Effect of particle geometry on the properties of binderless particleboard manufactured from oil palm trunk

Materials and Design 31 (2010) 4251–4257

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier .com/locate /matdes

Effect of particle geometry on the properties of binderless particleboardmanufactured from oil palm trunk

Rokiah Hashim a,*, Norhafizah Saari a, Othman Sulaiman a, Tomoko Sugimoto b, Salim Hiziroglu c,Masatoshi Sato d, Ryohei Tanaka e

a Division of Bio-resource, Paper and Coatings Technology, School of Industrial Technology, Universiti Sains Malaysia, 11800 Minden, Penang, Malaysiab Japan International Research Center for Agricultural Sciences, 1-1, Owashi, Tsukuba, Ibaraki 305-8686, Japanc Department of Natural Resource Ecology and Management, Oklahoma State University, Stillwater, OK 74078-6013, USAd Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1, Yayoi, Bunkyo-ku, Tokyo 113-8657, Japane Forestry and Forest Products Research Institute (FFPRI), Tsukuba, Ibaraki 305-8687, Japan

a r t i c l e i n f o a b s t r a c t

Article history:Received 24 February 2010Accepted 3 April 2010Available online 9 April 2010

Keywords:A. CompositesB. Particulates and powdersE. Physical

0261-3069/$ - see front matter � 2010 Elsevier Ltd. Adoi:10.1016/j.matdes.2010.04.012

* Corresponding author. Tel.: +60 46535217; fax: +E-mail address: [email protected] (R. Hashim).

Experimental binderless composite panels were manufactured using fine particles and strands of oil palm(Elaeis guineensis) trunks. Modulus of rupture (MOR), internal bond strength (IB), dimensional stability,and surface roughness of the panels made with a target density of 0.80 g/cm3 were evaluated. Strand typesamples had MOR and IB values of 24.95 and 0.95 MPa, respectively. Corresponding values for the fineparticle type samples were 4.04 and 0.49 MPa. Panels made from strands met MOR requirement statedin Japanese Industrial Standard (JIS). Enhanced bonding between strands observed by micrographs takenusing scanning electron microscopy (SEM) also supported the findings. However, the samples having fineparticles had lower MOR values than minimum requirement listed in JIS. Strand type panels had 41.6%thickness swelling which is only 4.6% lower than that of the panels made from fine particles. It appearsthat dimensional stability of both types of panels exhibited insufficient results according to JIS. Surfaceroughness quality of the samples made from fine particles had average surface roughness values compa-rable to those of panels made in past studies. Based on initial results of this work, raw material from oilpalm trunks can have some potential to be used to manufacture binderless panels without using anyadhesives. This study revealed that mechanical and physical properties of such experimental panels wereinfluenced by the particle geometry. It would be important to consider possible addition of chemical orwax in the particles to improve their dimensional stability in further studies.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Oil palm (Elaeis guineensis) which is native species from Africahas the largest plantation land among other agricultural crops inMalaysia. It has a significant commercial value in the form of oilthat can be produced from the mesocarp of the fruit. CurrentlyMalaysia is one of the largest producers of palm oil in the world.Increasing land area of oil palm plantation has been leaving sub-stantial amount of residue in harvesting sites. It is estimated thatoverall oil palm industry generates at least 30 million tonnes of lig-nocellulosic biomass per year in the form of trunks, fronds, emptyfruit bunches, and leaves [1]. Unfortunately the resource is notused very effectively, open burning and land filling are commonpractices to eliminate oil palm residue which cause environmentalpollution and put adverse impact on the ecosystem.

ll rights reserved.

60 46573678.

Oil palm being lignocellulosic material that could be consideredto produce value-added composite panels similar to other non-wood resources such as bagasse, wheat straw, or kenaf [2–5]. Com-posite panels including particleboard and fiberboards are widelyused as substrate for thin overlay to manufacture furniture units.

Manufacture of wood composite panels normally requires usingdifferent types of binder to have proper physical and mechanicalcharacteristics of the final product. The choice of the adhesive nor-mally depends on the end use of the composites. Urea formalde-hyde is the most common adhesive used in the industry [6]. It isalso the least expensive compared to other wood adhesives.Although it has a low cost, it still makes up about 60% of overallcost of particleboard production even if it is used only 8–10% basedon oven dry furnish weight [7]. Formaldehyde based adhesive suchas urea formaldehyde poses a problem of formaldehyde emissionfrom panels that could cause health concerns [6,7]. Therefore,manufacture of composite panels made without using any resinsknown as binderless panels is an alternative method to keep lowcost of the final product without having any adverse health

Fig. 2. Average modulus of rupture values of the samples.

Fig. 3. Average internal bond values of the samples.

4252 R. Hashim et al. / Materials and Design 31 (2010) 4251–4257

influence. Shen [8] made binderless boards using sugar cane andsorghum stalks in their natural form. The process for manufactur-ing binderless composite products from sugar containing such lig-nocellulosic materials was patented. Bonding of the particles wasaccomplished by the existence of free sugar, carbohydrates, or sac-charides that served as bonding and bulking agent pressed at atemperature of 180 �C. Mobarak et al. [9] studied the mechanismof binderless panels made from lignocelluloses of bagasse. Thework showed bonding of particles were achieved as a result ofthe ability of the particles to compress closely together whichcan be attributed by the high lumen to cell wall ratio of bagasse.

Information on binderless particleboard from oil palm trunk isvery limited and no solid data on properties of the panels havebeen reported yet. Suzuki et al. [10] prepared binderless boardsfrom oil palm fronds from steam exploded pulps. Similar techniquehas also been reported by Laemsak and Okuma [11] who at-tempted to use steam exploded oil palm fronds to manufacturebinderless medium density fiberboard.

Particleboard is wood based composite basically consists of par-ticles of varying shape and size bonded together with an adhesiveand consolidated under heat and pressure [12]. Particle geometryincluding shape and size is a major parameter which can create asignificant impact on the properties of the boards [13]. Accordingto Suchsland [14], particle geometry plays more significant roleon development of board properties than the actual mechanicalproperties of the fiber types panel. Variation of particle shapeand size significantly influenced overall panel properties [13].Study by Miyamoto et al. [15] showed that particle shape affectedthe linear expansion of particleboard. Sackey et al. [16] pointed outthat the fines content and the ratio of all particle-size fractionsstrongly influenced the internal bond strength of the samples.

The objective of the work was to investigate properties of pan-els made from oil palm trunk without using any binders focusingon the effect of particle geometry in terms of size and shape. Insuch case, binding naturally is developed as a result of pressingof the mat at high temperature so that free sugars, carbohydratesor/and saccharides or any other chemical present in oil palm canbe activated to form natural binding agents. Experimental panelswere produced using commercial manufacturing parameters. Bothmechanical and physical properties of such composite panels made

Fine Particles

Fig. 1. Samples of fine p

of different particle geometry were evaluated. Also microscopicstudy was carried out to determine bonding quality of the experi-mental panels as function of particle geometry.

Strands

articles and strands.

Fig. 4. Average thickness swelling values of the samples.

Fig. 5. Average water absorption values of the samples.

Strands

Fine particles

Fig. 6. Typical surface roughne

R. Hashim et al. / Materials and Design 31 (2010) 4251–4257 4253

2. Materials and methods

2.1. Panel manufacture and sample preparation

Oil palm trunks were supplied by a local plantation in NorthernMalaysia. The trunks were sawn into round section with 10 cmthickness before these sections were chipped into small pieceswith approximate dimensions of 5 cm � 2 cm � 1 cm employinga laboratory type hammermill. The chips were dried to 8% moisturecontent in a laboratory oven. Strands were produced manually byseparating vascular bundles with a length of 3–5 cm from the chipas shown in Fig. 1. Oil palm chips were reduced into smaller parti-cles and later into fine particles using a hammermill and a willeymill, respectively, also illustrated in Fig. 1.

A total of 20 panels, with a dimension of 20.5 cm � 20.5 cm, 10from strands and 10 from fine particles were made for the experi-ments. All mats were manually formed using a frame. In the case ofstrand type panels, strands were laid perpendicular to each otherin the form of eight layers similar to oriented strand board config-uration. The mats were then pressed in computer controlled hotpress using a pressure of 5 MPa and a temperature of 180 �C for20 min. Target density and panel thickness were 0.80 g/cm3 and0.48 cm, respectively. After samples were cut from each panelbased on Japanese Industrial Standards, JIS A-5908 [17], they wereconditioned in a chamber with a temperature of 20 �C and relativehumidity of 65%.

2.2. Test procedure

Nine MOR and IB samples were prepared from each panel. Bothtypes of tests were carried out on an Instron Testing System ModelUTM-5582 equipped with a load cell having a capacity of 1000 kg.Six samples with a size of 5 cm � 5 cm were used to evaluate

ss profiles of the samples.

fine strands

Fig. 7. Surface of binderless particleboard samples made from fine particles and strands of oil palm trunk.

Fig. 8. Average surface roughness values of the samples.

4254 R. Hashim et al. / Materials and Design 31 (2010) 4251–4257

dimensional stability of the samples in terms of their thicknessswelling (TS) and water absorption (WA) for 24 h water soaking.Thickness of each sample was measured at four points midwayalong each side 1 cm from the edge before and after they weresoaked into water.

Five samples from each panel type with dimension of5 cm � 5 cm were used to evaluate their surface quality. The profi-lometer consisted of a main unit and a pick-up which had a skid-type diamond stylus with 5-lm tip radius and 90� tip angle wasused for surface roughness measurement of the samples. Variousroughness parameters such as average roughness (Ra), meanpeak-to-valley height (Rz) and maximum roughness (Rmax) can becalculated from the digital information. Definition of these param-eters presented was discussed in detail in previous studies [18,19].

2.3. Scanning electron microscopy

Scanning electron microscopy (SEM) was employed to charac-terize the morphology of raw materials and the condition of fibersand parenchyma cells in the panels related to bonding quality be-tween the materials. Micrographs were taken from the cross sec-tion samples of 0.5 cm � 0.5 cm samples from fine particles andstrand samples. The samples were coated with gold by an ion sput-ter coater (Polaron SC515, Fisons Instruments, UK). A ScanningElectron Microscope LEO Supra 50 Vp, Field Emission SEM, Carl-Zeiss SMT, Oberkochen, Germany was used for microscopic study.The SEM analysis was extended to obtain the elemental composi-tion of the fine and strand particles by means of energy dispersiveX-ray analysis (SEM-EDXA).

3. Results and discussion

3.1. Modulus of rupture and internal bond strength

Figs. 2 and 3 illustrate MOR and IB strength values of the panelsmanufactured from strands and fine particles, respectively. Modu-lus of rupture (MOR) and IB values of 24.95 and 0.93 MPa werefound for the samples made with strands. These values were 84%and 49% higher than those made from fine particles. The strandsare made up mostly vascular bundles with parenchyma cells at-tached together. Larger surface contact area between strandsresulting in a better glueline probably would be responsible forbetter strength characteristics of such panels. Effect of particlegeometry on MOR and IB of experimental panels were also inves-tigated in various previous studies and it was concluded that largercontact area of particles showed higher strength properties of the

samples [20,21]. A lower value of MOR and IB of panels made fromparticle length less than 2.54 cm were determined by Lehmannand Geimer [22]. Badejo [23] also showed longer and thinnerstrands or a particle resulted in higher MOR and IB values of thepanels. It seems that the findings in this work are consistent withresults of above studies. Although no adhesive was used for panelmanufacture, both types of panels showed satisfactory strengthproperties based on Japanese Industrial Standard [17]. Viscoelasticnature of oil palm particles from trunk which are rich in carbohy-drate content would be considered one of the major reasons forhaving acceptable mechanical properties of the samples [24,25].Also possible chemical interaction between particles due to theircarbohydrate content by application of temperature and pressureduring pressing of the mats could have enhanced the strengthproperties of the samples [2,4]. Similar attributable influence ofself bonding of raw material was also determined in evaluationof strength properties of experimental panels made from particleswith high hemicelluloses content [8,25].

3.2. Physical properties

Thickness swelling and water absorption of the samples areshown in Figs. 4 and 5. Both types of panels resulted in relativelyhigh thickness swelling values. Strand type panels had 41.6% thick-ness swelling which is only 4.6% lower than that of the panelsmade from fine particles. Particle geometry, large surface area ofstrands having better contact to each other in the case of strand

R. Hashim et al. / Materials and Design 31 (2010) 4251–4257 4255

type panel is considered major parameter of having lower thick-ness swelling of such samples. Some additional processes such asheat treatment, chemical treatment, or steam treatment wouldbe considered to enhance such shortcoming of the binderless pan-els [2,3,9]. In this work, maximum temperature of 180 �C was usedfor panel pressing. Based on the results of a study carried out byOkuda and Sato [4], it was suggested that higher press temperatureimproved dimensional stability of binderless panels. Another pos-sibility to improve this shortcoming is probably by the addition ofwax. However, the addition of wax could reduce strength proper-ties of the panels. Using 1% or 1.5% wax in the panels could be ta-ken into consideration to enhance their dimensional properties[26]. Neither of the panels met minimum thickness swellingrequirement of 12% stated in JIS. Water absorption of the samplesalso followed similar trend of thickness swelling with high values.

Figs. 6 and 7 show typical roughness profiles of the samples andsurface of the panels. Average Ra, Rz and Rmax values of the samplesmade from strands and fine particles were 12.9 lm, 36.7 and

Compressed Fibers

Starch granules

Compressed Fibers

A

C

B

Fig. 9. Scanning electron micrographs of cross section of binderless particleboardmade from fine particles of oil palm trunk.

143.7 lm, and 6.1, 45.6 and 90.6 lm, respectively, as shown inFig. 8. Average Ra values of commercially manufactured particle-board made from rubberwood was found as 8.2 lm in a previouswork [18]. A typical particleboard has a Ra values ranging from 3to 10 lm depending on raw material, particle size and surface den-sification [27]. In the case of the samples made from fine particlestheir roughness values were found within above range. Coarser andrigid structure of vascular bundles used in strand type panel couldbe main reason of having relatively high Ra, Rz and Rmaz values forstrand type panels. In a previous work Ra value of strand type pan-els made from empty fruit bunch of oil palm was found as 13.8 lm[28] which is comparable to the results in this work. As mentionedpreviously no sanding was applied to the surface of the samples. Ifthey were sanded their surface quality would have been enhanced.It appears that panels made from fine particles and strands are notinferior in terms of their surface quality as compared to commer-cially manufactured and laboratory made panels evaluated in paststudies stated above. Based on the findings in this work such

Compressed Fibers

Compressed vascular bundle

Parenchymatous ground tissue

Parenchymatous ground tissue

Parenchymatous ground tissue

Compressed Fibers

A

B

C

Fig. 10. Scanning electron micrographs of cross section of binderless particleboardmade from strands of oil palm trunk.

strands

fine

Fig. 11. EDXA spectra of cross section of binderless panels made from fine particles and strands of oil palm trunk with their elemental compositions.

4256 R. Hashim et al. / Materials and Design 31 (2010) 4251–4257

panels would have some potential to be used as substrate for over-lay for furniture manufacturer without having substantial adverseimpact on their overall quality.

3.3. Scanning electron microscopy analysis

The interphase morphology of the binderless panels was inves-tigated as a function of particle size and shape. The SEM micro-graphs of panels made from fine particles and strands are shown

in Figs. 9 and 10, respectively. The oil palm has mostly parenchymacells and vascular bundles. Each vascular bundle consists of afibrous sheath, phloem cells, and xylem and parenchyma cells aspreviously described [24]. In both Figs. 9 and 10 the compressedcell walls during pressing were observed [5,7]. The SEM of the pan-els made from fine particles revealed a uniform homogenous blendof the fiber and parenchyma cells which were compressedtogether. This resulted in smoother surface of the panels madefrom the fine particles. The presence of starch granules either

R. Hashim et al. / Materials and Design 31 (2010) 4251–4257 4257

compressed or in its natural form of oval or elliptical shape mainlywith cut ends can be seen between the fibers. Previous studyshowed that starch from oil palm trunk melts at 67.3 �C. At hightemperature the hydrogen bonding system in starch and watermolecules present in the particles will breakdown providingpseudoplastic flow behavior contributing to the adhesion of panels[25]. The pressed vascular bundles and parenchyma can be clearlyobserved in samples made from strands as illustrated in Fig. 10.Each strand consists of vascular bundle surrounded by the paren-chyma as shown in Fig. 10a and b. Fig. 10c also shows thick wallvascular bundle and thin wall parenchymatic tissue. It appears thatlong fine strands made up of the vascular bundle contributed tohigher MOR, IB and improved thickness swelling values as com-pared to those samples made of fine particles [24]. The bondingas can be seen from SEM micrographs is generally in the form ofmechanical interlocking between fiber [13] in addition to the otherchemical bonding such as hemicelluloses component being hydro-lyzed due to heat and pressure [4,8,9]. The orientation of the com-pressed vascular bundles strands also influenced the mechanicalproperties [13]. The SEM results revealed that overall bonding ofthe binderless panels can be enhanced using strands.

The inorganic constituents of the binderless panels made fromfine particles and strands of oil palm trunk were studied by scan-ning electron microscopy equipped with energy dispersive X-rayanalysis. Fig. 11 shows the energy dispersive X-ray analysis(SEM-EDXA) spectra with the elemental composition of fine parti-cles and strands of oil palm trunk. No specific difference was notedbetween the two samples. Both specimens had carbon and oxygenin a large weight percentage while little amount of potassium,chlorine, calcium and silicon was determined [29]. Due to highweight percentage of oxygen it was postulated that potassium,chlorine, calcium and silicon could exist in oxide form [30]. It ap-pears that the presence of silicon may also influence the interfacialadhesion between the particles.

4. Conclusions

Effect of particle geometry on some of the basic properties ofbinderless panels made from oil palm trunks was investigated.The results showed that particle size and shape have an influenceon the properties of the samples. Overall the panels met strengthrequirements based on JIS standards. However, thickness swellingof both types of samples needs to be improved by using eithersome kind of chemical, steam treatment or adding some wax inthe furnish. Microscopic study also revealed that panels made fromthe strands had better bonding characteristics supporting en-hanced mechanical properties of such panels.

Acknowledgements

We would like to acknowledge Universiti Sains Malaysia for theFellowship grant awarded to Ms Nohafizah Saari and Japan Inter-national Research for Agricultural Science (JIRCAS) for partiallysponsoring the research.

References

[1] Sumathi S, Chai SP, Mohamed AR. Utilization of oil palm as a source ofrenewable energy in Malaysia. Renew Sust Energy Rev 2008;12:2404–21.

[2] Widyorini R, Xu J, Watanabe T, Kawai S. Chemical changes in steam-pressedkenaf core binderless particleboard. J Wood Sci 2005;46:296–302.

[3] Halvarsson S, Edlund H, Norgen M. Manufacture of non-resin wheat strawfibreboards. Ind Crop Prod 2009;29:437–45.

[4] Okuda N, Hori K, Sato M. Chemical changes of kenaf core binderless boardsduring hot pressing (II): effects on the binderless board properties. J Wood Sci2006;52:249–54.

[5] Sulaiman O, Salim N, Hashim R, Yusof LHM, Razak W, Yunus NYM, et al.Evaluation on the suitability of some adhesives for laminated veneer lumberfrom oil palm trunks. Mater Design 2009;30:3572–80.

[6] Hashim R, Hamid SHA, Sulaiman O, Ismail N, Ibrahim MH, Jais H, et al.Extractable formaldehyde from waste medium density fibreboard. J TropForest Sci 2009;21:25–33.

[7] Hashim R, Sarmin SN, Sulaiman O, Yusof LHM. Effects of cold setting adhesiveson properties of laminated veneer lumber from oil palm trunk in comparisonwith rubberwood. Eur J Wood Prod 2010. doi: 10.1007/s00107-009-0405-2.

[8] Shen KC. Process for manufacturing composite products from lignocellulosicmaterials. United States patent 4627951. 1986.

[9] Mobarak F, Fahmy Y, Augustin H. Binderless lignocelluloses composite frombagasse and mechanism of self bonding. Holzforshung 1982;36:131–5.

[10] Suzuki S, Shintani H, Park SY, Saito K, Laemsak N, Okuma M, et al. Preparationof binderless boards from steam exploded pulps of oil palm (Elaeis guineensisJaxq) fronds and structural characteristics of lignin and wall polysaccharides insteam exploded pulps to be discussed for self-bindings. Holzforshung1998;52:417–26.

[11] Laemsak N, Okuma M. Development of boards made from oil palm frond II:properties of binderless boards from steam-exploded fibers of oil palm frond. JWood Sci 2000;46:322–6.

[12] Charles C. Wood particleboard and flakeboard: types, grades, and uses. Generaltechnical report FPL; GTR-53. Madison, USA: USDA Forest Service, ForestProduct Laboratory; 1986. p. 1–9.

[13] Frybort S, Mauritz R, Teischinger A, Muller U. Cement bonded composites – amechanical review. Bioresources 2008;3:602–26.

[14] Suchsland O, Woodson GE. Fiberboard manufacturing practices in the UnitedStates. Agriculture handbook no. 640. Washington, DC: USDA Forest Service;1990.

[15] Miyamoto K, Nakahara S, Suzuki S. Effect of particle shape on linear expansionof particleboard. J Wood Sci 2002;48:185–90.

[16] Sackey EK, Semple KE, Oh SW, Smith GD. Improving core bond strength ofparticleboard through particle size redistribution. Wood Fiber Sci2008;40:214–24.

[17] JIS – A 5908. Particleboards. Japanese Standards Association, 1-24 Akasaka, 4Minato-ku, Tokyo 107, Japan. 2003.

[18] Hiziroglu S, Jarusombuti S, Fueangvivat V. Surface characteristics of woodcomposites manufactured in Thailand. Build Environ 2004;39:1359–64.

[19] Sulaiman O, Hashim R, Subari K, Liang CK. Effect of sanding on surfaceroughness of rubberwood. J Mater Process Tech 2009;209:3949–55.

[20] Sumardi I, Ono K, Suzumi S. Effect of board density and layer structure on themechanical properties of bamboo-oriented strandboard. J Wood Sci2007;53:510–5.

[21] Bhagwat S. Physical and mechanical variations in cottonwood and hickoryflakeboards made from flakes of three sizes. Forest Prod J 1971;21:101–3.

[22] Lehman WF, Geimer RL. Properties of structural particleboards from Douglas-fir forest residues. Forest Prod J 1974;24:17–25.

[23] Badejo SOO. Effect of flake geometry on properties of cement-bondedparticleboard from mixed tropical hardwoods. Wood Sci Technol1988;22:357–70.

[24] Lim SC, Gan KS. Characteristics and utilization of oil palm stem. Timbertechnology bulletin no. 35. Malaysia: Forest Research Institute Malaysia; 2005.

[25] Noor MAM, Dos AMM, Islam MN, Mehat NA. Physico-chemical properties of oilpalm trunk starch. Starch-Stacke 1999;52:293–301.

[26] Roffael E, Schneider T, Dix B, Buchholz T. Zur hydrophobierung vonmitteldichten fasserplatten (MDF) mit paraffinen Teil 1: Einfluss derchemischen zusammensetzung des paraffin und des emulgatortyps auf diehydrophobierung von MDF. Holz Roh Werkst 2005;63:192–203.

[27] Cai Z, Wu Q, Lee JN, Hiziroglu S. Influence of board density, mat construction,and chip type on performance of particleboard made from eastern red cedar.Forest Prod J 2004;54:226–32.

[28] Ratnasingam J, Nyugen V, Ioras F. Evaluation of some finishing properties of oilpalm particleboard for furniture application. J Appl Sci 2008;8:1786–9.

[29] Tono Y, Miyafuji H, Shibata M, Saka S. Characterization of inorganicconstituents in oil palm (Elaeis guineensis) by SEM-EDXA. J Jpn Inst Energy2007;86:973–7.

[30] Chin LH, Hameed BH, Ahmad AL. Process optimization for biodiesel productionfrom waste cooking palm oil (Elaeis guineensis) using response surfacemethodology. Energy Fuel 2009;23:1040–4.