percent. Table 1 summarizes the conditions used in the 2 x 2 x 2 x 2 experimental format. Four replications were made for each board type. Furnish calculations were based on the ovendry (OD) weight of a hydrated board assuming that the water of hydration amounted to 25 percent. As an example, a board constructed to a wood/c-r of 0.35 contained 1 part of cement, 0.35 part of OD wood, and 0.25 part of water of hydration, or 62.5, 21.9, and 15.6 percent, respectively. Two wood/c-rs, 0.28 and 0.35, and two board SGs, 1.2 and 1 The use of trade or firm names in this publication is for reader information and does not imply endorsement by the U.S. Department of Agriculture of any product or service. 7

8 1.4, were used in the principal study. All CO 2 -injected boards were produced with 5 percent of the required type I portland cement replaced with an agriculture-grade Ca(OH) 2. The CO 2 -injected boards were fabricated to water/c-rs of 0.25 and The water added to the furnish was based on the total weight of the cement and the Ca(OH) 2. This was in addition to a quantity of water (30 percent minus the measured MC, OD wood basis) necessary to saturate the wood. Materials were mixed in a small, 45-cm-diameter by 50-cm-long drum blender rotating at 40 rpm. First, the water, containing the dissolved Ca(OH) 2, was added to the wood. After mixing for 3 min, the cement was sifted onto the wet wood and the entire furnish mixed for an additional 5 min. Uniform blending was a prerequisite to obtaining even gas distribution in the press. Boards measuring 360 by 360 mm were fabricated to target SGs of 1.2 and 1.4 (OD hydrated board basis) and a thickness of 14.1 mm. A fine brass screen crimped within a steel frame was used as the lower caul plate. A dosed assembly time of 5 min between forming and pressing allowed time to prepare the press and eliminated major variations in initial hydration between boards. Boards were cold pressed in a computer-controlled hydraulic press (Fig. 3). Thickness was controlled electronically for those boards pressed in an unsealed mode. When the press was sealed, the sealing frame controlled thickness. A 120-s, two-sided gassing step was used in the formal experiment for both sealed and unsealed systems. Press schedules defining the gas injection stages for both systems are given in Tables 2 and 3. Pressurized cylinders of pure dry carbon dioxide were regulated to provide a pressure of 600 kpa for the 1.2 SG boards and 1406 kpa for the 1.4 SG boards. A 1500-W electric gas heater maintained a temperature of approximately 23 C in the platen manifold. Gas and board temperatures were monitored throughout the press cycle. Control boards made at SGs of 1.2 and 1.4 were all constructed at a wood/c-r of 0.35 and a water/c-r of No Ca(OH) 2 was added to the furnish. Instead, 5 percent of the cement was replaced 8

9 with aluminum sulfate (AISO 4 ). which acts as an accelerator. Control boards were pressed in a screwtype press for 24 h before release of pressure. Following removal from the press, the boards were measured for weight, tested for bending stiffness, and measured for thickness, in that order. The boards were then cured for at least 28 days in a controlled environment of 80 percent relative humidity (RH) at 27 C and finally were conditioned to equilibrium at 65 percent RH and 21 C prior to further testing. The initial (out-of-press) bending deflection was measured by placing a 1.8-kg weight on the center of the board supported on a 280-mm span. Bending modulus of elasticity (MOE) and modulus of rupture (MOR) (two specimens per board), internal bond (four specimens per board), and thickness swell and water absorption (one specimen per board) were determined on the fully hydrated boards according to ASTM D-1037 specifications (2). Press Schedules Results and Discussion Two attempts were made to construct control panels. The strength property values of the first eight boards after conditioning for 28 days were extremely low, which prompted us to fabricate a duplicate set of panels. These boards also had low bending and internal bond property values (Table 4), confirming the predictions made by the cement hydration tests (Fig. 2). The two press schedules (sealed and unsealed) used for the optimization work are identified in the procedure section and are shown in Tables 2 and 3. Typical curves depicting the press variables of board and gas pressure. temperature, and board thickness are shown in Figures 4 and 5 for sealed and unsealed boards fabricated to 1.2 SG, 0.35 water/c-r, and 0.35 wood/c-r. Temperature curves for the other board types were similar with the exception of a faster rise in curves for those boards containing less water or wood. A slight delay in reaching maximum temperature in the sealed boards was attributed to the lag in gas pressure rise. This in turn was due to leaks in the system and additional time necessary to charge the manifolds. 9

10 Press hydraulic pressure was allowed to vary in the unsealed system but was held constant in the sealed system to provide a pressure of approximately 3,000 kpa over the board and framing ring area. Maximum pressures needed to press unsealed 0.28 wood/c-r boards were lower with a 0.35 water/c-r than with a 0.25 water/c-r (Table 5) (statistical analysis of the data is reported by Souza (11)). This was true for both the low- and high-density series and indicated that at the 0.28 wood/c-r, wood plasticization and enhanced furnish flow caused by additional water had a greater effect on reducing maximum board pressure than the additional compaction of the wood (caused by the presence of the water) had on increasing pressures. The reverse was true with boards constructed to a wood/c-r of Additional water only increased the compaction pressure. Gas Consumption Weight gain, the increase in weight of the board during pressing, provided a relative comparison of CO 2 utilization. Weight gain is expressed as a percentage of total cement weight (Table 5). Those boards pressed in the sealed system absorbed more CO 2 than did those in the unsealed system. However, the pattern in weight gain changes between different board types was the same in both systems. Increases in either water or wood tended to reduce gas permeability and consequently reduced the CO 2 absorption, resulting in lower weight gain (Fig. 6). Within a pressing system, at the same water/c- R, the weight gain was estimated as a linear function of compacted wood SG. When the water/c-r ratio was increased from 0.25 to 0.35, the weight gain was severely reduced. In all cases, an increase in board density decreased the percentage of weight gain. This reduction in weight gain would have been even more severe if we had not elected to use twice as much gas pressure, 1,200 kpa compared to 600 kpa, with the high-density boards. Maximum board temperature is another indication of the rate and extent of the exothermic chemical reaction. However, the relation between weight gain and temperature is not well correlated (Fig. 7). Board temperature was also affected by the gas temperature, which tended to decrease with increased usage, by changes in board mass (that Is, SG), and by evaporation of the water. 10

11 The lower weight gains of those boards pressed in an unsealed system were attributed to less gas penetration. Presumably, some gas escaped through the edges of the board prior to reaching the core, whereas in the sealed system, gas entered through the edge. Total CO 2 consumption is compared to weight gain in Figure 8. No special effort was made in this study to optimize gas usage. The high consumption of gas shown for the 1.4 SG boards made in a sealed system was probably due to the leakage in the platen system. Gas flow rate should normally decrease after hydration temperature peaks because of the decreased chemical reaction as the available Ca(OH) 2 diminishes and a decrease in permeability caused by the deposition of water formed during the reaction. However, once the system became charged, flow rate was nearly linear throughout the exposure time even after the board reached maximum temperature. This indicated a constant leakage of gas. The ratio of CO 2 consumption to total cement weight (usage) and the ratio of CO 2 consumption to weight gain (efficiency) are shown for selected boards in Table 6. Gas usage varied from a low of 0.25 g CO 2/g cement for the unsealed, 1.4 SG, 0.35 wood/c-r, 0.35 water/c-r board to a high of 0.86 g CO 2/g cement for the sealed, 1.4 SG 0.28 wood/c-r, 0.25 water/c-r board. The efficiency is 100 percent if all the CO 2 reacted with Ca(OH) 2. The efficiency of the systems investigated varied over a wide range from 2.4 to 28.7 percent and was not directly related to gas usage. Gas usage was dependent on the extent to which gas loss could be restrained, whereas gas efficiency depended on obtaining gas penetration of the board. According to Equation (1) a 3.0-percent weight gain, based on total cement and Ca(OH) 2, is necessary to achieve full conversion of the 5-percent cement replacement of Ca(OH) 2 used in this study. Excess weight gain was obtained for all board types except the two made at a SG of 1.4 with a 0.35 water/c-r in an unsealed system. The greatest weight gain. between four and five times that accounted for by the added Ca(OH) 2, occurred in low-density boards made at 0.25 water/c-r in a sealed system. Initial Board Stiffness As expected, initial bending MOE values measured directly after pressing increased with board SG (Table 7). However, with the exception of the 1.2 SG unsealed boards, initial bending MOE values 11

12 decreased with an increase in compacted wood SG, which implies that MOE was more dependent on the amount of cement than on the compaction of the wood. This relation was similar to that reported by Moslemi and Pfister (7). Initial MOE values for the sealed and unsealed boards are combined and averaged in Figure 9. The bars are presented in order of increasing compacted wood SG, the same sequence (A,B,C,D) that occurs in Table 7. Obviously, MOE did not respond to the board variables in the same manner as weight gain. If the bars are sequenced in order of the relative amount of cement as given in Figure 9 (that is, B, A, D, C), bending MOE value ascends with increasing weight of cement. Increasing water/c-r decreased initial MOE values. This reduction of initial stiffness was attributed to a diminished reaction between the CO 2 and Ca(OH) 2, as indicated by less weight gain. It was not dear why the boards made in an unsealed system had higher bending stiffness values with less weight gain than those boards made in a sealed system. However, bending stiffness was considered to be adequate even in those boards that had relatively low weight gains. The required initial stiffness will depend in part on the type of board and the post-pressing handling equipment. More detailed research is necessary to establish the effect of CO 2 consumption and Ca(OH) 2 addition on accelerating initial stiffness. Final Board Properties A topic of major concern was how changes in construction and pressing variables influence final board properties and to what extent the CO 2 treatment enhances wood/cement compatibility. Bending properties, internal bond, thickness swell, and water absorption averages obtained after a minimum conditioning period of 28 days are given for each board type in Table 7. Modulus of elasticity values increased with an increase in SG but decreased with increases in wood/c-r (Pig. 10). This was similar to what occurred with the initial bending modulus and indicated that MOE values are highly dependent on the amount of cement. Modulus of rupture values were maintained or increased with an increase in both SG and wood/c-r (Pig. 11). Increasing SG or wood/c-r increased 12

13 both the total weight and compaction of wood. The compacted wood SG (as estimated from Fig. 1) is often more important than the total amount of the wood in affecting strength properties of composites. (5). Increases in water/c-r theoretically increase the compaction of the wood (Fig. 1). However, with the exception of 1.4 SG boards made in a sealed system, increasing water/c-r decreased both MOR and MOE property values. This appeared to be inversely related to the extent of CO 2 reaction as measured by weight gain. In general, comparison of the weight gain and bending strength indicated that changing variables to increase the permeability of the board and permit easier entry of gas decreased the compaction of the wood and the bending strength. Bending MOE values of CO 2 -injected boards fabricated with 0.35 wood/c-r and 0.35 water/c-r were 1.1 to 1.9 times higher than those of conventionally pressed boards. Bending MOR values for comparable boards were 1.8 to 2.5 times those of conventional boards. With the water/c-r decreased to 0.25, CO 2 -injected boards were produced that had MOE and MOR property values 2.4 and 2.8 times, respectively, greater than those of conventionally pressed panels. Internal bond was quite variable. Averages for the board types ranged from 0.64 to 1.47 MPa Compared to conventionally pressed boards, the CO 2 -injected boards fabricated with 0.35 wood/c-r and 0.35 water/c-r were 2.3 to 5.9 times stronger in internal bond. No definite pattern could be established for the effect of board fabrication variables on the internal bond of the boards. High variation between board types may have been the result of some combination of variables restricting gas permeation into the center of the board. Compared to boards bonded with thermosetting resin, the 24-h TS of inorganic bonded boards is quite low. The use of CO 2 injection pressing did not alter this relation (Table 7). Thickness swell ranged from 0.1 to 1.8 percent. increased with increasing SG and wood/c-r, and was highly dependent on both the amount of wood and the degree of compaction. Water absorption was influenced by board SG but showed no consistent relation to either wood/c-r or water/c-r. The relation between TS and WA was not as direct as normally experienced with organic resin bonded boards. 13

14 CONCLUDING REMARKS The CO 2 injection provides a method of reducing the pressing time of cement-bonded wood composite boards. Boards 14-mm-thick and fabricated with Southern Pine particles could be removed from the press in less than 4-1/2 min. Another major advantage of the gas system is the decrease in wood-cement incompatibility. The CO 2 -injected boards made in this study were up to 1.9 times greater in bending modulus of elasticity (MOE) and up to 2.5 times greater in bending modulus of rupture (MOR) than were similar boards pressed in the conventional manner. Weight gain, an indicator of gas utilization, varied considerably with board type but showed that, in most cases, the reaction progressed beyond that necessary to fully convert a 5-percent addition of Ca(OH) 2. Weight gain was highly dependent on gas penetration and was decreased with increasing wood compaction. Boards were pressed successfully in both sealed and unsealed systems. The sealed system provided slightly better gas utilization but produced boards that were slightly inferior in bending properties compared to boards made in the unsealed system. Bending MOR values were increased with increasing board specific gravity (SG) and wood/cement ratios and were highly dependent on compacted wood SG. Bending MOE values increased with an increase in board SG but decreased with an increase in wood/cement ratio. Modulus of elasticity was sensitive to changes in the amount and characteristics of the cement. A decrease in initial MOE and final MOE and MOR property values with increasing water/c-r was attributed to decreased weight gain. Relative to resin-bonded boards, thickness swell levels were quite low. As expected, thickness swell increased with addition in quantity or compaction of wood. RECOMMENDATlONS Research that is designed to replace or enhance established manufacturing techniques with new technology is prone at first to expose a multitude of associated questions. The results of this study 14

15 indicate that additional research in the following areas is needed: 1. Establish the effect of CO 2 consumption and Ca(OH) 2 addition on accelerating initial stiffness. 2. Establish the effect of CO 2 consumption and Ca(OH) 2 addition on the final board properties. 3. Determine if the increase in compatibility of wood and cement is relatively the same for different species of wood. 4. Define the time parameters of precure. 5. Determine the extent to which the water produced by the CO 2 reaction can be relied on to furnish necessary hydration moisture. 6. Investigate methods to obtain uniform gas distribution. REFERENCES 15

16 16

17 Table 1. Experimental plan (2 x 2 x 2 x 2) with four replications a Press system Unsealed, sealed Specific gravity 1.2, 1.4 Water/cement ratio (water/c-r) 0.25, 0.35 Wood/cement ratio (wood/c-r) 0.28, 0.35 a Conventionally pressed boards (four each (controls)) were made at specific gravities of 1.2 and 1.4. Control boards were all constructed using a water/c-r of 0.35 and a wood/c-r of

18 Table 2. Press program for unsealed system Segment no./ cumulative time (s) Press action Valve action Closing stage 1/0 Start at 130 mm No action 2/13 Linear closure at 6 mm/s until 40 mm No action 3/37 Parabolic closure at 1 mm/s until 14.1 mm No action Gassing stage 4/67 Hold at 14.1 mm Apply gas at top and vacuum at bottom for 30s 5/187 Hold at 14.1 mm Apply gas in both sides for 120 s 6/217 Hold at 14.1 mm Close valves for 30 s 7/220 Hold at 14.1 mm Vent at top and bottom for 3 s 8/250 Hold at 14.1 mm Vent at top and vacuum at bottom for 30 s Decompression stage 9/255 Parabolic opening at 500 kpa/s until 100 KPa Vent at top and bottom 10/260 Hold at 100 kpa for 5 s Vent at top and bottom 11/279 Linear opening at 6 mm/s until 130 mm No action 18

19 Table 3. Press program for sealed system Segment No./ cumulative time (s) Press action Valve action Closing stage 1/0 Start at 130 mm 2/13 Linear closure at 6 mm/s until 40 mm 3/31 Linear closure 2 mm/s until 200 kpa of pressure 4/45 Linear closure 500 kpa/s until 7,000 kpa of pressure Gassing stage 5/75 Hold at 7,000 kpa 6/195 Hold at 7,000 kpa 7/225 Hold at 7,000 kpa Decompression stage 8/241 Parabolic opening at 500 kpa/s until 100 kpa 9/247 Hold at 100 kpa for 5 s 10/266 Linear opening at 6 mm/s until 130 mm No action No action No action No action Apply gas at top and vacuum at bottom for 30 s Apply gas in both sides for 120 s on center Vent top and bottom for 30 s No action No action No action 19

20 Table 4. Physical and mechanical properties of control boards a Density (g/cm 3 ) IB (MPa) MOE (MPa) MOR (MPa) TS (%) WA (%) , , a Means of four board replications. Number of specimens per board: IB - 4; MOE, MOR - 2; TS, WS -1. Thickness swell (TS) and water absorbtion (WA) tested after 24-h soak. 20

21 Table 5. Press environment and out-of-press board measurements a Target SG Wood/ cement ratio Pressed wood SG b Canter temperature ( C) Maximum board pressure (kpa) Gas pressure (kpa) Weight gain c (%) Unsealed system - water/cement Unsealed system - water/cement Sealed system - water/cement Sealed system - water/cement , , ,525 1, ,343 1, ,763 2,293 3,884 4, ,310 1, ,304 1, ,283 1, a Measured values are average of four boards. b SG of compacted wood as per Figure 1. c Weight gain is expressed as percentage of weight of cement and Ca(OH) 2. 21

22 Table 6. Weight gain and carbon dioxide gas consumption for selected boards a Target SG Wood/ cement ratio Weight gain b (%) Unsealed system - water/cement Unsealed system - water/cement Sealed system - water/cement Sealed system - water/cement Usagec (%) CO 2 gas Efficiency d (%) a Values represent single measurement. b Weight gain is expressed as percentage of weight of cement and Ca(OH) 2. c CO 2 consumption is expressed as percentage of weight of cement and Ca(OH) 2 d Weight gain is expressed as percentage of CO 2 consumption. 22

23 Table 7. Mechanical and physical properties a Target SG Wood/ cement ratio Board SG b Initial MOE c (MPa) IB (MPa) MOE MOR TS WA (MPa) (MPa) (%) (%) Unsealed system - water/cement , , , ,660 Unsealed system - water/cement , , , ,210 Sealed system -water/cement , , ,120 Sealed system - water/cement ,600 4,800 6,300 5, ,300 6,800 6, , , , , ,500 4,000 4,900 4, a Values are an average of four board replications Number of specimens per board: IB - 4; MOE. MOR - 2; TS, WA - 1. Thickness swell (TS) and water absorption (WA) measured after a 24-h soak. b Specific gravity, ovendry basis, measured on full board after reaching equilibrium at 65 percent relative humidity and 27 C. c Measured on untrimmed board immediately after pressing. Values calculated on full width of board are for comparison only. 23

24 FIGURE CAPTIONS Figure 1--The specific gravity (SG) of compacted wood (SG W ) in wood-cement composite varies with board SG, wood/cement ratio, and water/cement ratio. Figure 2--Hydration rate curves. Southern Pine is classified as moderately inhibitory. Curves defining suitable and nonsuitable hydration rates are reproduced from Sandermann and Kohler (8). Figure 3--Press system used for CO 2 gas injection of wood-cement composites. Figure 4--Board and press conditions during pressing of CO 2 -injected wood-cement composite made in sealed system. Figure 5--Board and press conditions during pressing of carbon dioxide CO 2 -injected wood-cement composite made in unsealed system. Figure 6--Weight gain is dependent on compacted wood SG and water/cement ratio (water/c-r). Figure 7--Board weight gain is not well correlated to maxmum board temperature Figure 8--Gas consumption and board weight gain by board type. Specific gravity and wood/cement ratio, respectively: A, 1.2 and 0.28; B, 1.2 and 0.35: C, 1.4 and 0.28; D, 1.4 and 0.35 (water/cement ratio is water/c-r). Figure 9--initial modulus of elasticity (MOE) (average of sealed and unsealed system boards) is highly dependent on total weight of cement (values shown in bars). 24

25 Figure 10--Modulus of elasticity (MOE) (average of sealed and unsealed system boards) of fully hydrated board is highly dependent on total weight of cement (values shown in bars). Figure 11--Modulus of rupture (MOR) (average of sealed and unsealed system boards) of fully hydrated board is dependent on SG of compacted wood (weight of cement shown in bars). 25

26 APPENDIX A Development of equations to obtain the SG of compacted wood and the volume of voids in a cement-bonded wood composite follows: Assumptions: 1. Water of hydration H h is 25 percent of the weight of cement 2. Specific gravity of the cement is 3.0 and no voids exist between the particles of cement. This condition will be approached as the cement becomes mixed with water. 3. The SG of wood is only changed by the reduction in size of the cell lumens. It then follows that the amount of water needed to bring the wood to fiber saturation is of no consequence in determination of compressed wood SG. By design, the relative volume of a pressed board V B is equal to (1A) where Then weight is combined weight of components. where 26

27 Therefore, (3A) Also (4A) where Letting C = 1 again and substituting Equation (3A) into Equation (4A), (5A) or (6A) Since (7A) We can determine the relative volume of the unhydrated cement V and the relative volume of C u water V H (exdusive of water of saturation) in an unhydrated board. u (8A) or substituting Equation (6A) for SG B u in Equation (8A), (9A) 27

28 or Likewise, (11A) and (12A) Assuming no voids exist in the system other than in the cell lumens, the relative volume of the wood V W in the board is Substituting Equation (1OA) and Equation (12A) for V C u and V H, u (14A) so (15A) and 28

29 (16A) Replacing V W with Equation (13A) and substituting Equation (10A) and Equation (12A) for respectively, (17A) which reduces to (18A) While the SG of the woody substance is approximately 1.56 according to Tiemann (12), the saturated cell walls can only be compressed to approximately Let the relative volume of the woody substance = W/1.56 and the relative volume of the water of saturation = W (0.30)/1. Then the SG of the saturated wood (19A) which reduces to (20A) Likewise if the SG of the completely crushed saturated cell wall is 1.38, the total relative volume of voids V V in the cell is (21A) 29

30 Figure 1--The specific gravity (SG) of compacted wood (SG W ) in wood-cement composite varies with board SG, wood/cement ratio, and water/cement ratio. 30

31 Figure 2--Hydration rate curves. Southern Pine is classified as moderately inhibitory. Curves defining suitable and nonsuitable hydration rates are reproduced from Sandermann and Kohler (8). 31

32 Figure 3--Press system used for CO 2 gas injection of wood-cement composites. 32

33 Figure 4--Board and press conditions during pressing of CO 2 -injected wood-cement composite made in sealed system. 33

34 Figure 5--Board and press conditions during pressing of carbon dioxide CO 2 -injected wood-cement composite made in unsealed system. 34

35 Figure 6--Weight gain is dependent on compacted wood SG and water/cement ratio (water/c-r). 35

36 Figure 7--Board weight gain is not well correlated to maximum board temperature 36

37 Figure 8--Gas consumption and board weight gain by board type. Specific gravity and wood/cement ratio. respectively: A, 1.2 and 0.28; B, 1.2 and 0.35; C, 1.4 and 0.28; D, 1.4 and 0.35 (water/cement ratio is water/c-r). 37

38 Figure 9--Initial modulus of elasticity (MOE) (average of sealed and unsealed system boards) is highly dependent on total weight of cement (values shown in bars). 38

39 Figure 10--Modulus of elasticity (MOE) (average of sealed and unsealed system boards) of fully hydrated board is highly dependent on total weight of cement (values shown in bars). 39

40 Figure 11--Modulus of rupture (MOR) (average of sealed and unsealed system boards) of fully hydrated board is dependent on SG of compacted wood (weight of cement shown in bars). 40