Properties variation study of furniture grade M2 particleboard manufactured in Canada.

Abstract

In response to a lack of readily available comparative data on mechanical properties of Canadian-made particleboard, furniture grade M2 particleboard panels were obtained from six randomly selected manufacturing facilities (press lines) across Canada and tested for density

and strength properties: internal bond (IB), bending and elastic moduli, and face and edge screw withdrawal resistance (SWR). A total of 30 full-sized (1.22 by 2.44m) boards were obtained from suppliers and each divided into eight sub-panels measuring 61 by 61 cm to allow eight sets of test samples to be cut. All press lines except for one exceeded American National Standards Institute (ANSI) standard 208.1 recommended minimum IB of 0.45 MPa. Results were mixed for other strength properties; most press lines were borderline or below ANSI recommended minimum values for face and edge SWR and bending strength. Panels from all but one press line exceeded the ANSI minimum of 2250 MPa for elastic modulus. Screw type (10 or 16 threads per inch) and machine direction (particle orientation) of boards had no significant effect on SWR from the edge of particleboards; whereas there was an overall significant effect of screw type on face SWR with screws with 16 threads per inch having lower face SWR than screws with 10 threads per inch.

**********

Particleboard is the most common reconstituted wood composite panel used in the furniture manufacturing industry (Wu and Vlosky 2000). It has much greater structural homogeneity and surface smoothness than oriented strandboard (OSB) and is cheaper to produce and lower in density than its closest competitor, medium density fiberboard (MDF). The suitability and acceptance of particleboard for use in furniture has greatly improved since a feasibility study by Suchsland and Good (1968) identified some specific areas for product improvement. The major problems were difficult edge moulding and banding, fastener holding problems, low strength to weight ratio, poor water resistance, and dimensional instability. Most of these problems have been rectified by the industry-the characteristic three-layer structure, i.e., densified surface layers containing fine particles and less densified core containing coarse particles, gives furniture-grade particleboard the required surface smoothness and flexural strength-to-weight ratio for laminated modular components.

Despite these improvements in particleboard quality, recent surveys in the United States have identified high variation in the mechanical properties of particleboards across different mills, which in turn perpetuates the end-user perception of the product as low quality. A survey by Temple-Inland Panel Products (Bautista 1994, Zhang 1994) found wide variation in the mechanical properties of 40 particleboards from 25 mills representing 15 companies. The use of different wood species, particle geometries, and pressing strategies by mills were identified as the main causes of variation in properties. A survey by Cassens et al. (1994) across seven different U.S. suppliers of the same 3/8-inch M2 grade product also found significant variation in all properties except for internal bond (IB) strength. The products were supplied to a kitchen cabinet manufacturer who found that panels from two of the seven mills did not meet minimum standards. Wu and Vlosky (2000) point to a lack of information about specific customer requirements of particleboard and other furniture-grade panels hindering the development of new markets and making it difficult for existing suppliers to improve and tailor their products to customer requirements.

Most particleboard strength properties are improved by increasing panel density, which results in better inter-particle contact during pressing (Maloney 1993), but this is offset by the greater weight and dimensional instability of highly densified particleboard. The U-shaped density profile of modern particleboard is a compromise designed to provide sufficient bending strength and dense, smooth surfaces, but at the same time keeping overall panel weight and dimensional instability down. The lower density core, however, has the effect of reducing the edge fastener holding ability of particleboard (Kelly 1977, Wong et al. 1999). Over-driving of screws further compromises fastening strength (Carroll 1970), and a low-density core can be easily damaged by the insertion of screws into pilot holes of an inappropriate size (Rajak and Eckelman 1993), further contributing to low screwholding ability. Because conventional screws require higher strength embedment materials, a range of alternative 'non-screw' connectors such as joining plates, dowels, or plugs made from wood, plastic, or metal have evolved for assembling particleboard furniture components (Schmidt 1986). However in North America, screws and pre-drilled pilot holes are still the most common method of connecting particleboard furniture components.

In contrast to the United States, there is no publicly available comparative study of the variability in physical and mechanical properties of furniture-grade particleboards manufactured in Canada, which is the source of most raw particleboard used in the manufacture of ready-to-assemble (RTA) furniture in Canada. This study attempts to address this lack of information by sampling 5/8-inch M2 grade particleboards and measuring their strength and screwholding properties. The study compares panel properties across press lines and provides information on the within and between panel variation of properties, including face and edge screw withdrawal resistance (SWR).

The objectives of this study were to:

1. compare the density, IB, face and edge SWR, and flexural strength and stiffness of furniture grade M2 particle-boards produced on six press lines across Canada and

2. determine the effects of screw thread pitch on SWR and of particle orientation (with respect to the machine direction of mat) on SWR and flexural properties.

Materials and methods

Sourcing particleboards and sample preparation

A letter of request for five 15.87-mm- (5/8-in) thick furniture grade M2 particleboards was sent to companies known to produce M2 grade particleboard in Canada. Five out of nine mills contacted agreed to participate in the study, with these mills covering a wide geographical area. The five mills thereafter referred to as press lines) represented four different companies and were randomly assigned a code A to F, with one mill producing panels from two different press lines. Of the six press lines from which panels were sampled, four were batch presses and two were continuous presses.

Five replicate sheets of 4-by 8-foot particleboard wrapped in plastic were received from each mill, for a total of 30 panels. Each of these was cut into eight sub-panels each measuring 61 by 61 cm (2 by 2 ft) using a vertical panel ripping saw. Each sub-panel was labelled with the press line code (A to F), panel replicate (1 to 5), sub-panel position number (1 to 8), and machine direction of the mat (parallel to, [parallel], or perpendicular to, [perpendicular to]). These sub-panels were stacked between stickers in a conditioning room at 20 [+ or -] 1[degrees]C and 65 [+ or -] 5 percent relative humidity (RH) for a minimum of two weeks prior to cutting test samples in accordance with ASTM D 1037 (2000) testing procedures. They were then cut into test samples as shown schematically in Figure 1 and listed in Table 1; the numbers in the specimen ID column of the table correspond with the test specimen codes on Figure 1. The two samples for screws A and B for edge SWR (machine direction), edge SWR (perpendicular direction), and face SWR were side-matched pairs but locations of the pairs and those of all of the other test pieces were randomized within each sub-panel to avoid any location bias. For those press lines with batch presses, the machine direction corresponded to the long edge of the panel. For the two press lines with continuous presses, the machine direction corresponded to the short edge of the panel; therefore the overlaid specimen location template for each sub-panel was rotated 90[degrees] clockwise. Test samples were conditioned again after cutting at 20 [+ or -] 1[degrees]C and 65 [+ or -] 5 percent RH for an additional two weeks prior to testing of mechanical properties. Density (ovendry), moisture content (MC), and strength properties were tested in accordance with ASTM D 1037 (2000).

[FIGURE 1 OMITTED]

Experimental design and statistical analysis

The experimental design consisted of five full-sized (1.22 by 2.44 m or 4 by 8 ft) particleboard panels from each of six 'press lines'. Due to lack of control over which mills were sampled, 'press line' as a main effect is considered a random factor. Each panel (replicate) contained eight test samples (one per sub-panel) for properties 1 to 10 listed in Table 1. In the case of density and IB, properties 11 and 12, respectively, there were two samples per sub-panel, giving 16 samples per panel. The treatment structures (fixed and random factors) for each mechanical property are summarized in Table 2.

Each treatment structure involved a nested design with one or more fixed factors depending on the response variable tested. In the case of density, IB, and face SWR, data was modeled as a split-plot design with panel replicate as the treatment block, i.e., specimens (8 for face SWR, 16 for density and IB) within each of eight sub-panels within each of five replicate panels for each press line. Edge SWR and modulus of rupture (MOR)/modulus of elasticity (MOE) had the extra effect of machine direction and were modeled using a split-split plot design, i.e., eight specimens each for both machine directions for each of eight sub-panels within each of the five replicate panels.

The sample testing was blocked by replicate (panel), i.e., for each mechanical property the samples from all of the press lines for replicate 1 were tested in random order followed by all of those for replicate 2, and so on. The order in which the sub-panels were tested was also randomized for each panel replicate and press line. The testing of mechanical properties took place over a period of approximately three months.

The main effects of, and interactions between, each set of factors for a particular property were tested for significance by an appropriate analysis of variance (ANOVA) model at the 5 percent significance level using Genstat 5 release 4.21 (Lawes Agricultural Trust 2001). Before the final analyses, diagnostic checks on the normality of the data sets and equality of variances were undertaken. Significant results are compared graphically and the Least Significant Difference (LSD) for p [less than or equal to] 0.05 is included on graphs to facilitate comparison of means.

Measurement of panel density

Each IB sample (number 11 on Fig. 1) had a matching ovendry density/MC sample (number 12) cut adjacent to it. The ovendry density of these samples was determined using the water displacement method using the ovendry weight and the undried specimen volume (at 10% equilibrium moisture content [EMC]). The minimum core density was determined from the vertical density profiles measured for each IB specimen prior to IB testing. An X-ray density profilometer (Quintex Measurement Systems; Model QDP-01X) was used to measure the density of IB specimens at intervals of 0.1 mm through the thickness of the specimen. A cross-sectional U-shaped density profile was compiled for each IB specimen. The minimum core density was taken as the minimum density of the middle 6-mm portion of each specimen.

Measurement of SWR

Screw A was a 1-inch No. 10 Type AB sheet metal screw with 16 threads per inch (tpi) specified for screw pull tests in ASTM D 1037 (2000), and screw B was a 1-inch No. 10 T-304 (18-8) sheet metal screw with 10 tpi. Screw B was selected since it was of the same length and diameter as screw A, but had the fewest threads per inch for this particular screw diameter. The exact screw specified for use in ASTM D 1037 (screw A) was difficult to obtain from fastener suppliers, and a closely matching screw was chosen. Root diameter, [D.sub.r], was 3.18 mm for screw A and 3.16 mm for screw B, and total shank diameter, [D.sub.t], including the threads was 4.73 mm for screw A and 4.86 mm for screw B. The thread height, [D.sub.t] - [D.sub.r], for screw A was 1.55 mm and 1.70 mm for screw B.

Pilot holes measuring 3.2 mm (just over 1/8 in) in diameter (specified by ASTM D 1037) (2000) were first drilled using a drill press. The pilot hole diameter was 68 percent of the total diameter of screw A and 66 percent of the total diameter of screw B. For face SWR the hole was drilled through the middle of the test piece. In the case of edge SWR samples, the hole was drilled to a distance of 24 mm mid-way through the specimen cross section and parallel to the surface. The screw was then inserted into the pilot hole and screwed into the sample using an electric hand-held drill until only the top 8 mm of screw shank protruded beyond the edge of the test specimen. Embedment depth for the screws was 17 mm, and measurement of SWR in N was undertaken according to ASTM D 1037 (2000).

Results and discussion

Summary of effects

The main effects of press line, machine direction, and screw type on particleboard physical and mechanical properties are summarized in Table 3. The effect of press line on panel density and strength properties was highly significant (p < 0.001). Face SWR was affected by screw type, whereas edge SWR was not. Interestingly, the machine direction of the mat had no discernible effect on edge SWR; however, it did influence flexural properties (MOR and MOE) as expected. There were no significant interactions between any of the main effects influencing properties, and so interaction terms are not shown in Table 3. Means and coefficients of variation (COV, %) for each measured property for the press lines are given in Table 4.

Physical properties: Density, core density, MC, and IB

The variation in panel density and IB by press line is shown in Figure 2a and b. Mean density of panels from each press line are shown in Figure 2a, and ranged from 650 kg/[m.sup.3] for press line E to 710 kg/[m.sup.3] for press line B. The MC of samples ranged from 9.83 percent in press line D (COV = 2.6%) to 10.46 percent in press lines B and C (COV = 2.1%). Boards from press line A had the highest IB (averaging 0.73 MPa) and those from B the lowest (averaging 0.42 MPa) (Fig. 2b). IB of panels from press lines C to F were all similar (0.58 to 0.65 MPa). Overall, the IB of panels from all press lines met the voluntary ANSI A208.1 (1999) standard for M2 grade particle-board of 0.45 MPa, except those from press line B. This was despite the high density of samples from this press line, as can be seen from Figure 2a. Note from Table 4 the low variation within press lines for panel density. COV is mostly between 2 percent and 5 percent, but is higher in the case of IB; mostly above 9 percent. Variability in IB was also highest for press line B (COV = 18.9%).

The large variation in the IB of panels from different press lines occurred despite the fact that they were of similar density, especially core density. Core density had a narrow range of 530 to 545 kg/[m.sup.3], except for press line D which had an average minimum core density of 500 kg/[m.sup.3]. These results suggest that the low IB of panels from press line B is not caused by insufficient compaction ratio, which is a common cause of low IB, but by other possible unknown factors such as variation in resin content between press lines, inadequate curing, and/or non-homogeneous distribution of resin throughout the core furnish. According to Maloney (1993) process-related factors such as inappropriate pressing schedule or insufficient cooling of urea-for-maldehyde-(UF) bonded panels can also adversely affect IB.

[FIGURE 2 OMITTED]

Variation in IB in boards from the same manufacturer could reflect underlying variability in resin distribution within the furnish material used in the core of the panels, although this is speculative since resin content and formulation in the core furnish is not known. Another factor that could partially explain the variation in bond strength both within and between manufacturers is that environmental regulations of formaldehyde emissions have resulted in the widespread use of less efficient low mole ratio UF resins, making bond quality much more sensitive to physical and chemical variations in the wood feed stock (Gylseth and Maylor 1996, Graves 1999). Variation in particle physical and chemical characteristics arising from different wood species either within or across particleboard plants is a critical source of variation in bond strength and panel strength properties (Kelly 1977, Xu and Suchsland 1998). Inadequate monitoring and control of the species/wood density mix entering a particleboard plant in the form of chips or secondary processing wastes is a also major cause of unacceptably high property variation within panels from the same plant (Cassens et al. 1994, Sjoblom et al. 2004). Wood species used in the particleboards sampled from across Canada in this study ranged from pure spruce to mixtures of hardwoods (birch and maple) and/or softwoods (spruce and fir).

[FIGURE 3 OMITTED]

Face SWR

Since face SWR was significantly affected by screw type (p < 0.001), the mean face SWR values are plotted as a function of press line and screw type in Figure 3. Similar to the trend for IB, face SWR was highest for panels from press line A and lowest for press line B. Values for press lines C to F were similar, ranging from 950 to 1050 N. Mean values for screw B (10 tpi) were, without exception, higher than those for screw A (16 tpi). The COV within press lines for face SWR with screw A ranged from 7 to about 12 percent, while the COV for screw B was slightly lower, ranging from 6.6 to 11.2 percent.

The ANSI A208.1 (1999) minimum requirement for face SWR of M2 particleboard is 1000 N. This is for tests made according to ASTM D 1037 (2000) using the No. 10 Type AB 1-inch screw with a pitch of 16 tpi (i.e., screw A used in our study). All press lines except for B and F met the standard with screw A. Boards from press lines A, C, D, and E had average face SWR over 1000 N with screw B. Face SWR was significantly lower for screw A in the cases of press lines A, B, D, and F; whereby the difference was greater than the LSD of 41 N. The difference between screws A and B was less than 41 N in the cases of press lines C and E. Differences in particle size, geometry, and bonding in particleboards from different press lines is likely to have influenced screw holding behavior of screws with different thread configurations.

The findings of this study are in agreement with those of a study by Superfesky (1974) who found that screws with fewer threads i.e., 12 tpi as opposed to 16 tpi, had slightly but consistently higher face SWR While the embedment depth and screw dimensions in our study were similar to those used by Superfesky (1974), average face SWR loads in his study were higher (1375 N for the 12 tpi screws and 1299 N for the 16 tpi screws).

Our values for face SWR correspond better with predicted values from a model based on specific gravity (SG) by Eckelman (1975) for face SWR from softwood particleboard. In that model, the force required to pull a screw for the particleboard face, [F.sub.face], is given by:

[F.sub.face] = 2655[D.sup.0.5](L - [D/3])[.sup.1.25] [G.sup.2] (lb) [1]

where:

D = the shank diameter,

L = the embedment depth in inches, and

G = the SG of the sample at a MC of 10 percent.

The grade of particleboard for which this model was derived was not specified.

For example, for a panel with a SG of 0.6 and MC of 10 percent at time of testing, Eckelman's model predicts [F.sub.face] values of 979 N for Type A (16 tpi) screws and 988 N for type B (10 tpi) screws. The predicted difference between the two screw types is based solely on the difference in shank diameter, and is minimal at 9 N or 1 percent, whereas average observed face SWR for screw B was 3.6 percent higher than screw A. This suggests that much of the difference between the screws in their effect on face SWR may be explained by factors other than shank diameter, such as thread pitch and thread height.

Edge SWR

Since there was no significant effect of machine direction or screw type on the edge SWR from sampled particleboards; the results for edge SWR were pooled across machine direction and screw type and plotted as a function of press line only in Figure 4. Edge SWR values were approximately 25 percent lower than face SWR, in agreement with the findings of Eckelman (2003).

The ANSI A208.1 (1999) minimum requirement for SWR (edge) is 900 N. SWR from the edge of particleboard is also more variable than from the face due to the lower density of the core and greater structural heterogeneity from the presence of coarser particles in this layer (Superfesky 1974). Only panels from press line A exceeded the standard value (mean edge SWR = 973 N) while average edge SWR in panels from other press lines were significantly lower; between 630 N (press line B) and 790 N (press line F). Note from Table 4 the higher variability within press lines for edge SWR (COV values ranging from 11.0% to 17.4% in the case of press line B).

The high edge SWR of panels from press line A corresponds with the high IB strength of its core, from Figure 2b. Other studies suggest that edge SWR is likely to be a function of both IB and density (Johnson 1967; Eckelman 1973, 1975; Rajak and Eckelman 1993; Wong et al. 1999). Relationships between panel density, IB, and SWR from the data collected in this study will be investigated in a subsequent publication.

Data from the National Particleboard Association (1968). Eckelman (1975), and Barnes and Lyon (1978) suggest that edge SWR in particleboard can be reliably predicted using only SG, screw thickness, and depth of penetration. Eckelman (1975) gives a predictive model for the force required to pull a screw from the edge of a particleboard sample, [F.sub.edge], based on its SG as:

[FIGURE 4 OMITTED]

[F.sub.edge] = [2055.sup.0.5](L - [D/3])[.sup.1.25] [G.sup.2] (lb) [2]

where:

D = the shank diameter.

L = the embedment depth in inches, and

G = the SG of the sample at a MC of 10 percent.

Predicted [F.sub.edge] values for panels of similar SG to ours, i.e., 0.6, are 757 N for screw type A and and 765 N for screw type B. These values are below the 973 N for press line A and above the 630 N for press line B, but agree reasonably well with measured edge SWR of the other press lines (C to F).

By comparison, edge SWR values in particleboard measured by Superfesky (1974) were higher: 1107 N for 10 tpi screws (Screw A) and 1119 N for 12 tpi screws. The grade of particleboard used by Superfesky was unspecified, but that work may have used a higher grade of particleboard than the panels used in this study. In agreement with our findings, Superfesky (1974) found no consistent difference between the two screw types for edge SWR, a finding that was attributed to the possible masking effect of greater heterogeneity of structure and bonding within the core region of particleboard. The screw holding ability of particleboard is lower than that of solid wood, plywood, OSB, or MDF. For example, the SWR of particleboard is around 40 percent of that for hardwood MDF (Eckelman 2003). Results from Erdil et al. (2002) for Douglas-fir plywood of the same thickness, screw type, and embedment depth as used in this study gave edge SWR values of approximately 1802 N. Values for No. 10 AB screws in the edge of 19-mm-thick OSB were on average 1686 N (Erdil et al. 2002), although they used an embedment depth of 25.4 mm, compared with only 17 mm used here.

Flexural strength (MOR and MOE)

As expected, machine direction had a significant effect (p < 0.001) on MOR and MOE (Table 3). The means for each press line and machine direction are compared in Figure 5a for MOR and Figure 5b for MOE. The ANSI A208.1 standard minimum values for MOE and MOR of M2 grade particle-board are 2250 MPa and 14.5 MPa, respectively. Most panels exceeded the requirement for MOE, but not MOR (except for those from press lines A and D). The variability in MOR and MOE differed by press line (Table 4). Variation was smallest for press line A. COV ranged from 2.5 to 6.8 percent for MOR and MOE, whereas for press line B the COV for the same properties ranged from 11.8 to 16.7 percent. This higher variability in flexural strength for press line B may be due to the greater within press line variation of density and IB observed for press line B.

[FIGURE 5 OMITTED]

It was observed that the mean MOR of panels from press line B was not very different to other press lines despite their lower IB. It is speculated that a faster press closure rate may have been used in the manufacture of the panels sampled from press line B. It is well known that faster press closure rate improves bending strength but is at the expense of IB strength and screw holding ability (Bismarck 1974, Rice et al. 1967). This is because press closure rate, in combination with surface and core furnish MC, determines the shape of the vertical density profile (Kelly 1977, Wang and Winistorfer 2000, Dai and Wang 2004). Vertical density profiles will be presented and discussed in a subsequent paper in which MS and M2 grade particleboards from two press lines are compared.

Conclusions

M2 furniture grade particleboards produced on six different press lines varied significantly in physical and mechanical properties; the main findings from the survey were:

1. All press lines except for one exceeded ANSI 208.1 (1999) requirements for IB of 0.45 MPa or greater.

2. The face SWR of most press lines was just equal to or below the ANSI minimum of 1000 N, except for one which exceeded this value by 10 percent. Screw type (10 or 16 tpi) had a significant effect on withdrawal resistance. SWR was inversely proportional to thread pitch with the 10 tpi screws having higher face SWR than those with 16 tpi.

3. Five of the six press lines failed to meet the ANSI minimum of 900 N for edge SWR. Edge SWR was not affected by thread pitch.

4. Only two press lines exceeded the ANSI minimum of 14.5 MPa for MOR, but five of the six press lines exceeded the minimum of 2.25 GPa for MOE.

5. Panels from one press line were consistently lower in all strength properties despite having high average density. The variation in properties for this press line was also higher than for other press lines.

Literature cited

American National Standards Institute (ANSI). 1999. ANSI A208.1 Particleboard, 11 pp.

American Society for Testing and Materials (ASTM). 2000. Annual book of ASTM standards. Section 4. ASTM D-1037 Construction (Wood). Vol. 04.01.

Barnes, H.M. and D.E. Lyon. 1978. Fastener withdrawal loads for weathered and unweathered particleboard decking. Forest Prod. J. 28(4):33-36.

Bautista, C.Q. 1994. Particleboard industry properties survey. Part I. Survey results. In: Proc. of the 48th Forest Prod. Soc. Annual Meeting, Portland, ME. Panel Products Technology Centre. Temple-Inland Forest Prod. Corp., Diboll. TX. 22 pp.

Bismarck, C. 1974. Optimizing the pressing of particleboards. The manufacture of particleboards with urea-formaldehyde binders using special automated regulation systems for the pressing process. Holz-Zentralblatt. 100(80):1247-1249.

Carroll, M.N. 1970. Relationship between driving torque and screw holding strength in particleboard and plywood. Forest Prod. J. 20(3):24-29.

Cassens. D.L., J.P. Bradtmeuller, and F. Picado. 1994. Variation in selected properties of industrial grade particleboard. Forest Prod. J. 44(10):50-56.

Eckelman, C.A. 1973. Holding strength of screws in wood and wood based materials. Agr. Res. Bulletin 85. Purdue Univ. 15 pp.

__________. 1975. Screwholding performance in hardwoods and particleboard. Forest Prod. J. 25(6):30-35.

__________. 2003. Chapter 6: Strength of screws in wood composites. In: Textbook of Product Engineering and Strength Design of Furniture. Purdue Univ., School of Agriculture. pp. 56-67.

Erdil, Y.Z., J. Zhang, and C.A. Eckelman. 2002. Holding strength of screws in plywood and oriented strand board. Forest Prod. J. 52(6):55-62.

Graves, G. 1999. Urea formaldehyde resins: Yesterday, today and tomorrow. In: Proc. of the 1998 Resin & Blending Seminar, 29-30 Oct., Portland OR. J. Bradfield, Ed. Composite Panel Assoc., Gaithersberg, MD. pp. 3-10.

Gylseth, B. and R. Maylor. 1996. New melamine modified binders for low formaldehyde emission wood panels. Asian Timber, 9:52-55.

Johnson, J.W. 1967. Screw-holding ability of particleboard and plywood. Forest Research Lab. Rept. No. T-22, School of Forestry, Oregon State Univ., Corvallis, OR.

Kelly, M. W. 1977. Critical literature review of relationships between processing parameters and physical properties of particleboard. Gen. Tech. Rep. FPL 10. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 64 pp.

Lawes Agricultural Trust. 2001. Genstat--A General Statistical Program 5 Release 4.21. Lawes Rothhamsted Experimental Station, UK.

Maloney, T.M. 1993. Modern Particleboard and Dry Process Fiberboard Manufacturing. 2nd ed. Forest Prod. Soc., Madison, WI. 681 pp.

National Particleboard Association (NPA). 1968. Screw holding of particleboard. Tech. Bulletin No. 3, NPA, Washington DC.

Rajak, Z.I. and C.A. Eckelman. 1993. Edge and face withdrawal of large screws in particleboard and medium density fiberboard. Forest Prod. J. 43(4):25-30.

Rice, J.T., J.L. Snyder, and C.A. Hart. 1973. Influence of selected resin and bonding factors on flakeboard properties. Forest Prod. J. 17(8):49-57.

Schmidt, H. 1986. Industrial particleboard: Application requirements and special manufacturing features. In: Proc. of the 20th WSU International Particleboard/Composite Materials Symp. T.M. Maloney, Ed. Washing State Univ., Pullman, WA, pp. 343-354.

Sjoblom, E., B. Johnson, and H. Sundstrom. 2004. Optimization of particleboard production using NIR spectroscopy and multivariate techniques. Forest Prod. J. 54(6):71-75.

Suchsland, O. and W.S. Good. 1968. The selection of panel materials by furniture and cabinet manufacturers. Michigan Agr. Expt. Sta. J. Article No. 4378. Michigan State Univ., East Lansing, MI. 11 pp.

Superfesky, M.J. 1974. Screw withdrawal resistance of types A and AB sheet metal (Tapping) screws in particleboard and medium density fiberboard. Res. Paper No. 239. USDA Forest Serv., Forest Prod. Lab., Madison, WI. 8 pp.

Wong, E.D., M. Zhang, Q. Wang, and S. Kawai. 1999. Formation of the density profile and its effects on the properties of particleboard. Wood Sci. Tech. 33(4):327-340.

Wu, Q.L. and R.P. Vlosky. 2000. Panel products: A perspective from furniture and cabinet manufacturers in the southern United States. Forest Prod. J. 50(9):45-50.

Xu, W. and O. Suchsland. 1998. Variability of particleboard properties from single and mixed-species processes. Forest Prod. J. 48(9):68-74.

Zhang, J. 1994. Particleboard industry physical properties survey. Part II. Analytical technique. In: Proc. of the 48th Forest Prod. Soc. Annual Meeting, Portland, ME. Panel Products Technology Center, Temple-Inland Forest Prod. Corp., Diboll, TX. 27 pp.

Kate Semple*

Emmanuel Sackey*

He Jun Park

Gregory D. Smith*

The authors are, respectively, Post Doctoral Fellow (ksemple@forestry.ubc.ca), PhD Student (esackey@interchange.ubc.ca), Visiting Scientist (phjun@iksan.ac.kr), and Assistant Professor (gregory.smith@ubc.ca), Dept. of Wood Science, The Univ. of British Columbia, Vancouver. BC. Canada. This paper was received for publication in November 2004. Article No. 9955.

*Forest Products Society Member.

Table 1. -- Mechanical properties, specimen ID, and the number of
samples for M2 particleboard.

                                              Number of samples per
                                       Specimen  Press
Property                               ID        line   Board  Sub-panel

Edge SWR ([parallel] to machine         1        40      8     1
  direction) Screw A
Edge SWR ([perpendicular to]            2        40      8     1
  to machine direction) Screw A
Edge SWR ([parallel] to machine         3        40      8     1
  direction) Screw B
Edge SWR ([perpendicular to]            4        40      8     1
  to machine direction) Screw B
Face SWR Screw A                        5        40      8     1
Face SWR Screw B                        6        40      8     1
MOR ([parallel] to machine direction)   7        40      8     1
MOR ([perpendicular to] to              8        40      8     1
  machine direction)
MOE ([parallel] to machine direction)   9        40      8     1
MOE ([perpendicular to] to             10        40      8     1
  machine direction)
IB                                     11        80     16     2
Board ovendry density                  12        80     16     2

Table 2. -- Treatment structures for board physical and mechanical
properties.

Property     Specimen ID  Fixed factors      Levels

Edge SWR      1 to 4      machine direction  [parallel],
                                               [perpendicular to]
                          screw type         A,B
Face SWR      5 and 6     screw type         A,B
MOR, MOE      7 to 10     machine direction  [parallel],
                                               [perpendicular to]
IB, density  11 and 12    none               --

Property     Random factors  Levels

Edge SWR     press line      A to F
             panel           1 to 5
             sub panel       1 to 8
             specimen        1 to 4
Face SWR     press line      I to F
             panel           1 to 5
             sub panel       1 to 8
             specimen        1,2
MOR, MOE     press line      A to F
             panel           1 to 5
             sub panel       1 to 8
             specimen        1,2
IB, density  press line      A to F
             panel           1 to 5
             sub panel       1 to 8
             specimen        1,2

Table 3. -- Main effects on properties of M2 particleboards from six
press lines. (a)

Effect                 Density    IB         Face SWR   Edge SWR

Press line (P)         p < 0.001  p < 0.001  p < 0.001  p < 0.001
Machine direction (D)  n.a.       n.a.       n.a.       n.s.
Screw type (S)         n.a.       n.a.       p < 0.001  n.s.

Effect                 MOR          MOE

Press line (P)         p < 0.001    p < 0.001
Machine direction (D)  p = 0.0004   p < 0.001
Screw type (S)         n.a.         n.a.

(a) n.a. = not applicable for that mechanical property; n.s. = not
significant at the 5% confidence level; p < 0.001 = significant at the
0.1% level.

Table 4. -- Means and COV (%) of physical and mechanical properties for
boards from press lines A to F.

                                                Press line
Property                                     A        B       C

Density                 Mean (kg/[m.sup.3])   681.3   706.9    702.0
                        COV (%)                 1.6     4.9      3.0
IB                      Mean (MPa)              0.71    0.43     0.65
                        COV (%)                11.2    18.9      7.6
Face SWR                Mean (N)             1098.2   837.2   1020.4
  Screw A               COV (%)                12.3    11.5      8.8
Face SWR                Mean (N)             1166.1   883.1   1031.0
  Screw B               COV (%)                 6.6    11.2      7.0
Edge SWR                Mean (N)              972.9   634.4    776.2
                        COV (%)                12.4    17.4     11.9
MOR [parallel]          Mean (MPa)             16.0    13.5     14.7
                        COV (%)                 5.4    16.7     14.7
MOR [perpendicular to]  Mean (MPa)             15.0    12.6     14.2
                        COV (%)                 3.7    14.3      6.6
MOE [parallel]          Mean (GPa)              3.1     2.3      2.6
                        COV (%)                 6.8    15.8      9.2
MOE [perpendicular to]  Mean (GPa)              2.8     2.2      2.4
                        COV (%)                 2.5    11.8      5.5

                                                Press line
Property                                     D        E        F

Density                 Mean (kg/[m.sup.3])   657.6    646.7   648.3
                        COV (%)                 3.7      3.1     3.0
IB                      Mean (MPa)              0.59     0.58    0.61
                        COV (%)                11.1      9.1     9.0
Face SWR                Mean (N)             1035.2   1015.6   950.0
  Screw A               COV (%)                10.8      7.0    10.6
Face SWR                Mean (N)             1076.4   1026.3   991.5
  Screw B               COV (%)                 9.6      6.7     8.4
Edge SWR                Mean (N)              732.5    770.8   790.1
                        COV (%)                14.0     10.9    11.8
MOR [parallel]          Mean (MPa)             16.6     12.3    13.4
                        COV (%)                12.6      6.3    12.6
MOR [perpendicular to]  Mean (MPa)             15.3     12.2    13.0
                        COV (%)                13.7     11.3    13.8
MOE [parallel]          Mean (GPa)              3.0      2.6     2.8
                        COV (%)                 9.3      8.6    12.1
MOE [perpendicular to]  Mean (GPa)              2.8      2.4     2.7
                        COV (%)                10.2      8.9    10.9