Copyright Forest Products Society Feb 1993

The main advantage of laser cutting is the ability to produce narrow kerf widths, which result in less waste (7, 16). Lasers can cut rectangular or curved parts, and blind cuts are possible. There are no mechanical stresses placed on the workpiece since material is removed by thermochemical means. Other advantages of laser cutting wood are low noise level, low power per lineal foot cut, no sawdust, and easy automation because the process is computer controlled (7,9,12,13). However, the disadvantage of laser cutting wood is the limited depth-of-cut, the carbonized charred surface, and the exposure of wood to elevated temperatures (3,4,6,9,13,14,17).

The charred surface left by laser cutting is caused by thermochemical decomposition (TCD) of wood (1,3,4,11,14,19), which has significant effects on glueline performance. Previous work indicated a loss of bond strength without partial or complete removal of the charred layer (19). The optimization of laser processing parameters for wood may lessen the problems of TCD (2).

The purpose of this study was to investigate the laser cutting of wood materials and how it affects the gluability of wood. This study seeks an alternative method to rough-cut wood, which will reduce the waste found in conventional rough-cutting and defect manipulation. The information gathered will be important in the development and understanding of how laser-cutting parameters affect profiles, waviness, flatness, perpendicularity, discoloration, and average surface roughness (ASR) of the laser-cut kerf (1,5,6,8, 15,19) and as these factors affect wood gluabillty.

STUDY OBJECTIVES

1. Study the characteristics of laser-cut wood surfaces and determine the effect of laser cutting upon adhesion by testing glued samples for shear strength and wood failure.

2. Determine the most economical method of improving the laser-cut wood surface for gluing, including methods of handling.

3. Determine the characteristics of glues that would be best suited for laser-cut wood surfaces.

EXPERIMENTAL PROCEDURE

MATERIAL

Sugar maple, medium density fiberboard (MDF), and ponderosa pine were selected for comparison in the laser-cutting experiment. The selection of samples was done on the basis of specific gravity (SG), straightness of grain, and freedom from defects. Samples represented, as nearly as possible, ideal quartersawn (radial sections) and flatsawn (tangential sections) pieces. SG of the samples was identified for both species. MDF samples were selected to eliminate the SG variation that is a typical phenomenon within the annual rings of wood. Prior to the experiment, the samples were brought to equillbrium moisture content (MC) at 40 percent relative humidity and a temperature of 22degC. Mean MC was measured with a Gann Electronics H#35 moisture meter.

PREPARATION OF SAMPLES FOR LASER CUTTING

Solid wood samples were cut to provide tangential and radial sections for bonding. Test materials were selected for grain configuration from strips 810 mm (32 in.) long. Strips were planed to thickness and width dimensions of 20 by 42 mm. Samples 381 mm long were cut from the strips using a crosscut table saw. The MDF test material was cut from 19-mm-thick panels, cut to a width of 102 mm, and a length of 381 mm. All samples were then weighed within +/- 0.0 1 gram on a Mettler electronic balance, and the dimensions were measured to within +/- 0.1 mm. Using these data, the SG of each sample was calculated.

LASER CUTTING PROCESSING PARAMETERS

In order to determine how laser cutting affects adhesion, samples of sugar maple were laser cut from the tangential and radial sections. The experiment used a Trumpt continuous wave carbon dioxide (CO sub 2 ) laser rated at a mmum power of 2.8 kW, with reflective focusing optics. Settings for power, focal length, speed of cut, and cover gas pressure were adjusted for apparent optimum surface conditions based on previous experiments of Khan and Mukherlee (10).

PARAMETERS OF LASER CUT I

Focusing optics: 254 mm focal length

Air pressure: 0.58 MPa

Power setting: 1.850 watts

Frequency: 10 kHz

Focal point: adjusted to 3 mm beneath the surface of the samples

Feed rate: 254 cm/minute for cutting sugar maple with SG varying from 0.60 to 0.71 g/cm sup 3 . For samples with a higher SG (0.72 to 0.84). the feed rate had to be decreased to 238 cm/minute.

PARAMETERS OF LASER CUT II

Focusing optics: 127 mm focal length

Air pressure: 0.35 MPa

Power setting: 750 watts

Frequency: same as laser cut I

Focal point: same as laser cut I

Feed rate: 90 cm/minute

PARAMETERS OF LASER CUT III

Focusing optics: 254 mm focal length

Air pressure: 0.35 MPa

Power setting: 1,750 watts

Frequency: same as laser cut I

Focal point: same as laser cut I

Feed rate: 180 cm/minute

PARAMETERS OF LASER CUT IV

Focusing optics: 254 mm focal length

Air pressure: 0.58 MPa

Power setting: 2,600 watts

Frequency: same as laser cut I

Focal point: same as laser cut I

Feed rate: 230 cm/minute for cutting MDF with SG 0.74 g/cm sup 3 ; 305 cm per minute for cutting sugar maple with radial and tangential sections with SG varying from 0.67 to 0.81 g/cm sup 3 ; 380 cm per minute for cutting ponderosa pine with tangential sections with SG varying from 0.48 to 0.51 g/cm sup 3 .

PREPARATION OF SAMPLES FOR GLUELINE SHEAR STRENGTH TESTS

Samples cut along the grain were used to compare glueline shear strength and wood failure. The materials were prepared by conventional table sawing and planing. Comparative samples were laser cut. Sample preparation for glueline shear strength tests followed the recommendations in ASTM D-905-89 except that samples were cut to a 20-mm width for sugar maple and ponderosa pine, and a 19-mm width for MDF instead of the standard 51 mm. These adjustments were necessary because the laser beam provides a limited depth-of-cut.

The adhesives used in the experiments were:

1. A phenol-resorcinol (PR) liquid resin (100 parts) mixed with a powder catalyst (20 parts). Viscosity at 25degC was 200 +/- 50 cp with a pH of 7.70 +/- 0.20, and a solids content of 54 +/- 1 percent. This adhesive has a long open time (5 min.) and long closed time (15 min.). The extended times should allow the glue to wet the substrate and penetrate it. This was an important factor in the selection of this adhesive because laser-cut samples are difficult to wet during the application of adhesive (1,3,11,19).

2. A pre-catalyzed crosslinking polyvinyl acetate (PVAc) emulsion adhesive that is recommended for hot and cold press applications. The performance of this adhesive on laser-cut samples was expected to be less satisfactory than PR since the cure rate is faster. Viscosity at 28degC was 3,500 +/- 500 cps, with a pH of 3 +/- 0.50, and a solids content of 48 +/- 2 percent.

SURFACE PREPARATIONS

The following approaches were used on samples prior to gluing the laser-cut surface:

1. Conventional surface preparation (sawing, jointing, or planing) and wiped-of with tissues.

2. Laser-cut surfaces with 1-mm carbonized layer planed off along the grain. The laser-cut surface became white and flat.

3. Laser-cut surface with loose particles brushed off.

The same approaches were used for MDF sample preparation. Grain orientation, SG, and MC of the wood samples were closely monitored. Bonding conditions for the parameters of laser cut I, II, and III were as follows:

PVAc adhesive:

Spread rate of glue (single side spread): 250 g/m sup 2

(The amount of glue when using laser cut parameters was 500 g/m sup 2 , glue was applied to each side of the parts to be bonded (double side spread); this enabled the glue to penetrate the substrates.)

Open assembly time: 30 seconds

Closed assembly time: 90 seconds

Cure temperature: 22degC

Clamping pressure: 1.4 MPa

Clamping time: 2 hours

Post-gluing conditioning: 7 days

PR adhesive:

Spread rate of glue (double side spread): 500 g/m sup 2

(The glue was applied on each side of the parts to be bonded.)

Open assembly time: 5 minutes

Closed assembly time: 15 minutes

Cure temperature: 22degC

Clamping pressure: 1.4 MPa

Clamping time: 8 hours

Post-gluing conditioning: 7 days

RESULTS AND DISCUSSION

Visual and olfactory surface examination of the samples before gluing revealed interesting information about the laser cutting effects. In all cases, samples had the characteristic odor of carbonized wood. These characteristics are more pronounced and clearly shown in laser-cut red oak. Figure 1 shows the tangential section of a laser-cut red oak sample. (Figure 1 omitted) Note the waviness and differences in the roughness between spring and summerwood. Figure 2 shows the radial section of laser-cut red oak, note the washboard effect produced. (Figure 2 omitted) Figure 3 shows a laser cut made across the end grain of the radial section of oak. (Figure 3 omitted) The kerf was opened and shows that larger volumes of wood were removed from the less dense spningwood than from the denser summerwood. In addition, the bottom of the kerf shows where the laser beam (power density, feed rate, and air pressure held constant) penetrated the springwood thickness, but left the summerwood partially intact. This is evidence of the fact that laser cutting is a mass removal process, since denser materials were removed at a slower rate (8,9, 14,19).

A Bendix Surface Profilometer was used to determine the ASR (mu inch). A Sigma Scan Measurement System was then used to digitize these data. The ASR data are presented in Figure 4, which shows that the planed surfaces had the lowest ASR. However, the planed tangential surface is significantly (in this paper significant refers to statistical significance at the .95 confidence level) smoother than the planed radial surface for sugar maple. The planed tangential surface of sugar maple is also smoother than the planed tangential surface of ponderosa pine, and the as-received surface of MDF. The ASR of the top edge of the laser-cut kerfs, measured along the grain, was significantly larger than planed surfaces for each tested material. ASR increased with the depth-of-cut. Consequently, the ASR at the bottom of the kerf was extremely high; all measurements exceeded the range (+/- 8000 mu inch) of the profilometer.

SURFACE CHARACTERISTICS OF LASER-CUT WOOD

The surface quality of laser-cut wood depends on the diameter of reflective optics. The 127-mm-diameter optics produce a hollowing near the center of the kerf. The 254-mm-diameter optics produce a V-shaped kerf shown in Figures 5 and 6. (Figure 5 and 6 omitted) The average "V" angles produced by using the 254-mm-diameter optics are shown in Table 1. (Table 1 omitted) ASR and waviness are demonstrated in Figures 1 through 4. As previously reported (8,14,19), ASR and waviness vary with the distribution of SG in the annual growth rings.

Figure 7 shows an open kerf with dried glue and a TCD layer on a hollowed laser-cut radial surface. At the center of the kerf, voids in the glueline are obvious. (Figure 7 omitted) Figure 8 shows a thick PR glueline on the laser-cut tangential surfaces of sugar maple (A) and a thin PR glueline on the planed tangential surfaces of sugar maple (B). (Figure 8 omitted)

Table 2 shows the width of the laser beam kerf measured on radial and tangential cut samples of sugar maple. (Table 2 omitted) The measurements were taken from the top and bottom, at the beginning, the middle, and the end of the kerf. The width of the kerf was not significantly different between the tangential and radial samples. However, it is clear from Table 2 that the kerf is significantly wider at the top of the samples than at the bottom. These variations were found in both radial and tangential samples at each of the three measurement points. The width variations of the kerfs are consistent with the findings presented in Table 1.

SUGAR MAPLE GLUELINE STRENGTH FOR PARAMETERS OF LASER CUT I

Table 3 shows the shear strength of the PR glueline by type of surface for sugar maple samples. (Table 3 omitted) The data for glueline shear show no significant differences. No wood failure was produced at the glueline during testing with the one exception of the planed radial sections. This is logical since the radial section is the weakest of the three orthotropic wood sections (18,20). In order to improve understanding of PR gluing of laser-cut surfaces, parameters of laser cut IV are repeated on sugar maple.

MDF GLUELINE STRENGTH FOR PARAMETERS OF LASER CUTS II THROUGH IV

MDF glueline shear strength test results for samples prepared with laser-cut parameters II and III and glued with PVAc adhesive are presented in Figure 9. (Figure 9 omitted) The conventional sawing (sample #1) produced 11.02 MPa glueline shear strength, while the laser cut II samples with the TCD layer removed by planing (sample #2) produced glueline shear strength of 10.69 MPa. The difference between these two samples was not significant. Laser-cut samples with the carbonized layer brushed off (sample #3) produced 6.48 MPa glueline shear strength.

MDF samples prepared with the cut III parameter and glued with the laser cut resembling a "V" (sample #4) produced a glueline shear strength of 6.61 MPa. These samples had one of the glued pair rotated 180 degrees to improve surface contact (sample #5). This matching produced glueline shear strength of 7.71 MPa. Inspection showed that the contact area was improved, but still the center of the glueline on the MDF samples failed due to a thicker glueline. This occurred despite the fact that the thickness of glueline and the flatness of matched paired surfaces was improved when compared to cut II. This part of the experiment shows the necessity to improve flatness when laser-cut surfaces are used for gluing applications.

Test results for the MDF glueline shear strength using PVAc and PR adhesives for samples prepared with laser cut IV parameters are presented in Figure 10. (Figure 10 omitted) Data for the PVAc adhesive showed no significant difference among the brushed-off laser-cut surface, laser-cut surface with 1 mm planed off, and planed surface. Data for PR adhesive showed the brushed-off laser-cut surfaces had significantly lower shear strength than the other two surfaces.

It is important to emphasize that the PVAc and PR adhesives were spread on each side of the parts to be joined when using laser cut IV parameters. This allowed the PVAc to penetrate the substrate and develop a strong glueline. The high cure rate and short open and close time were initially thought to be insufficient for this application. PR adhesive produced a glueline shear strength equal to those of PVAc on planed surfaces and the laser-cut surface with 1 mm planed off. The brushed-off laser-cut surfaces were significantly weaker. The glueline shear strength for the brushed-off laser-cut samples using PR adhesive was 7.5 MPa with a standard deviation of +/- 2.29. This was the largest variance of the six sets tested.

Close inspection of Figure 11 shows that the lasercut surfaces were hollowed near the center. (Figure 11 omitted) This phenomenon was due to characteristic decreasing gradient of SG from the surface to the center in MDF.

This reduction of SG caused the laser beam to remove a larger volume of lower density material toward the center of the board (8-10,14,15). Surface contact in the glued areas ranged from 40 to 60 percent. The glueline shear strength on the brushed off laser-cut surface with PR adhesive was 60 percent of planed MDF.

The experimental data show that PVAc adhesive (double side spread) on carbonized laser-cut surfaces was able to develop glueline shear strength equal to the glueline strength produced by the planed surfaces of MDF. However. Figure 11 shows that the laser-cut surface requires improvement in flatness in order to increase the surface-to-surface contact. Good surface-to-surface contact is necessary for uniform glueline thickness and strong gluebonds.

PONDEROSA PINE GLUELINE STRENGTH FOR PARAMETERS OF LASER CUT IV

Shear strength and wood failure for sets of PVAc and PR adhesives on specimens prepared with parameters for laser cut IV are presented in Figure 12. (Figure 12 omitted) The shear strengths of the brushed-off surfaces (samples #1 and #4) did not differ significantly. Wood failure had greater variability than the shear strength. Sets of PVAc (sample #1) and PR (sample #4) adhesive applied to brushed-off laser-cut surfaces resulted in just 8.67 and 6.11 percent wood failure, respectively. The wood failure data indirectly indicate that the TDC layer reduced glue penetration into the substrate (3,11,19). Glue was encapsulated by the carbon in the TDC layer and dried on the surface of the wood. In addition, lack of surface flatness influenced the low wood failure results. PVAc and PR adhesives on planed ponderosa pine surfaces (samples #3 and #6) produced 90 percent and 77 percent wood failure; laser cuts with 1 mm planed off the surface produced wood failure of 44 percent (sample #2) and 81 percent (sample #5), respectively.

SUGAR MAPLE TANGENTIAL SECTION GLUELINE STRENGTH FOR PARAMETERS OF LASER CUT IV

Sugar maple tangential section glueline shear strengths and wood failure for specimens prepared with parameters of laser cut IV are presented in Figure 13. (Figure 13 omitted) The PVAc adhesive shear strengths for planed surfaces (sample #3) and laser-cut surfaces with 1 mm planed off (sample #2) produced 21.4 and 27.0 MPa, respectively. On similarly prepared surfaces (samples #6 and #5), the PR adhesive produced only 16.4 and 10.3 MPa, respectively. These differences are significant.

PVAc and PR adhesives on brushed-off laser-cut surfaces (1 and 4) produced glueline shear strengths of 7.2 MPa for both adhesives. The tangential section of sugar maple produced low wood failure with PVAc adhesive on the planed surfaces (sample #3) and laser-cut surfaces with 1 mm planed off (sample #2). Earlier works (18,20) suggest that the tangential section of sugar maple is significantly stronger in shearing than the radial section. Laser-cut surfaces where the loose particles were brushed off produced no wood failure, which indicates a problem using glues on these surfaces. The TCD layers resulting from laser cutting prevented sufficient glue penetration. The lack of surface flatness from laser cutting prevented proper surface-to-surface contact, which is necessary for maximum bond strength.

SUGAR MAPLE RADIAL SECTION GLUELINE STRENGTH FOR PARAMETERS OF LASER CUT IV

Results of sugar maple radial bonded glueline strengths and wood failure for sets of PVAc and PR adhesives prepared with parameters of laser cut IV are presented in Figure 14. (Figure 14 omitted) Radial bonded sugar maple samples have lower glueline strengths and higher percentage of wood failure than tangential bonded specimens (Fig. 13). The lowest glueline strengths were obtained on brushed off laser-cut surfaces (samples #1 and #4). Wood failure resulting from the use of PVAc adhesive was significantly higher when compared to PR adhesives. The samples glued with PVAc adhesive also produced higher glueline strengths on planed (sample #3) and laser-cut surfaces with 1 mm planed off(sample #2) than the samples glued with PR adhesive (samples #6 and #5). The experimental results for shear strength on laser-cut surfaces with 1 mm planed off suggest that planing off the TCD layer gave flatness and perpendicularity to the kerf surface. These factors provided surface conditions that re stored wood failure and glueline shear strength in sugar maple.

SUMMARY AND CONCLUSIONS

1. The kerf surface on laser-cut woods is covered by a lava-like layer that prevents glue from wetting and penetrating the wood.

2. Laser-cut wood surfaces have the characteristic odor of carbonized wood and discoloration on the entire surface of the kerf. There was a lighter color at the beam entrance and a darker color at the laser beam exit.

3. Woods with lower SG produced a lighter color, while higher SG woods produced a darker color at the laser-cut kerf.

4. Profile and waviness of laser-cut wood is affected by: 1) focusing optics: 127-mm-diameter optics produced hollowing near the center thickness of the specimens, while 254-mm-diameter optics produced a V-shaped kerf with a hollowing near two-thirds of the specimen thickness. The width of a laser kerf was wider at the entrance and narrower at the exit of the laser beam; 2) SG distribution in the annual growth rings (spring/summerwood) or in man-made materials; and 3) the distribution of structural elements in wood: wood rays, middle lamella, springwood, secondary walls, and summerwood.

4. ASR for laser-cut wood measured along the grain at the top edge of the kerf was significantly greater than the ASR for planed wood. ASR of the laser-cut kerf increased with the depth of cut. Extreme ASR was found at the bottom of the laser cuts.

5. Laser processing parameters for MDF and sugar maple specimens indicated that power density of the laser beam, air pressure of cover gas, and feed rate affected the glueline shear strengths developed with either PVAc or PR adhesives.

6. PVAc adhesive (double side spread) on brushed-off laser-cut surfaces was able to develop glueline shear strength equal to the glueline strength produced by the planed surfaces of MDF.

7. On the brushed-off laser-cut surfaces of sugar maple and MDF, a double-side spread of PVAc adhesive significantly improved glueline strength.

8. The glueline shear strength for the laser-cut surface of ponderosa pine showed no significant difference when PVAc or PR adhesive was used. Ponderosa pine glueline shear strength for laser-cut surfaces had no significant differences among glueline shear strengths for brushed-off laser-cut surface, laser-cut surface with 1 mm planed off, and conventionally planed surface.

9. Ponderosa pine had low wood failure results on the brushed off laser-cut surfaces. This was due to low glue penetration into the TCD layer and the unevenness of the laser-cut kerf.

10. The brushed-off laser-cut surfaces of radial and tangential bonded sugar maple specimens exhibited lower glueline shear strengths.

11. Planing off the TCD layer restored flatness and perpendicularity of the kerf surface. These surface conditions are necessary to improve wood failure and glueline shear strength when joining laser-cut wood substrates.

RECOMMENDATIONS FOR FUTURE RESEARCH

This project generated many questions that could be answered by future research. To facilltate this, the following recommendations have been made:

1. Determine the chemical composition and properties of the TCD layer through materials science research.

2. Determine if alterations in laser-cutting parameters can reduce or eliminate the TCD layer.

3. Further laser-processing research is necessary in order to learn if there is a relationship: 1) between laser cutting parameters and the gluabillty of wood; 2) among laser beam power density, SG of wood, and the gluabillty of wood; 3) among laser-processing parameters, orthotropic wood structure, MC of wood, and the gluability of wood; and 4) among the TCD layer, SG of wood, depth of cut, and the gluability of wood.

4. Conduct comparative research for various other wood species and wood composites.

LITERATURE CITED

1. Arai, T., H. Kawasumi, and D. Hayashi. 1977. Studies on laser machining of wood. 1. Infrared spectra and x-ray diffraction of heat-affected zone. Japan Wood Res. Soc. 23(7):317-321.

2. --, --, and N. Kinoshita. 1978. Three-dimensional graphic analysis to determine optimum laser cutting conditions of (solid wood) board. Japan Wood Res. Soc. 24(5):281-286.

3. --, S. Shimakwa, and D. Hayashi. 1976. Studies on laser machining of wood: measurement and observation of the heat-affected zone. Japan Wood Res. Soc. 22(12):655-660.

4. Babiak, M., J. Orech, and M. Kleskenova. 1986. Temperature distribution in wood heated by scanning Gaussian laser beam. 1. Surface temperature distributions. Drevarsky Vyskum 108:1-15.

5. Barnekov, V. 1991. Laser wood machining research in Europe. In: Proc. The First Inter. Conf. on Automated Lumber Processing Systems and Laser Machining of Wood. Michigan State Univ., East Lansing, Mich.

6. Hattori, N., T. Matano, H. Okamato, and K. Okamura. 1988. Microscopic observations of the solids products deposited on the edge of papers by carbon dioxide laser cutting. Japan Wood Res. Soc. 34(5):417-422.

7. Huber H.A. 1991. The development of ALPS (Automated Lumber Processing Systems) in the USA. In: Proc. The First Inter. Conf. on Automated Lumber Processing Systems and Laser Machining of Wood. Michigan State Univ., East Lansing, Mich.

8. Juza, F. 1989. Cutting wood and nonmetals with CO sub 2 laser beam. Wood Industry 11/12:33-35.

9. Kahn, P.A.A. and K. Mukherjee. 1991. Laser machining of hardwoods, current status and control. In: Proc. First Inter. Conf. on Automated Lumber Processing Systems and Laser Machining of Wood. Michigan State Univ., East Lansing, Mich.

10. -- and --. 1990. Laser cutting of hardwoods. Presented at a symposium at Michigan State Univ., East Lansing, Mich.

11. Klobetzova, T. 1980. Effect of laser radiation on some physical and mechanical properties of wood and wood-based materials. Drevarsky Vyskum 24(4):13-28.

12. McMlllin, C.W., R.W. Conners, and H.A. Huber. 1989. ALPS--A potential new automated lumber processing system. Forest Prod. J. 34 (1):13-20.

13. Mukherjee, K. 1991. Industrial laser machining. In: Proc. First Inter. Conf. on Automated Lumber Processing Systems an Laser Machining of Wood. Michigan State Univ., East Lansing Mich.

14. Orech, J. and F. Juza. 1987. Specific cutting energy and its determination during the interactlon of the laser beam with wood. Drevarsky Vyskum 114:29-40.

15. -- and M. Kleskenova. 1978. Some experience in wood cutting with lasers. Drevarsky Vyskum 23(1):49-62.

16. -- and M. Kleskenova. 1975. Laser as a wood cutting tool. Drevo 30(11):324-326.

17. --, P. Rusnak, and F. Juza. 1988. Model of the formation of the cut in the process of cutting wood by laser beam. Drevarsky Vyskum (119):63-74.

18. Rabiej, R.J. and H.D. Behm. 1992. The effect of clamping pressure and orthotropic wood structure on strength of glue bonds. Wood and Fiber Sci. 24(3):260-273.

19. --, --, and P.A.A. Khan. 1991. Optimizing glueline strength of laser cut hardwoods. Proc. Wood Adhesives 1990. Forest Prod. Res. Soc.

20. --, --, and A. Hughes. 1991. Effect of Orthotropic Structure on Wood Mechanics. Presented at Forest Prod. Res. Soc. Annual Meeting.

The authors are, respectively, Associate Professor, Instructor, and Assistant, Dept. of Engineering Tech., Western Michigan Univ., Kalamazoo, MI 49008-5064. This project was sponsored by the USDA/Cooperative State Res. Serv. and was awared grant #90-34158-5674. It is a continuation of work done under the Eastern Hardwood Util. Project at Michigan State Univ. This paper was received for publication in June 1992.

(C) Forest Products Society 1993.

Forest Prod. J. 43(2):45-54.