COMPOSITES AND MANUFACTURED PRODUCTS

MEDIUM DENSITY FIBERBOARD MADE FROM EUCALYPTUS SALIGNA

ANDRZEJ M. KRZYSIK JAMES H. MUEHL JOHN A. YOUNGQUIST* FABIO SPINA FRANCA

tion of hardboards. Other fiber-based A BSTRACT The production of industrial wood from natural forests is predicted to decline in the composite products, like medium den­ sity fiberboard (MDF), appear to be an future. Factors that will contribute to this decline include changes in land use patterns, attractive alternative for this fiber. depletion of resources in some parts of the world, and the withdrawal of forest areas from industrial production in order to provide for environmental, recreational, and MDF is a nonstructural wood-based other social needs. There is a shortage of information on the suitability of fiber from panel that is composed of wood fibers many plantation-grown species for alternative composite products. This research was bonded together with resin under heat conducted to determine the suitability of plantation-grown Eucalyptus saligna from and pressure. In recent years, great Brazil as a raw material for medium density fiberboard (MDF). Test panels of varying changes have taken place in the MDF thickness (6,13, and 19 mm) were made with 10 percent urea resin and 1.5 percent wax. industry. Production of this product has increased dramatically and new plants Mechanical, water resistance, and dimensional stability properties were tested accord­ are planned worldwide. In 1996, MDF ing to American Society for Testing and Materials standards (ASTM). The results shipments from U.S. plants set another showed that nearly all mechanical properties of the panels at all thickness levels were annual record in an unbroken series, to­ above minimum requirements for MDF as specified in the ANSI-AHA and Euro MDF taling 2.1 million m3, an 8.5 percent in- standards. These results indicate that MDF-type panels can be made from wood fiber crease from 1995. U.S. production for derived from Eucalyptus saligna. Additional work is needed to ascertain the perfor­ 1997 was forecasted to be 3 million m3. mance of MDF panels from this species through pilot- and production-scale trials. Canadian plant capacity was predicted to increase 19 percent, to 1.2 million m3. In 1996, European production of MDF jumped 18 percent to 4.5 million m3, The Food and Agricultural Organi- Brazil is the largest plantation grower continuing an unbroken upward trend zation (FAO) of the United Nations esti- of various species of eucalyptus, with in Europe. mates that the production of industrial 2.7 million hectares. Other countries The popularity of this relatively new wood from plantations will be an increas- that grow this species include South Af­ panel product is due to its ability to be ingly important source of industrial fiber rica, Congo, India, and Burundi. Most of produced in molded form, as well as in worldwide (6). At present, countries like the eucalyptus fiber from Brazil is used straight-edged flat panels, for a host of for pulp and paper production, although South Africa and New Zealand derive industrial markets. MDF is used exten­ some is currently diverted to the produc- sively in factory-assembled and ready- 100 percent of their industrial wood from forest plantations. Other countries that derive a high proportion of industrial wood from plantations include Chile The authors are, respectively, Research Specialist, Forest Products Technologist, and Pro­ (95%), Spain (81%), Brazil and Argen- ject Leader (retired), USDA Forest Serv., Forest Prod. Lab., One Gifford Pinchot Dr., Madi­ son, WI 53705-2398; and General Manager, Votorantim-Siderugica Barra Mansa SA, Capao tina (60%), and Japan (55%). In 1987, Bonito, Sao Paula, Brazil. The use of trade or firm names in this publication is for reader infor- Sedjo (12) predicted that by the year mation and does not imply endorsement by the U.S. Dept. of Agriculture of any product or ser- 2000, half the industrial wood produced vice. This paper was received for publication in August 2000. Reprint No. 9175. *Forest Products Society Member. in Latin American countries would come ©Forest Products Society 2001. from forest plantations. Forest Prod. J. 51(10):47-50.

FOREST PRODUCTS JOURNAL V OL. 51, NO . 10 47 to-assemble furniture, as well as in cab­ The drainage time test method could M ATERIALS AND METHODS inets, underlayment, drawer fronts, be used as an efficient tool for the evalu­ Wood chips of E. saligna were ob­ molding, and countertops. Finishes and ation of eucalyptus hardboard pulps. tained from Votorantim-Siderugica Barra overlays can be used to provide a grain The dimensional stability of hardboard Mansa SA (Forest Business Unit) of pattern typical of lumber, and many made from high-density Eucalyptus mac­ Brazil. Initial moisture content (MC) wood finishing components such as ulata improved as the drainage time in- was 25 percent (percentage based on door edgings, decorative trim, frames, creased, while the opposite effect was ovendry weight). The chips were con­ and cornices are being made from MDF. found for boards made from low-density verted into fibers for fabricating the Moreover, MDF is replacing thin ply- Eucalyptus obliqua (9). MDF panels. The physical characteris­ wood and wet-process hardboard in the In Australia, use of eucalyptus fiber in tics of the fiber were controlled or modi­ production of molded and flush door- reconstituted wood products is concen­ fied by varying the chip retention time skins. trated primarily in the hardboard indus­ within the digester, varying the gap be- New MDF products include generic try, with smaller amounts used in ply- tween the refiner plates, and selecting and proprietary panels. One example is a wood and particleboard. Hardboards refiner plate patterns. super-refined board in which fine fibers made from eucalyptus fiber have major The chips were defibrated in a Sprout- are distributed throughout the board to advantages compared with those made Bauer Model 12-1CP (Andritz, Inc., facilitate deep routing and machining. In from other species. No supplementary Muncy, Pennsylvania) 305-mm thermal some countries, panels are being made bonding resins are needed to provide a mechanical single-disk refiner. Defibra­ from many different hardwood and soft- high level of strength, in contrast to the tion was done in a batch process, with wood species as well as from nonwood­ resins required for softwoods. Because each batch limited to a maximum of 4 based lignocellulosics from raw materi­ eucalyptus fibers are short, they do not kg (ovendry basis) by the capacity of the als such as bagasse and cotton stalks. In flocculate as readily as long fibers. Con­ receiver tank. Before refining, the chips South America and Australia, hardboard sequently, boards made with eucalyptus were placed into a digester, ahead of the panels have been successfully produced fiber have better surface properties and refiner, to soften them for obtaining a from a variety of wood species including are preferred worldwide as the substrate higher quality fiber. The chips going Eucalyptus grandis and E. saligna (3,7). for prefinished hardboard products (7). into the digester were held for 20 min­ In the United States, some panels are be­ Chauhan and Bist (4) concluded that utes under 586 kPa of saturated steam ing produced from recycled fibers from unbarked eucalyptus hybrid (mainly E. pressure. Defibration occurred as the postconsumer wood waste. tereticornis ), including tops, twigs, and chips passed between the rotating and Continuing concerns for panel manu­ branches, is a suitable raw material for stationary plates of the refiner. Sprout- facturers are regional shortages in soft- hardboard manufacture. Oil-tempered Bauer refiner plates (D2B503 type) with wood fiber, the need to find suitable, boards do not require the addition of siz­ surface dams and closed periphery were low-cost hardwood fibers for the manu­ ing agents to the pulp suspension in the used for the refining process; the plate facture of various fiber-based products, semichemical wet process. gap was 0.36 mm. Refining each batch and the need to determine if various fast- Pranda (11) reported that MDF made took approximately 4 minutes (from 3 growing plantation species are suitable from Eucalyptus globulus required higher min. and 45 sec. to 4 min. and 10 sec.). for MDF and hardboard panels. The ob­ amounts of adhesive to reach the same The adhesive (a water-soluble, liquid jective of the research reported here was mechanical properties as those of MDF urea-formaldehyde resin) was obtained to develop the material and process for made from Pinus pinaster. However, from Neste Resins Corporation (North fabricating MDF composite panels from water absorption and thickness swell Bay, Ontario, Canada). The resin, Ba-255, wood fibers of E. saligna. values of MDF from E. globulus were had a solids content of 65 percent, vis­ SELECTED LITERATURE REVIEW higher than those of similar panels made cosity of 0.19 Pa·s at 25°C pH of 7.62, Literature pertaining to the manufac­ from Pinus pinaster. A possible expla­ and SG of 1.281. The wax, Cascowax ture of various fiberboards from differ­ nation for higher resin consumption EW-3100P (a paraffin wax emulsion), ent species of eucalyptus was reviewed may be the higher content of fine parti­ was obtained from Borden, Inc. (Co­ and is briefly discussed here. cles and relatively high specific surface lumbus, Ohio). This material had a sol- area in eucalyptus fibers after defibra­ ids content of 58 percent and pH of 8.3. Products of hydrolysis can affect the tion of the chips. properties of hardboard, but they can be PROCESSING removed from the stock by washing. In EXPERIMENTAL DESIGN laboratory experiments, washing the eu­ Fibers of E. saligna were used to FlBERS calyptus fibrous stock produced in the make MDF with a specific gravity (SG) After defibration, fibers had an ap­ defibrator process increased modulus of of 0.77 in three selected thicknesses: 6, proximate MC of 125 percent. Before rupture (MOR) and internal bond strength 13, and 19 mm. The fiberboards had 10 further processing, the fibers were dried and decreased linear hygro-expansivity percent resin content and 1.5 percent to a final MC of 3 to 4 percent in a of the resulting hardboard. There was wax content (percentage based on solids steam-heated tray dryer at 94°C for 24 also a slight adverse effect on the prop­ content and ovendry fiber weight). Each hours. The drying process caused the fi­ erties of water absorption, thickness panel thickness was considered a repli­ bers to clump together as a result of hy­ swell, and machinability, but the magni­ cated set that consisted of five individual drogen bonding. Therefore, further pro­ tude of change was too small to be of panels. In total, 15 panels were made for cessing was needed to make the fibers concern in end-use applications (8). this experiment. suitable for blending with resin and wax.

48 OCTOBER 2001 The ovendried fibers were processed to 152 mm, which allowed for easy in­ This test was performed in conformance through a hammermill using a 19-mm sertion of the mat into the hot-press. with ASTM D 1037. Linear expansion screen opening. The purpose of the All panels were consolidated using a tests were conducted on length mea­ hammermilling process was to separate manually controlled, steam-heated press. surements made at equilibrium condi­ clumps of fibers, not to shorten fiber The press temperature was approxi­ tions of 50 and 90 percent relative hu­ length. This procedure resulted in a mately 180°C. Maximum panel pres­ midity and 27°C. The linear expansion high-quality fiber furnish with few no­ sure during closing ranged from 3.05 to test was done in conformance with ticeable fines. 6.10 MPa and was lowered to 0.11 MPa ASTM D 1037. Mechanical and physi­ WAX AND ADHESIVE after the target thickness was reached. cal property data for the three panel thicknesses are presented in Table 1. The wax emulsion and urea-formal­ In the manufacture of all panels, a dehyde resin were mixed together using Each value is an average of 10 tests for thermocouple was inserted in the center static bending MOR and MOE, 25 tests a high-speed laboratory mixer. The of the formed mat to determine the core resin-wax mixture was sprayed onto the for internal bond strength, 10 tests for temperature. Press cycles were con- thickness swell and water absorption, wood fiber at 25°C as it rotated in a trolled by the core temperature. drum-type blender. The mixture was ap­ and 10 tests for linear expansion. plied with a single pneumatic spray gun TESTING For reference purposes, the minimum applicator. All the blended furnish was Mechanical and physical property property standard requirements for MDF then hammermilled again. In this case, tests were conducted on specimens cut panels as specified in ANSI A208.2 and the purpose of the hammermilling pro­ from the selected experimental panels. in the Euro MDF standard are included cess was to break up any balls of fiber All panels were weighed and measured in Table 1. The results in Table 1 were formed in blending with wax and resin. and the SG was calculated. The panels statistically analyzed and are reported as The same 19-mm screen opening used were selected on the basis of the target the mean and the coefficient of varia­ to separate the ovendried fiber clumps SG (0.77 ± 0.05) and target thickness. tion. Each stage of the research is pre­ was used here. This method of panel selection allowed sented separately. BOARD MANUFACTURE us to narrow the variability in SG be- RESULTS AND DISCUSSION Mats were hand-formed in a 508- by tween individual experimental panels. Bending strength (MOR) values in- 508-mm deckle box attached to a vac­ Prior to mechanical and physical prop­ creased as panel thickness increased, uum. The fiber was manually forced, us­ erty testing, the specimens were condi­ and values exceeded the Interior ANSI ing a brushing motion, through a 6-mm tioned at 65 percent relative humidity A208.2 MDF standard (10) for all three screen on the top of the box. This al­ and 20°C. Three-point static bending thicknesses. For the Euro standard (5), lowed individual fibers and fiber bun­ MOR and modulus of elasticity (MOE) strength of the 6-mm panels was below dles to pass through the top screen and and internal bond strength tests were the specified minimum bending strength collect at the bottom of the box. When performed in conformance with ASTM value of 40 N/mm2. The 12- and 19-mm all the fiber had been put into the deckle D 1037 standards (2) and ANSI A208.2 panels met the minimum Euro standard box, the mats were manually precom­ standards (10) using an Instron Model requirements. Bending stiffness (MOE) pressed. Depending on the target thick­ TTCLM1 (Instron Corp., Canton, Mas­ values for all panel thicknesses ex­ ness of the board, the average height of sachusetts) testing machine. Thickness ceeded both the ANSI and Euro MDF the mat was 203 to 356 mm. To reduce swell and water absorption measure­ minimum property requirements. Inter­ the mat height and to densify the mat, ments were made by immersing speci­ nal bond strength decreased as panel the mat was cold-pressed. This proce­ mens in water in a horizontal position thickness increased; nevertheless, inter­ dure reduced the mat height to about 127 for 24 hours at ambient temperature. nal bond strength exceeded the ANSI

TABLE 1. — Mechanical and physical properties of panels made from Eucalyptus saligna fiber compared with minimum standards of ANSI and Euro MDF.a Static bending Static bending Internal Thickness swell, Water absorption, Linear MOR MOE bond 24-hr. 24-hr. expansion ------(N/mm2 )------(%)------Formulation (88.5% fiber/10% resin/l.5% wax) b 6-mm panel 37.4 (14) 3140 (18) 1.12 (14) 11 (4) 24 (4) 0.30 (9) b 12-mm panel 39.8 (13) 4010 (10) 0.87 (14) 6 (8) 15 (12) 0.25 (17) b 19-mm panel 42.2 (12) 4420 (11) 0.66 (13) 3 (22) 8 (26) 0.24 (15) c ANSI MDF standard (10) 24.0 2400 0.60 NA NA NA Euro MDF standard (5) d 6-mm panel 40 2600 0.70 22 NA 0.50 d 12-mm panel 35 2500 0.65 15 NA 0.40 d 19-mm panel 30 2500 0.60 10 NA 0.40 a Values in parentheses are coefficients of variation (%). b Panels conditioned at 50 to 90 percent relative humidity. c NA = not specified in test requirements. d Panels conditioned at 35 to 85 percent relative humidity.

FOREST PRODUCTS JOURNAL V OL. 51, No. 10 49 and Euro MDF standard requirements ceed levels specified in the appropriate 5. European Association of Medium Density for each panel thickness. existing standards. Further experimenta­ Fiberboard Manufacturers. 1990. Medium density fiberboard (MDF) industry stan­ The ANSI A208.2 MDF standard tion with other plantation-grown euca­ dard. 2nd ed. D-6300. EAMDFM, Glessen, does not specify performance require­ lyptus species must be conducted to Germany. ments for thickness swell, water absorp­ confirm our findings with E. saligna and 6. Evans, J. 1998. The sustainability of wood tion, and linear expansion. These water allow comparisons. Additionally, pilot- production in plantation forestry. Unasylva scale and full production trials must be 49(192):47-52. sensitivity tests were conducted on the 7. Hillis, W.E. and A.G. Brown. 1978. Euca­ experimental MDF panels to provide a conducted to confirm our laboratory re­ lyptus for wood production. Chapter 14. In: basis of reference for others doing work sults. These research results are promis­ Reconstituted Wood Products. Common- in this area. Maximum thickness swell ing and indicate that plantation-grown wealth Scientific and Industrial Research and water absorption properties are E. saligna can be successfully used for Organization (CSIRO), Melbourne, Victo­ the production of MDF, a value-added ria, Australia. pp. 317-321. specified by ANSI for other fiber-based 8. Holokiz, H. 1971. The effect of stock wash­ products like basic hardboard (1) as 25 panel product. ing on the properties of hardboard. Appita and 35 percent, respectively. The Euro LITERATURE CITED 25(3):194-199. standard specifies maximum 24-hour 1. American Hardboard Association. 1995. Ba­ 9. 1973. Effect of drainage time thickness swell values for 6-, 12- and sic hardboard, ANSI/AHA A135.4-1995. on the properties of eucalyptus hardboard. AHA, Palatine, IL. Appita 26(6):437-443. 19-mm MDF panels as 22, 15, and 10 2. American Society for Testing and Materials. 10. National Particleboard Association. 1994. percent, respectively. The thickness swell 1994. Standard test methods for evaluating Medium density fiberboard. ANSI A208.2- and water absorption values of all the the properties of wood-based fiber and parti­ 1994. NPA, Gaithersburg, MD. MDF panels in our experiment were cle panel materials. Ann. book of ASTM 11. Pranda, J. 1995. Medium density fiber- well below these maximum specified standards. ASTM D 1037-94. ASTM, West boards made from Pinus pinaster and Euca­ Conshohocken, PA. lyptus globulus wood. Part I. Chemical levels. 3. Anonymous. 1977. Brazilian mill uses euca­ composition and specific surface area of de­ lyptus. World Wood 18(4):30-31. CONCLUDING REMARKS fibrated wood. Drevarsky Vyskum 2:19-28. 4. Chauhan, B.R.S. and J.P.S. Bist. 1987. 12. Sedjo, R.A. 1987. Forest resources of the The results from this experiment indi­ Hardboard from unbarked eucalyptus hy­ world: Forests in transition. In: The Global cate that laboratory MDF panels made brid. Composite Wood Branch, Forest Res. Forest Sector: An Analytical Perspective, Inst. and Colleges, Dehra Dun, Uttar from plantation-grown E. saligna can be M. Kallio, D.P. Dykstra, and C.S. Binkley, Pradesh, India. Indian Forester 113(3): eds. John Wiley & Sons, Chichester, West fabricated with panel properties that ex­ 185-190. Sussex, UK.

50 OCTOBER 2001